On Bicycles, and.... what else is there?

simple Strava to CSV ride decoder with Perl

2012-01-28T20:51:00.000-08:00

Slight detour from my project... Just wanted confidence I could decode JSON data, at least in Perl.

I wrote a little decode for Strava ride data using Perl. The code, which I call "Strava_to_csv", requires a command line option specified as follows:

Strava_to_csv -activity activity-number

where activity-number is the number of the activity.

It's nothing fancy, and not very robust. It expects all data to be scalar except for "latlng", which is an array of two numbers, the first the latitude, the second the longitude. It worked for me.

This code uses the JSON module version 2.53 from CPAN. There's a lot of stuff there about encoding type, but since Strava is unicode, it worked simply.

The code is written for Linux and maybe OS/X: it uses a system call to the "wget" command to download the URL (a Perl package for HTML could be used instead).

Anyway, here's my code. This hardly took any effort at all, maybe 30 minutes, which makes me wonder why it's taken me until now to get around to doing it. Since my blog has such narrow columns many of the lines may wrap around (sorry).

Okay, back to Java...


#! /usr/bin/perl
use strict;
use JSON;
use Getopt::Long;

my $activity;

my %options = (
  "activity=s" => \$activity
);

die "$0 : illegal command line options specified: @ARGV\n"
  unless GetOptions( %options );

die("$0: must specify activity with -activity option")
  unless(defined $activity);

my $url = "http://app.strava.com/api/v1/streams/$activity";

# slurp
undef $\;

open FP, "wget -O - $url |"
  or die("ERROR opening url $url\n");

my $s = ;

my $json = JSON::PP->new;

my $data = $json->decode($s);

# check to make sure array lengths are the same

my $ndata;
my $k0;
my @keys = keys %$data;
for my $k ( @keys ) {
  my $l = scalar @{$data->{$k}};

  die("zero length data element $k found\n")
    unless ($l);

  unless (defined $ndata) {
    $ndata = $l;
    $k0 = $k;
  }
  elsif ($l != $ndata) {
    die("keys $k0 ($ndata) and $k ($l) reference arrays of different length\n");
  }
  # special case: we expect latlong to be an array of length 2
  if ($k eq "latlng") {
    unless((ref $data->{$k}->[0] eq "ARRAY") && (@{$data->{$k}->[0]} == 2)) {
      die("$k expected to reference data arrays of length 2\n");
    }
  }
  elsif (ref $data->{$k}->[0] ne "") {
    die("data element $k expected to be scalar; reference found instead.");
  }
}

# print label line
my @labels;
for my $k ( @keys ) {
  push @labels, ($k eq "latlng") ? "lat,lng" : $k;
}
print join(",", @labels), "\n";

# print data lines
for my $n ( 0 .. $ndata - 1 ) {
  my @data;
  for my $k ( @keys ) {
    if ($k eq "latlng") {
      push @data, @{$data->{$k}->[$n]};
    }
    else {
      push @data, $data->{$k}->[$n];
    }
  }
  print join(",", @data), "\n";
}

writing my Strava Android app, part 2

2012-01-28T08:54:00.000-08:00

More progress...

First, I did more sketch work on my proposed pages for the app. One of these involves data plots: plotting the altitude, and on a separate plot, speed from a Strava activity on a graph. This is the biggest challenge of my GUI design: the rest consists of a text widget, a bunch of button widgets, and some labels, with more simple widgets in the "configuration" screen. Plots with limited pixels are a challenge: I believe the screen resolution is only 480 × 800 for which only a small subset, for example 400 × 100, will be available for each plot. Not so bad, actually, but real estates needs to be used efficiently. I'll avoid doing anything fancy: no scrolling or zooming, for example. The goal is just to identify portions of the ride in a clearly identifiable fashion. For this I can either do my own plots with graphic primitives or use a more general purpose plotting package. I'm not yet decided on this.

Second, I looked into the Strava API itself (it's a work in progress so I resist putting the link here). It's not nearly as bad as I feared. The Strava API consists of specific GET and PUT transactions, using the appropriate URL's, to read and write JSON data to the servers. JSON stands for "JavaScript Object Notation". The advantage of this is it is effectively language-indendent. For example, CPAN has a nice JSON package for Perl, my favorite scripting language. However, I will resist the temptation to do this project in Perl: Java is definitely the way to go. Perl is great for quick-and-dirty projects but for stuff intended for distribution on Android Market I should toe the line and follow best practices, which means Java. (Note: this won't stop me from at some point whipping together a quick script in Perl to export ride data to CSV format, something which I'll find personally useful). Previously I'd been a bit intimidated because the Strava API examples are in Ruby, a language of which I know virtually nothing. But Java should work equally well. The challenge with the API will be writing data: I still need to experiment with that. Worst case is the app will write data to a new activity, then leave it to the user to delete the old one to avoid duplication. But I strongly suspect I'll be able to work out activity deletion or even direct data substitution as well.

Here's an example of what version 1 of the Strava API produces: JSON data of a recent short ride I did. It's fairly straightforward, if inefficient. For example, were I to want to download all data for all crossings of the Golden Gate Bridge for statistical analysis (which would be interesting, to get a statistical distribution of speeds), that would be extremely slow; I'd want a compressed data format like FIT (used by Garmin: a Perl library and a link to the Garmin SDK are here) for that purpose. But for a single ride the rich text format provided by JSON is fine.

As an aside, version 2 of the API claims to allow data to be filtered before download (similar to Garmin "smart sampling"). This should help a lot on the bandwidth hit. But using FIT or similar would probably be at least a 10× improvement.

There is another issue: user authentication. Hopefully this can be handled by the standing Strava app. But I may need to add a log-in page to mine. This isn't something with which I've dealt before.

On the number-crunching side, Java is a fairly simple, sequential language and the sequential array processing should be simple enough.

So things are looking okay. The goal on any programming job is to take complex tasks and break them down into simpler sub-tasks. Then take these sub-tasks, and if necessary, break them down into even simpler sub-tasks. You keep breaking tasks down until the remaining tasks are so simple, the actual coding of them is trivial. It's a lazy shortcut to try to take too much in one bite: to do a complex task directly, and attempting to do so invariably turns the project into a mess. The actual coding part for each sub-task should be simple. The challenge is in organizing the tree of tasks, in knowing exactly what you want to do before you try doing it.

writing my Strava app: part 1

2012-01-27T19:49:00.000-08:00

The first step in writing my Strava app was to make a note of exactly what I wanted it to do. Well, perhaps not exactly: I don't know the capabilities yet of the Strava API. In particular, does the Strava API allow you to replace the data on a ride? Apparently it does, since StravaHaHa allows modification of ride data. Honestly I don't know where the API is documented, but I am deferring that until later. No rush.

So then when I knew what I wanted the program to do, I set out to design some pages. So I measured the screen on my HTC Incredible phone and found it was 5 cm wide by 8 cm high. So I set up an image file using GIMP with 500 pixels wide by 800 pixels high, and for each screen, began "sketching out" how I wanted the page to look. It's different assembling a page from a widget tool kit, so details aren't critical here: I just the look and feel of each page.

When I got tired of doing that, and I was on-line, I started downloading some of the stuff I want for writing the app. For example, this tutorial page provides some nice guidance. First I went to upload my Ubuntu version on my laptop, but my CD player is being cranky, and I can't get the installation CD (written from my work machine: shhh!) to run. Hmm... I'm not going to invest any money in this laptop, however, as it's from 2007: ancient. Instead I think I'm ready to dip my toes in an Apple's OS/X. I'll work on that...

The tutorial recommends Eclipse as a development platform. An alternative is to use a simple text editor. But that seems so old-school: better to use a proper development environment. So I'm just finishing up downloading Eclipse now.

Next I need to download the Java Development Kit. Android apps can be written in C, but Java is more common, and I should follow the best practice. I really like Java, as I noted. But it's been quite a long time since I've done any Java coding.

So time for a Java tutorial. That's easy: Oracle (formerly Sun) has some nice tutorials. So I'm going through those now.

A lot needs to come together for this to work. Honestly I'm not sure I'm going to make my deadline. But I need to stick to it, and by writing here my intent puts some pressure on me to make it work. If I can, I'll feel pretty good about it.

added: Sure enough, Eclipse crashed and burned on my system, which is Ubuntu 9.04: it states it requires 10.04 at least. Hopefully I can get my CD drive working so I can install 11.10. Otherwise I need to wait for that Apple.

writing an Android app

2012-01-27T06:26:00.000-08:00

I've long wanted to write an Android app so now seems a fairly good time to give that a shot. I want to port over some Perl stuff I've done, not directly necessarily, probably in Java, which I've done some work in before but long ago. I really liked Java, though: a proper object oriented environment rather than the layered kludge-fest which is C++, or for that matter "object-oriented Perl", neither of which I've ever liked. I end up with hybrid object oriented - sequential code. Better to have the environment optimized for the object oriented paradigm and stick with that.

Step one didn't go too well... upgrade my Ubuntu to the 11.10, but then my CD drive on my old Thinkpad isn't working well. I tried to clean it but now it's working worse...

Anyway, I got a book on the subject ( Android Applications for Dummies ), but am not waiting for that to arrive. There's plenty of on-line reference material and tutorials. Book form is just more convenient for my train commute, and there's something to be said for the coherence of printed books. When Cara and I were in New Zealand we had an old Lonely Planet printed, a new one on Kindle, and after one attempt at the Kindle version we were back to the old printed copy the rest of the time. Kindles are fine for linear access, or in principle for text-based searching, but for random access of pages with a good index and table of contents printed still wins.

First thing on any programming project is to clearly define what you want to do, so that's where I'll start. I find once you have that down, the actual coding becomes nearly trivial, the details of the language barely relevant.

I want to use the Strava API for this project. Cosmo's done some nice stuff with it. For example, his TCX exporter is my method of choice for getting data off Strava, better than the native GPX exporter Strava eventually added.

Cosmo's stuff is web-based, which is another option. That's another option: it would also give me a chance to expand my web skills, which would help for the Low-Key Hillclimbs. It would be fun to write something to browse historical data. But that's a second priority, and time is so limited as it is.

Goal for this project to be functional is 29 Feb.

rating New Zealand climbs

2012-01-26T17:00:00.000-08:00

I've described a decent number of climb profiles from my trip to New Zealand, with profile data collected from a Garmin Edge 500 or, in the case of Haylocks, by Strava using topological data in conjunction with positions recorded by the Strava Android app. It's then obvious to ask: how do the climbs compare? Which was the toughest?

Good question.

First, some data. I plot here the altitude gained versus the climbing distance for each of the climbs. I define "the climb" as the portion of road which maximizes my climb rating, which I've described in this blog before. Here's the result, where I normalize each by the number associated with Old La Honda road:

What makes a climb tough? To me, it's continuous time spent at a big grade, with a lot of altitude gained.

click on image for expanded view

profile of Haylocks from Maori church. Ignore wiggles, a mapping artifact; climb is continuously up. Click on image for larger view.

Using an objective combination of grade, continuity, and meters of elevation difference I'd have to rank Long Bay Road as the toughest. Haylocks Road was definitely the steepest: it was an effort just to keep the bike moving on that one, but it was fairly short. I did have the advantage on Haylock of having not climbed it continuously: I stopped at the junction with Hamiltons, went up Hamiltons, then descended back to the junction and did the portion of Haylocks from there. But even accounting for that I felt Long Bay was the greater endurance challenge.

click on image for expanded view

You can see from the first plot that both sides of Crown Range Road gained more vertical than Long Bay. But the east side of Crown Range is gradual, only a few percent, for a considerable distance. Long kilometers on the road can be difficult and tiring, but the difficulty I'm talking about here isn't about hours in the saddle, it's about the portion of the difficulty due to the climbing itself. The western slope of Crown Range is far steeper. However, there is considerable recovery provided: approximately 5 km where the grade barely touches 7%, including descending. These kilometers dilute the rest of the climb, where the grade sustains between 10.1% and 12.6% by my determination. Long Bay Road also has recovery sections, but they are much shorter: 500 meters and 200 meters, approximately, most of the rest of the climb hovering about the 12-13% range.

So which is harder? A short grunt of a climb (Haylocks) flirting continuously with the nasty side of 20%, or a much longer climb (Long Bay) which tends to hang out in the 12-13% range? The answer depends, of course, on what gears relative to your strength-to-weight ratio.

My climb rating formula includes an exponential difficulty factor proportional to the exponent of the ratio of the grade to 8%: when the grade gets much over 8% difficulty per unit altitude change increases rapidly. With this assumption, the difference between a given amount of climbing at 12% and at 20%, sustained, is around a factor of e, or around 271%. This pays off big-time for Haylocks.

So here's a plot of the derived ratings, where I applied a 100 meter dual-sided exponential smoothing function to the Haylocks profile to get rid of those anomalous wiggles before applying the formula:

Haylocks, despite being the least total climbing of any of the climbs I plotted, wins on the difficulty scale. Hmm... I still disagree, from my perception, but with different perceptions (for example, had I been stuck with a 39-23 low gear instead of a 34-26) it might have been an undisputable conclusion.

Here's a summary of the numbers, all of which have been normalized to Old La Honda Road (5.55 km, 393 meters, rating formula = 1005.5 meters):

climb	km	m	rating
Pigeon Bay Rd, Banks Peninsula	0.603	0.736	0.908
Dyers Pass Rd (S), Governors Bay	0.529	0.731	1.012
Crown Range Pass (E), South Island	7.048	1.966	1.057
Okains Bay Rd, Banks Peninsula	0.835	0.952	1.070
Arthurs Pass, Westland, South Island	0.718	0.998	1.602
Crown Range Pass (W), South Island	2.019	1.753	1.716
Long Bay Rd, Akaroa	0.893	1.120	1.750
Spur Rd - Haylocks Rd, Onuku	0.207	0.524	2.306

climbing Onuku hills

2012-01-25T17:00:00.000-08:00

With climbs, there's a tendency to focus on those which gain big altitude and are frequently ridden, preferably in top-level races. However, sometimes it's the back-road little-known compact climbs are equally memorable, or in this case, surprisingly challenging.

The day after my big Banks Peninsula hill ride, I went for a short scenic ride with Cara south of Akaroa. The goal was to check out Lighthouse Road. From the distance Lighthouse road is clearly visible, etched in the side of the treeless hill, the occasional car snaking its way slowly up the road's numerous switchbacks. It looked steep. But before our trip I'd set Cara's bike up with a 34/32 SRAM 1070 cassette with a medium cage SRAM Rival rear derailleur, a very nice combination which seemed to work very well (I detected no reduction in shift quality relative to her previous SRAM Red short-cage). So I figured we'd go check it out and see how steep it really was. Looks can be deceiving.

In central Akaroa, she zipped away in her frequent fashion, thinking I'd meant the road to the present location of the lighthouse. But the lighthouse was moved years ago when the actual heavy lifting of keeping boats off the rocks was taken over by an automated beacon: the old residential lighthouse was now just another tourist attraction. Lighthouse Road went to the old site, where the beacon now lives.

The coastal route Cara led us on didn't pan out as expected, however, as first it turned to dirt, then ended in overgrown brush at an exit point for a treated sewage line. Not a tourist hot-spot, I decided, although I hate to discriminate against sewage pipes. I was reminded of the proposal in San Francisco to rename the sewage treatment plant after George "W" Bush, but it was decided that was unfair to the workers there. In any case, we backtracked and continued the way I'd intended, towards Lighthouse Road.

The Banks Peninsula is volcanic, and like Maui, another volcanic land mass we'd visited (in October), the land rises up from the sea without much pause: if you go inland, you go up, at least until you start going down towards the opposite shore. So moving inland to intersect Lighthouse resulted in some non-trivial climbing on Onuku Road at a fairly stiff grade.

But Lighthouse Road was in a different league. I turned briefly left onto the road and immediately recognized this wasn't the sort of thing I'd planned to take Cara on, especially since she was recovering from toe pain at the time.

So instead we continued down Onuku Road which leads to the town of the same name. Onuku is a Maori village whose primary tourist attraction is the 19th century church there: a melding of Christian and traditional Maori iconography. Prominent is a big face with it's eyes bulging and tongue extended, a trademark Maori expression.

The church was in fact very nice, right near the shore. As we were there, a tour guide riding a motorbike with a client in an attached sidecar came by, paused, U-turned, and returned to Akaroa. But we weren't ready to turn back.

The road cut inland from the church, tilting upward before turning a corner. "Uh, oh," I said, "this looks steep!"

"We'll see," Cara replied, and up we went.

Around the corner it kept climbing... and climbing... and climbing. I wondered how far it would go. Once a road starts going up, it's hard to turn back; the top might be just around that next corner!

Eventually I got to a junction, the road continuing steeply in both directions. One direction, labeled Hamiltons Road, appeared to be a more major road than Haylocks Road, the other direction. I paused for a bit, not wanting Cara to get confused about the direction I'd gone, but then I decided she'd probably recognize Hamiltons was the "obvious" route, so that's how I went.

the junction

Wow -- now this was really steep. Cara, I realized, probably wouldn't climb this far, so I wished that the road would end, even if I wasn't ready to turn around voluntarily. My wishes were met, however, when I reached a gate marking the entrance to Onuku Farm Hostel. A car which had recently passed me was there, the passenger opening the gate, but I didn't try to slip through. I declared this the top of the public road, and turned back.

gate at top of Hamiltons

When I reached the junction with Haylocks, I was surprised to see Cara waiting there. Not only had she climbed to that point, but she'd gone further: up Haylocks a bit, before turning back. I looked once again up Haylocks, only now appreciating how pathologically steep it was. Just by climbing a part of that, even with her new climbing gears, Cara proved I'd underestimated her. I told her I'd check it out myself (not wanting to be out-done) and would be back soon.

Haylocks was too steep for me to clip in directly, so I did a little loop at the junction to get into my Speedplays. Then I started to climb.

I was in my 34-26 (34-23 being simply out of the question), but at times I wondered if I'd be able to keep that gear turning. This road was really, really steep. Not Filbert Road in San Francisco levels of steep (31.5%), that road I could only handle for a short distance, but this road was close to my limits for sustained climbing.

After a bit, I noticed a house in the distance, well up the hill, a road leading to it. Roads not being in huge supply around here, the reasonable conclusion was the one I was on led to the house. I steeled myself for the life of pain to which I would be condemned for the foreseeable future... But then, salvation arrived. The road crossed a rough cattle guard at a gate.

gate marking the end of my Haylocks climb

I briefly considered "sprinting" to a speed which would give me a chance to get over the wide gaps in the guard, as the gate was open, but then I noticed a sign proclaiming "Lambing: keep out" and marking it as land for a Bed and Breakfast. I decided to follow orders, and returned to the junction where Cara was waiting. This descent gave me another hint about the grade: it couldn't be much over 20%, because at around 25% such as I encounter on several blocks near home, I have problems controlling my speed, while here I was in no danger. After I joined Cara we then rode back to our holiday park, a journey which took us over three significant steep climbs, but nothing more of the level of Haylocks.

The ride ended with an epic fail as I went hard at the base of Old Coach Road, intending to set a good Strava time to the holiday park on Morgan, but I simply had nothing in the tank, my motivation snapped mid-pedal-stroke, and I came to a virtual stop seconds later before recovering enough to crawl the rest of the way up the steep hill.

For the ride I was using only the Strava Android app on my phone. My personal Garmin Edge 500 was lost before the trip (stolen? misplaced? Maybe I'll never know), and so I had a Garmin only when Cara wasn't using here. The app records only longitude and latitude, with Strava using map data for determining altitude. This works surprisingly well, but not as well as the altimeter built into the Garmin Edge 500, so there tend to be fluctuations in altitude about the true value.

Here's that result, where I show the profile for riding from the Maori church to the junction and onward up Haylocks to the gate. I constructed this from my climb first to the junction, and later my climb from the junction up Haylocks:

profile of Haylocks from Maori church. Ignore wiggles, a mapping artifact; climb is continuously up. Hamiltons profile had greater artifacts and is not shown.

The line shows a 20% grade, and Haylocks has no problem keeping that pace. It would have been interesting to see how it continued on to the bed & breakfast. I suspect it was more of the same.

As for Cara, I'm super-impressed she was able to ride up Haylocks at all, especially after riding the climb to the junction. She's getting much stronger in her long return to her fitness levels she had before her mountain bike crash in 2009.

Later in the day I considered returning and exploring Lighthouse myself. However, I was simply too tired from all the recent activity. I needed some rest.

Climbing Dyers Pass Road

2012-01-24T00:46:00.000-08:00

Christchurch was the beginning and end of my just-completed trip to New Zealand. Christchurch consists of a dense central district, a suburban ring, and beyond that, agricultural land dramatic mountains. The central district has been largely destroyed in two days of quakes: 4 Sept 2010, and a highly destructive aftershock on 22 Feb 2011. The damage to the suburbs is less visible but just as real: many houses were condemned due to structural damage.

But the riding outside the city limits remains truly excellent. There's three popular passes just south of the city: Gabbies Pass, Dyers Pass, and Evans Pass. My route book, "Classic New Zealand Road Rides" by Jonathan Kennett and Kieran Turner described a ride over Gabbies Pass and Evans Pass, and shorter ride over Dyer Pass and Evans Pass. Each of these rides went over Evans Pass, but I was told as I staged at Zero's Cafe near Princess Margaret Hospital in Cashmere (part of Christchurch) that Evans was closed. So my alternate plan was to ride Gabbies to Dyer.

But two separate riders independently suggested an alternative: at the top of Gabbies Pass, ride "The Bastard", a climb along the local Summit Road (not to be confused with the one near Akaroa) to the top of Dyers Pass. Then I could descend Dyers back to Christchurch.

This presented a bit of a problem: I didn't want to deny myself some classic local climbing. But as I was riding towards Gabbies, a solution occurred to me: ride The Bastard, but then descend and reclimb Dyers (towards Governors Bay, the good side), then descend to Christchurch.

I asked someone at a cafe along the route if this seemed reasonable... there's a lot of cafes in this cycling-crazy region. "I'd doubt your sanity" he replied. That sealed it.

My derailleur cable slipped when I was riding Gabbies Pass, ruining that effort, and as I climbed the Bastard, the climbing itself took a back seat to admiring the spectacular views. To one side, Lyttelton Harbour, and to the other, the patchwork of extensive agricultural fields surrounding Christchurch. It was great.

But after climbing to the peak of Summit, then riding the rolling descent beyond, I came to the "Sign of the Kiwi", an historic cafe at Dyers Pass. One rider after another ascended from the right, the Governors Bay side. Others at the top cheered on those approaching the summit. It appeared to be some sort of large group ride, if not a formal event. I wanted my turn!

So without further delay, I set off down the descent, recovered a bit at the bottom, and began my climb.

Other than the car traffic, I loved climbing Dyers Pass.

On the down-side, it's a two lane road without much shoulder, blind corners resulting in occasionally uncomfortably short sight lines for drivers to see cyclists. But it has the advantage of popularity: with so many cyclists on the road, drivers seem to be generally aware, and I had no issues during my ride there.

But the climb itself is a straight uphill grind. Gaining close to 300 meters to its 331 meter peak, it's just a put-it-in-a-single-gear-and-go sort of climb, 10.2% average according to Strava. At that distance, you can really go out at a good, hard pace without fear of digging yourself too deeply. Sure, it will start to hurt, but then you just ride through the pain for the few extra minutes it takes to reach the top. The finish is fun: it lets up for a short distance, then makes up the deficit in the final meters. Doing a 50 meter smooth on altitude versus position, however, results in only a 1.2% increase in grade here about the mean, so it's easy to power through this in the same gear. If you have anything left, however, this is a good chance to upshift, build up some speed, and finish hard. In the big picture, though, Dyers is an easy climb to make an effort on, since there's not much thinking involved. If I lived in Christchurch I'd love doing this climb.

Given how unfresh I was from my recent hard rides, and how much extra mass I had on my already relatively heavy Ritchey Breakaway, I felt really good with my effort. It was good enough for the Strava KOM, with what for me is a solid VAM number.

I contemplated also climbing the Christchurch side. However, as I descended, it was obvious this was less interesting: the grade was generally shallower (a bit steep only near the bottom) and the influence of the shifting winds would be greater. So I decided enough was enough for the day, and so headed back to Zeroes to meet Cara, who was finishing up her flatter ride.

Climbing Pigeons Bay and Okains Bay Roads

2012-01-23T19:09:00.000-08:00

After climbing Long Bay Road on that second ride, I continued along Skyline Road, west towards Hill Top.

Summit Road was absolutely gorgeous. Never flat, it's alternately up and down the whole way as it offers fantastic views of both sides of the ridge. It's reminiscent of Skyline Road in San Mateo and Santa Clara counties of California, Summit Road and Highland Road down near Loma Prieta, and Ridgecrest Boulevard on Mount Tamalpais, or El Camino Ciel above Santa Barbara. Each is gorgeous, each similar yet unique. The pavement on Summit on Banks Peninsula was quality chip-seal, no issues at all, and traffic was light, although a high fraction of the traffic was trucks. It's a rural place, with ranches on either side of the road, ranches taking up much of New Zealand land.

So I passed first the road to Le Bons, then Okains Bay Road, then Pigeon Road, in addition to several "unsealed" roads along the way. But then I emerged finally at Hill Top, the terminus of the "tourist" route from Akaroa to Christchurch.

At Hill Top I descended 75 to Barry's Bay at the water's edge. The descent was fantastic: certainly it could have been taken a lot faster than I did. The roadway is excellent and there's no surprises in the corners, with turns presenting even the slightest challenge well-marked with recommended speed limits. The standard speed limit on even two-lane roads in New Zealand seems to be 100 kph, so any turn which can't be taken this fast is marked thus.

Traffic on 75 can be a bit heavy. It's especially bad if you catch the tour buses. On the descent, however, this isn't a bit issue. The road was certainly empty on my first descent, relatively early in the day for traffic originating in Christchurch.

Approaching Devauchelle, I reached my next target: Pigeon Bay Road. Here's the profile. Pigeon Bay has some challenging sections, but since it gains only 290 meters, it passes relatively quickly. But it's actually a bit more than this if you consider all of the roads between 75 and Summit begin at peaks in the undulating Highway 75, and (except for Long Bay) intersect Summit at dips. So there's some climbing to be done before reaching the base of Pigeon Bay.

At the top, I turned left on Summit, returned to Hill Top, then descended 75 again. This time there was substantially more wind and more car traffic. People were beginning to arrive from Christchurch. I even had a vehicle tailgate me a bit to let me know it wanted to pass. It was really the closest thing I'd had to that point to an "incident" in New Zealand. But really it's nothing anti-cyclist; were I in a car at the same speed he'd have done the same thing.

This time after passing Barry's Bay I continued through Devauchelle, climbing again to the next climb in the series, Okains Bay Road. There are two other prominent roads inland between Pigeon Bay and Okains Bay, but they aren't paved routes to Summit.

I was getting tired by this point and Okains Bay Road wasn't a very good effort for me. It's longer than Pigeon Bay, gaining around 371 meters. Indeed it seems to go on for quite awhile, Summit always visible above, yet since Summit is descending towards its intersection with Okains Bay, it's never clear how much climbing remains. To make things worse, or perhaps more interesting, the grade increases in segments, so if one isn't aware of this it's easy to think fatigue is winning the battle. In my case, perhaps it was.

But the intersection is only part of the way. There's still over 200 meters of climbing between Okains Bay Road and Long Bay. Tired and hungry, I stopped for a sandwich along this segment, admiring the view while I ate.

Finally I descended Long Bay back to Akaroa. This is the finale of "Le Race", the annual bike race between Christchurch and Akaroa. I can't imagine ending a race on such a steep descent. I'm told there's crashes every year.

Climbing Long Bay Road

2012-01-22T16:31:00.000-08:00

Akaroa on the Banks Peninsula, southeast of Christchurch on South Island, New Zealand, provides for excellent cycling: probably the best encountered on Cara on our vacation.

First up on the list is Long Bay Road, which goes from the northern boundary of Akaroa up to Summit Road. Summit runs along the rim of a volcano, Akaroa Harbour forming the cone, and Long Bay (among others) on the outer edge of the rim. There's a series of roads which run from the harbor (the cone) over the rim. From east to west the paved ones are Long Bay Road, Okains Bay Road, and Pigeon Bay Road. I rode all three while in Akaroa last week, but Long Bay received most of my attention.

Long Bay is the longest and toughest. Upon arriving in Akaroa, having done a fantastic 10 km trail run at Lake Takape that morning, I went for a "preview" ride of the hill. I'd not planned on doing the whole climb, but with Summit Road dangling in the distance, the hills mostly stripped of sight-blocking trees by ranchers in the 19th century, it was virtually impossible to stop climbing. Actually the climb extends a bit, "unsealed" gravel, beyond summit to a place called Cloud Farm. The name makes sense when you realize it's open space land purchased by the Queen Elizabeth II National Trust to preserve it from development. Here's that preview ride:

The next day I was feeling okay, so I set off to do a stronger effort up Long Bay to begin a longer ride.

There are two primary routes of access to Long Bay Road from central Akaroa. The "obvious" route is Highway 75, the main highway between Akaroa and Christchruch. There can be a relatively large traffic on this road, so it's not my recommended route unless you are gunning for a Strava segment which includes it, like this one, or this one which I created based on my "preview" ride.

A better option is to instead climb Old Coach Road. This road forms a semi-circle, starting and ending at Highway 75, the upper intersection adjacent to the start of Long Bay Road. Along the way Old Coach intersects with the bottom of Morgan Road, which is paved for the short segment to the Top 10 Holiday Park, but afterwards becomes rough gravel until it reaches Long Bay Road part-way up. I wanted to investigate this road but never made time for it: the gravel at the top seemed fairly rough; it was perhaps better at the holiday park entrance.

From the base, Long Bay Road begins climbing moderately steeply: steep by most standards, but not to the typical New Zealand standard. It's a fairly steady 8.2% for 1.0 km. After this the road levels out for another 500 meters. But then the real fun begins. The road heads upwards...

The next 2.6 km is a series of steep pitches and brief reductions in grade. I did a regression on the data from a 2 km segment, yielding a 12.2% grade, but the actual grade fluctuates about this value. My first ride up I tried to hold onto my 34/23 here, but in the second ride I more quickly retreated into my 34/26. Perhaps not surprisingly, the second ride was faster.

Finally the road once again relents. My first time, with a tailwind, I even thought there was a descent here, but on my second ride this perception was gone. The altitude data shows no descent, but compared to what you've just been through, it may as well be. Looking ahead, however, reveals that there's still more to come: the road is clearly etched into the hillside above the present altitude, even if the climbing directly to follow is hidden around the corner. Look for the tower, however: that's the goal, the tower past Skyline.

And once around the corner, up the road goes again. I have this section at 12.9%, but it's basically the same as the upper portion of the previous segment. The altitude difference just mentione is closed quickly: only 400 meters of suffering and the intersection suddenly appears. Long Bay is done.

But wait; there's more! Cross Summit Road and an easily ridden gravel road continues to climb, somewhat steeply but without the pressure of timing yourself here, not bad. At the top are the best views of Akaroa below, and the Cloud Farm Open Space.

This was the end of my preview ride. The ride the following day, I skipped the gravel and instead headed west on Skyline for more adventure.

climbing Crown Range pt 2 : Queenstown side

2012-01-19T17:00:00.000-08:00

Having climbed the east side of the pass, it was time to descent to the west. Every new turn brought a fresh view of the scene laid out before me, and it was a challenge to keep my focus on the road. After a wonderful series of sweeping turns, one punctuated by an active road construction crew, I entered the final portion of this descent, a series of tight hairpins reminiscent of my descent of the Poggio on the Ligurian Coast last fall. I knew I was riding these too slowly, but taking the turns a bit easier had the benefit of extending their duration, and they were worth savoring.

The descent done, I was ejected onto Highway 6. I expected this would take me to Arrowtown just down the road: a few km at most, based on my reading of the route profile before I'd left.

But I'd misunderstood the route profile. What had been so close to the Crown Range was Arrowtown Junction, not Arrowtown. It was still 5 km to the town. So down the side road I went.

When I arrived at the town, it started to rain, lightly but steadily. I first went into a Holiday Park (RV Park) for directions, which were trivial, then continues into the quaint, Los Gatos-like downtown where I rolled along, counterflow on the one-way road, until I found a bakery. Looking past all the meat pies I noted they sold French bread. I got a wonderful baguette, which would serve as my calorie source for the return trip.

It was still raining as I returned to Highway 6, but cleared by the time I'd returned to Crown Range Road. And so I began the highlight of this ride: the ride of its western slopes.

click on image for expanded view

It was a hard climb, longer than I'd remembered from the descent. Finally I made a push for what I thought was the summit, but it was not... several more steep segments remained before finally I reached the lookout at the pass. It was never super-steep, but often uncomfortably so, and I spent plenty of time in my 34/23 even if I resisted the 34/26. But the summit finally arrived, as it usually does.

A curious characteristice of this climb are the flat portions, which take up close to 3 km of the 10.8 km total distance. These really hurt VAM numbers for the climb: far better would be a steady grade of the same distance and total altitude gained. With the exception of a 200 meter break, what I had initially thought to be the finish, the final 3.0 km are sustained at 10% or more.

This would make a fantastic Low-Key Hillclimb. My rating algorithm scores it at 172% Old La Hondas of difficulty. This puts it between Montebello Road and Sierra Road in the climbs I've rated from the Low-Kehy Hillclimbs.

As I'd approached the top, I'd noticed a dirt road snaking up past the summit lot, so when I reached the finish of the paved climb, I went to check it out. It was fenced off, but a stair provided pedestrian access past the fence. After a bit of recovery, I carried my bike over the stair, and walked it up the first portion of dirt: the surface was too rought and steep to clip in. At the first switchback, it was evident the road only got worse further on, so I admired the view there and returned the way I'd come.

I'd hoped for tailwind on the return to Wanaka, but to the contrary, I battled a headwind the whole way. When I finally arrived, my baguette long since eaten, battered and a bit broken, I pulled into the Racer's Edge bike shop to fill my tires to their target 80 psi (they were down around 40... so much for getting an easy 100 psi from my hand pump). I asked the mechanic whether it was always this windy in Wanaka.

He looked outside, paused. "It's not windy today," he finally responded.

That basically answered my question.

Even the winds, and getting lost heading back to the holiday park, couldn't dampen the euphoria of the Crown Range. If you come to New Zealand, this climb is a must.

climbing Crown Range pt 1 : Wanaka side

2012-01-18T20:19:00.000-08:00

During a particularly good SF2G, I head the rider called "Space" say he'd been to New Zealand. I asked him what the best ride he'd done there. "Wanaka to Queenstown", he said. At that moment this became a "must-do" for my trip to the country.

There's two routes connecting the two cities. The main route, Highway 6, goes around the mountain while the more direct route, Crown Range, goes over it. Obviously Space has been referring to the latter.

Crown Range was only paved relatively recently. Indeed, the profile in Scott Kennedy's book marks it as dirt. The text description, on the other hand, makes no mention of a dirt surface (although gravel on shoulders is described in the summary directions). Obviously Lonely Planet had failed to update the profile in the edition where the text was updated with the new road work. The book has many issues.

Strava verified Space had ridden the Crown Range. So when Cara and I arrived in Wanaka, after a fun stop to run the maze at Puzzle Town, the ride was on my short list of to-do's.

We were staying at the Outlet Holiday Park on the shore of the lake. It was a wonderful place, within walking distance of the Sticky Forest, home to a virtual amusement park of bike-legal single track. Not only were the trails legal, but they were labeled with such evocative names as "Super G" and "Woo Hoo", marked in signs with a Bike logo. While I didn't have a mountain bike, the trails made for a most excellent trail run.

click on image for expanded view

From the park it was a bit of a ride to get to the Crown Range Road. But once there the route was simple: ride west on the road over the pass, turn right at the "T" with Highway 6, go to Arrowtown, stop for food, return.

The route began nicely enough but unspectacularly. After an initial little rise, the road began its initial gradual ascent. I stopped to take a photo of a sheep, one of many this trip. It was to be my last photo of what was to be a day of spectacular scenery: my camera battery emptied with this shot.

The first goal was Cardrona, approximately 25 km from Wanaka, 22 km from the start of the "climb". After passing through a construction zone and the entrance to the popular winter ski resort, I was there. As far as I could tell, there was no active mid-summer commerce here: a historical general store/post office, and an old hotel. I did see a woman in what appeared to be a kitchen of the hotel, so that was probably open. There was an antique car parked on the street.

Onward... I knew from Kennedy that the road eventually kicked up from here, increasing in grade. Indeed as soon as I left town I began climbing in earnest. This section is around 5.9%, but doesn't last long. Soon after the road levels out again, beginning a series of crossings of the Cadrona river: ten, I think.

I crossed a cement cattle guard. I noted I'd not before seen a cement one: they're almost always metal. Soon after this musing, I noticed something else: my front tire was most definitely going flat. So much for any Strava heroics on this side of the hill... at least starting in Cardrona.

Before the trip, based on advice from my SF2G brethren relayed by me, Cara had purchased two Lezyne "Road Drive" mini-pumps. 100 psi without problem, I'd been told. When we arrived here, our tires deflated by Security, I'd inflated them with around 150 pumps of the little pumps. This was the pressure where pumping became difficult. I'd not really assessed how much pressure this was, but when after this ride I stopped at Outside Sports in Wanaka to inflate my tires fully, I realized they'd probably been only around 40 psi. No wonder I pinch-flatted.

I repaired the flat, perhaps unwisely using a patch instead of swapping tubes. Another Lezyne purchase before the trip: their expensive but apparently high quality "glueless" patches. I put one of these on a puncture hole, reinstalled, and reinflated. No luck. So I removed the patch, which seemed to still be intact, applid some patch kit glue, and moved the patch to cover both holes. This worked: the patch held, and the tube held air when I reinflated it. 17 minutes for the stop... but it seemed like less, as it really was a nice place.

My incident marked the beginning of the true climb From here the grade began to increase, never decreasing until the summit. The final 2 km was surprisingly steep: sustaining around 11.4% before finally relenting in the last hundred meters.

And when I reached that summit, what i saw was absolutely spectacular.

I can't really describe the view from Crown Range Pass. It was so panoramic, so dense in detail, it was just too much to really all absorb: the road winding down the opposite slope to an expansive valley, mountains in multiple layers of distance, dirt roads winding up the closest slopes. It was an overwhelming excess of color and shade: various hues of green marking the valley and hillslides, extending to white snow on the distant peaks, a light patching of clouds on the otherwise blue sky. Really incredible.

And my camera battery was empty.

Enough, I decided: I'd told Cara I'd be back by 2 pm. So I got back on my bike and continued down the western slope.

Much later, when I'd run my rating formula on the numbers from Crown Range (E), I got a rating for this climb of 106% Old La Honda, with a distance of 705% OLH and an altitude gain of 197% OLH. So it gains more altitude but does it so gradually, at least initially, the net "difficulty" comes out about the same.

Hass Pass returned from the dead

2012-01-17T11:05:00.000-08:00

Weird...

After not seeing the Haas Pass data on several log-ins, it appeared on my account... not sure where it was hiding, but glad to see it back!

Actually, I think there's an issue with Synchronization when running in "airplane mode" when running the Strava app on a phone. I'm running airplane mode because data and voice are too expensive overseas, and I want to avoid any risks apps will check for updates or new mail or whatever and incur $1.99/minute data charges. And when I'm on vacation, spending time dealing with the internet becomes a low priority: it's liberating to do things like read actual newspapers, play Scrabble, or just sit and enjoy the chirping of the birds in the morning.

So Haas is back. A few other activities are now mysteriously missing, but I have confidence eventually my phone will "sync" them with my account. Patience, patience....

On my trip: not riding much, but every day is something: runs, hikes, even a fun day of riding a rental downhill bike in Queenstown. Cara, on her first mountain bike ride in 33 months since badly injuring her knee in a mountain bike crash, was dusting me on the beginner course. I salvaged some self-esteem by doing the intermediate course as my last run, after she'd declared herself tired and ready to stop... good time, but I will never be good at throwing myself at the mercy of gravitational acceleration.

More pass reports to follow... presently I'm heading back towards Akaroa near Christchurch where I'll get to sample some of the best road climbing on the island. The island has huge potential for epic climbs, but a lot of land is private ranch land with crude dirt roads which are accessible to "trampers" only. In fact, this is a hiker's paradise, with wonderful multi-day routes with a regular spacing of comfortable huts. Many of these routes cross private land, but people here aren't as fanatical about the delineation between private and public property as self-absorbed Americans.

Anyway, back to life... time to log off and do a quick trail run from Lake Takepo.

climbing Haas Pass

2012-01-15T10:00:00.001-08:00

Next on my list of New Zealand climbs was Haas Pass. In a way, this was a ride destined to end in tears. But more on that at the end.

Cycling New Zealand, a guide written by Scott Kennedy for Lonely Planet, describes the route from Makorora to Haast Junction, crossing the pass. However, if you look carefully at profile, shown on page 268 of the Sept 2009 edition, it's clear that only three data points were used for the profile between Haas Pass and Haas junction, covering 82 km. There's a point at the junction, of course, then one at 11 km from the pass at 100 meter elevation, then one at the pass itself, at what appears to be 560 meter elevation. So that's 460 meters gained in 11 km, an average of 4.2%. No problem, one might decide: this calls for an 11-20 straight block.

Heh.

I rode with Cara from our D.O.C. (pronounced "doc") campsite at Boundary Creek. Riding was effortless, it seemed, a good sign that there was going to be a heavy price to be paid later, with interest. it was around 15 km to the Makorora, then another 8 km to the Blue Pools, our goal. We locked our bikes, switched to our walking shoes, then did the wonderful little hike to the pools, which are surprisingly, well, blue for their depth. Perhaps the lack of particulate matter in the water results in reduced optical attenuation, increasing the vibrancy of the color. In addition to the view, the pedestrian suspension bridges are very nicely built. New Zealand takes its tramping seriously.

Back to our bikes, Cara set off back to the camp, while I went on to ride the pass. Of course high on my mind was Strava: when in foreign lands, K.O.M. opportunities abound.

The southern side is relatively easy, since according to Lonely Planet the real climbing here is only around 180 meters over 3 km. That seemed about accurate. It's a nice effort, but nothing epic.

The descent was surprisingly quick. Not only is it much steeper than the 4.2% indicated by Kennedy's profile, but the tailwind reduced the relative headwind, increasing my rate of acceleration on the straights. And occasionally I'd feel the instability of the shifting winds as I went around corners. As is often my practice, I took the descent carefully.

At the bottom, I wanted to make sure I'd cleared the Strava segment, so continues onward effortlessly.

I turned a corner to a spectacular view of distant mountains. I stopped and took a photo. A look at my phone showed the distance so far: 44.5 km. This would be a good time to turn around. So after a few pictures, I turned.

Instantly it was made clear to me why the riding on the flats had been so effortless: I was going straight into a brisk headwind. Ah, well...

The climb began right past the lookout for McFarlane Falls. And it was instantly steep: the grade meter was definitely pegged above 10%. At least the wind was less here due to the mountain's shelter.

And what a climb! Views of the gorge on the left, waterfalls cascading down the sheer cliff on the right, the occasional cattle guard to keep things interesting, all cooperated in keeping my mind off the effort. The road snaked along, rising above the Haas river, until it left the river for the final climb to the pass.

I kept the effort up here, increasing it when I saw the opposite-facing sign welcoming travelers to Westmore, not letting up until I cleared the simple pass sign on the left, the older, stone-etched county line divider on the right. I felt this was a Strava-worthy effort.

After the rapid descent came the long run back to Boundary Creek. Near Cameron Flat I looked again: 59 km covered, so 30 km remaining based on my turn-around. It was the last time I would look at the distance reading on the unit, but I knew at least I had both crossings of the pass in the bag. No other significant climbs remained.

So 30 km left, but what a 30 km... winds were occasionally tempered by lines of trees which marked the boundaries between sheep-grazing fields, but otherwise, it felt as if every pedal stroke was being heavily taxed: 2/3 for the wind, 1/3 for forward progress. A part of my issue was likely food: I hadn't brought enough.

I finally reached Makorora and it's surprisingly expansive Cafe/General Store/Information Center, which I entered after parking my bike outside. I resisted getting a $5 smoothy (well, fruit-flavored milkshake, but I assumed a smoothy could have been procured) and took the $1.50 banana instead: it was only 15 km further. I stopped to check on the Siberian Experience, a combination plane ride - hike - jet boat adventure. $320/person.... I decided really the hike was the part of it I wanted, and that can be had gratis. So off I went.

The final kilometers opassed, as they will, although much slower than wanted due to the wind. I'd hoped for a banana-induced kick, but none was granted. But turn over the pedals enough time and a goal is reached, and here I finally saw the 4 km to go sign, then eventually the campground itself. I decided I've gotten headwind-soft living in San Francisco after my time in Austin, Texas. I let the headwinds beat me down. Cara, by the way, later told me she hadn't minded the headwinds at all.

Here's where things went to tears: there was a problem with Strava and, although my app had been showing full distance, only the first 18 km of my ride had been recorded. These had been ridden at a relaxed pace with Cara, of no Strava interest, so I didn't bother saving them. My climb efforts were lost in the void of unrecorded data, the place we call /dev/null.

I was stunned. Shocked. A small bit of my life drained away in a spasm of despair. Surely I could have avoided this... had I not lost my Edge 500 before the trip... had I replaced it before the trip... The Strava app is great, but there's no beating specialized hardware for reliability.

Ah, well. There would likely be no further climbs of Haas in this life, but at least I had the memory of what was a surprisingly challenging climb, especially given Lonely Planet's neglect. Indeed, this is just one of many examples where the book fails to do justice to New Zealands fantastic cycling roads. I hope a better cycling guide for New Zealand is published soon. John Summerson, are you listening?

Postscript: Later, in Queenstown, I found a much improved guide to New Zealand cycling: the domestic "Classic New Zealand Road Rides" by Jonathan Kennett and Kieran Turner. It's a bit pricey at NZ 39.90, but the route profiles and descriptions are clearly superior.

Climbing Arthur's Pass

2012-01-09T22:13:00.001-08:00

New Years Day meant two big events for me: riding the San Bruno Hill Climb for the first time in four years, and going to SFO to fly to New Zealand.

After a night in Christchurch, viewing the sobering post-apocalypse which was the central business district and the contrasting beauty of the botanical gardins, Cara and I got in our rented camper van and headed out on national highway 73. Our destination: Jackston's Retreat, a very nice holiday park in Jacksons, west of Arthur's Pass through the New Zealand Alps. I'd read about Arthur's Pass using Strava Explorer. It's part of the Five Passes Tour, a combination bike tour + state race where riders tour the southern island and are timed over 5 climbs along the way. Arthur's Pass is rated a 5: the maximum points of the five.

The profile was impressive. The net stats are fairly close to Kings Mountain Road: 6.87 km and 490 meters of total climbing, no intermediate descents: average grade = 7.16%. Kings mountain is 6.95 km, gaining 469 meters, a 6.77% average. The key is that Arthur's Pass starts out quite gradually: 2 km in and you've climbed only 73 meters, a 3.7% average. The rest of the way is an 8.60% average, with most of the climbing steeper than that.

I set out from Jacksons onto a gradual uphill, but the upward grade was largely canceled by the typical western wind. The southern island of New Zealand is in the "Roaring 40's", the 40 to 50 range of southern latitudes well-known for its strong westerlies. On this side of the pass, there was little to abate the winds coming in from the ocean, providing a nice tail wind for eastern riding.

I was still feeling sluggish from the 16 hours plane journey and four hour time zone difference from San Francisco (four time zones plus summer time in New Zealand), so wasn't in a hurry for the climg to arrive. But soon enough I reached Otago, the town marking the base of the real climbing. I'd noted from Google satellite view via Strava that the Strava segment started somewhat east of a bridge just outside the town.

I was a bit worried about the weather on the pass. Low-hanging clouds obscured some of the tall peaks nearby, but the pass ahead seemed clear. It looked like a good day to climb it.

Not long after passing through the town I arrived at the bridge. Many bridges on the New Zealand south island are one-lane, two way, with one direction having the right of way. I would later encounter a bridge that was one-lane, 4 way: two vehicle directions, and remarkably, two rail directions all sharing the same space. This one wasn't quite that dramatic, but for a one-lane bridge it was fairly long. So I looked carefully to make sure no motor vehicles were approaching, then crossed the bridge. It was time to start the serious riding!

At the other side of the bridge I deposted one of my two water bottles at the side of the road, avoiding having to carry that weight up the hill. I was fairly heavily loaded, with my big saddle bag, a sandwhich in my pocket, my cell phone, and Cara's Garmin (mine I managed to lose somehow...). Still, every bit helps.

Then I started. This was clearly "climb", so I ramped the effort up to something appropriate. My gear started to skip almost immediately. Damn... I experimented with the shifter to find a gear I could pedal continuously.

Normally the thing to do here would be to adjust the barrel adjusters on the down-tube. But because of the high placement of the cable housing braze-on, there was no room for my barrel adjusters, so I omitted them: cable tension is adjusted either at the rear derailleur, or with a bit of risk, at the cable splitter. Neither of these is accessible when riding. I could have descended and started again, but since I'd told Cara I'd be back around 11 pm, I didn't want any extra delay. I was committed.

I was able to guess the cause: when I'd swapped the front housing for my rear derailleur cable a few days before leaving, the cap for the hold housing had been mangled, and I'd installed the housing without a cap, too rushed to go to a shop to get a new cap, and with inadequate support from the slotted braze-on, the housing separates, allowing cable tension to relax. It was a stupid mistake, one I'd made before.

But I was stuck doing the best I could. Not long after, however, the road steepend... dramatically. I shifted into my 34-26, a gear where the derailleur position is set by the limit screw rather than the cable tension, and the bike operated fine again.

The road here is a real grind: 15.5% for around 800 meters, with fluctuations around that mean. But I love this sort of thing: on a new climb on the opposite side of the world, absolutely gorgeous views everywhere. I knew the discomfort wouldn't last long, but I'd always remember this climb.

I turned a corner and hit a dramatic view, from the opposite direction, I'd seen from the camper van the day before: an open tunnel, similar to what one would expect in the French Pyrennees or Alpes, with a small aquaduct carrying water clear of the road, creating a waterfall under which I rode. I forgot instantly about the discomfort: this was great.

Then a landmark of this road: the Otira Viaduct. The viaduct was biult in the early 1990's, a considerable piece of engineering, to replace the roadway along the "scree slope" which was prone to periodic collapse. Riding the viaduct is really special, even if there's a bit of concern about the huge drop if one were to get bumped over the edge by a passing truck or car. But I had the viaduct to myself.

After passing the viaduct, the road leveled out a bit, passing a viewpoint where Cara and I had had a Kia perch itself upon our van of the same name, apparently enjoying tastes of our roof. But I wasn't stopping now... Here again I was forced to fumble with my shifting, trying the big ring to see if that would help, holding the lever partially depressed to put the derailleur in a position where it wouldn't shift.

Another steep slope followed. This was it, the last section, maxing out at around 18%. But it's short, only 60 vertical meters.

Then the road flattens for a sprint before a final rise to the plaque marking the summit. I initially missed this, riding onward, until the road had clearly begun to descend, and I was at a memorial obilisque for Arthur himself. I then turned around and began my ride back.

I stopped to climb a steep road to a viewpoint of the viaduct, a detour Cara and I had missed the day before. It was quite a view. It appeared this had been the old roadway before the viaduct had provided a safer route. A stone wall blocked cars from going beyond the view point, although there was little sign of a roadway extending past the wall.

The viaduct was quite an experience descending. Gusting winds, seeming far stronger than those I'd encounted while climbing, caused me to ride in the center of the vehicle lane when descending to provide adequate buffer from the guar rail. Two cars passed me in the opposite (right) lane.

As I descended, the wind got lighter, though, and the ride once again became pure joy. I stopped at the bridge to pick up my water bottle, then continued onward until I met Cara who had ridden east from Jackson Retreat an hour or so after I'd left.

Well, that hadn't been the way I'd hoped to have climbed Arthur's Pass. Little shortcuts on bike maintenance can come back to haunt, and in this case I was stuck with a self-destructing derailleur housing. My Strava time: 26:15, well off the target, but good for third in the rankings so far. All of the other rides have been from the Five Passes Tour: I expect Strava to catch on more here in the next year. Too many good climbs.

San Bruno Hillclimb

2012-01-03T13:49:00.000-08:00

San Bruno this year was a striking contrast to the Low-Key Hillclimbs in which I participated for six consecutive Saturdays from October to November: the nature of the climb and the wind made for an especially tactical race.

San Bruno Hillclimb route profile

I'd registered for E3 instead of my master's 1-2-3 category because, considering the start list at that time, I thought my podium chances were better in E3. To that point, E3 registrations were typically light, but the masters was already heavily populated with big guns. In the "clearly stronger than me" department were Cale Reeder and Kieran Sherlock. Carl Nielson was there: I'd beaten him in our last head-to-head, on Kings Mountain, but that had been the only time in my recollection and the big money was on him to finish ahead. Add in Clark Foy and Tim Clark, both of whom had been training well, who were also in contention for the podium, not to mention the riders with whom I was unfamiliar, and an "in the money" top 3 was looking very unlikely for me there.

In the E3: the two names I knew were Adrien Costa and Zeke Mostov, both riding for the Chipotle Development Squad. I knew they'd keep it together. I was fairly confident against Zeke, as he'd not been able to keep the pace on the Kings Mountain Low-Key. Adrien, on the other hand, while he'd finished only seconds ahead of me at Kings, clearly had more punch than me and would likely win this one, I felt. So I only had to worry about two unknowns in E3. The goal was to follow Adrien.

My pre-ride the day before had confirmed the weather forecast for wind being a factor. There was tailwind on Guadalupe Canyon, then mixed wind on Radio Road. Guadalupe Canyon would be a good place to get away for a strong rider, as the intermittent head and cross-winds on Radio Road would disrupt the coherence of a chance among uncooperative riders. But it was also clear that no effort was too great to stay within contact of a group on Guadalupe, as there would be ample opportunity for recovery on Radio Road with the wind factor.

I'd planned on pre-riding the entire climb. However, my train from San Francisco had been 15 minutes late, and there just wasn't time. So I had to do with "sampling" the beginning of Radio Road, but I didn't specifically know the wind conditions near the finish.

Our group headed out at a nice brisk pace at the base of Guadalupe Canyon. The pace was clearly unsustainable but not super-red zone, and I was able to sit in the group without much issue. I knew that everyone knew there was recovery on the false flat by the intersection with Carter Street (ordinarily a source of delay but during the race we didn't need to worry about the light phase), so I had no fear anyone else found the pace particularly comfortable. Racing is racing and it's all about sharing the pain.

Not long into the section, a rider launched himself off the front. But his progressed quickly stagnated as the Chipotle boys kept the pace. It was still early going and I wasn't ready to bury myself in a

Sure enough, when the grade leveled out, I was able to restore some physiological sanity. All I had to do was to stick for the shorter second steep pitch on Guadalupe Canyon, then it would be mission accomplished for part one.

But then Gregory Coleman of Dolce Vita launched a hard attack. I later spoke with him and he said he thought that was a good place to dig deep because the entry to the park would provide recovery. It was a good tactic: the Chipotle pair chased, as well as a few others. I knew I didn't have the top end to go with these accelerations, and things would very likely regroup on the entrance to the park.

This was a section I'd worried about: not lose time on the semi-technical right turn past the park entry kiosk. I'd practiced this several times, and I wasn't going to take chances, but at least I knew the line I wanted. Here things slightly unraveled, however: the rider ahead of me slowed ridiculously. A second gap to the leaders was lost in an instant. I probably should have more assertively passed him, but instead I let his mistake cost me time.

We started the climb and when the fog cleared from my brain, there were three riders up the road: Gregory in his Dolce Vita kit, Ned Britten in his Davis Cycling kit, and a single Chipotle ride who had to be Zeke, since Adrien was the better climber of the two. Adrien, it turned out, was long gone, but I was clearly closing on the other three. I had an advantage since all I had to do was to catch them and I'd get recovery, but the leader of that group had to keep to a more sustainable pace.

And bit by bit the gap did close. First I passed Gregory, who was perhaps suffering from his early attack, then I caught the other two. The win might have been up the road, but I thought second place was going to come from this group. All I had to do was beat one of them, and I was good for my goal of top 3.

Again I things are a bit of a fuzz here. The winds seemed to be all over the place, sometimes head, sometimes cross, sometimes briefly tail. I tried to use that to my advantage, but none of the other two guys in our group was willing to take charge.

As I finished a pull (I think this is when it happened), Zeke attacked. He got a quick gap, and I should have responded, but instead I waited for Ned to respond. Ned stayed on my wheel, however, so I upped the pace. I glanced back and there seemed to be a gap, so all I had to do was catch Zeke. But the gap wasn't coming down.

Soon we hit the last right-weak left-right combination to the finish. I thought I was good for third here, but Ned came around me and I just couldn't. Time collapsed into a singularity as the cross-wind changed into an apparent tail, shooting us forward to cross the line in our present order.

I was fourth. Again.

Starting with the Low-Key Hillclimbs, my results in races have been 4th, 4th, 4th, 4th, 4th, 5th, and now.... 4th. I was 4th overall in the Low-Keys. Crossing the line, I felt like I'd worked hard, but lacked that totally spent feeling from having dipped mhy feet into the "running from the pedator" reserve. I could have done more.

That's how it is with tactical racing. I could have worked harder, riding hard from early in the Radio Road, and Ned could have sat in for the ride and out-sprinted me. Or he would crossed his personal threshold, and been dropped. I would have liked to have been able to conduct that experiment, because my indulgence "game theory" was a failure. I hadn't met my goal of top three.

Ah, well. At least I'm happy with my time: 16:03. It's said San Bruno is 30 seconds faster than OLH, so that's about consistent with my most recent Old La Honda time, despite my Christmas travel and eight days off the bike (running only). I'd love to put this one in the bank and look forward to my next one, but the next mass-start hillclimb around here is Mt Tam (canceled two out of the past three years) in September. Sure, there's Mt Diablo in June, but that's a time trial: no tactics there, just ride.

For years I've contemplated organizing a USA Cycling sanctioned hillclimb. I've never been able to move ahead with the idea on my own. I need to get the right collaborators: people I know and can work with to make a truly good event. San Bruno is such an event: better today than when I first rode it so long ago. They even had post-ride food for riders, a feature of the Low-Keys.

In any case, while I failed in tactics this day, I am super-happy with my fitness, so time to set another goal and move on. I'm not sure what that goal will be, but I'll have time to think about it during my vacation, where I am now.

posting slackitude and SF2G

2011-12-31T06:50:00.000-08:00

For the first time this year since I started my blog in 2008, total posts have dropped.

Here's a plot of the rate of net posts during the year since 2008, when Cara convinced me to start the blog:

I started a bit slowly this year, then rallied, catching my 2010 schedule. But then I started losing ground, finishing in a dead heat with 2009.

So what happened?

The answer is I started riding more. This is evident from the plot on my Strava page. Barely surviving through April, then my hours start to take off in May. Normal commute means I am in the train around 100 minutes with my laptop. SF2G means I leave home early, then have only my cell phone for my train ride home. More riding = less posting.

And that is a very good thing.

Here's another look at my SF2G schedule. I counted the number of Strava activities I had with SF2G in the title. Before the period of the plot I didn't have GPS so I'd need to check my old training logs. I may have been occasionally remiss in tagging my rides SF2G, so there may be a few not included here. But you can definitely see I started riding into work a lot more often.

Through September 2010, at a previous job, I was more readily able to combine both, since I was more able to squeeze in lunch rides and even the occasional post-commute morning ride. I was on flex time, and could often do actual work on the train since much of what I did didn't require an internet connection (Caltrain has none). Now I need to be on the secure network to do "work", and in any case "face time" is more important, so working on the train then blowing out to join one of the Wednesday morning groups simply doesn't work any more. Working on the train in the morning often meant I spent the evening commute on the blog, but now if I ride in both commutes are blogless.

It's all good: I try for quality, not quantity. What really runs up the numbers is when I get hooked onto a topic, as I did with the Metrigear Vector. Serial posts on a topic become more efficient to get out.

A bit more on the SF2G rate: my SF2G's peaked in August: good weather, plenty of daylight for early starts. September tailed off, in part because I needed train time to prepare for the upcoming Low-Key Hillclimbs (scoring code, web pages). Then October-November my SF2G's were suppressed by not wanting to go into Saturday's Low-Key Hillclimb fatigued. December included holidays, and was also off.

For 2012? Who knows how the blog will go? I'll try not to set any goals. At some point I will have said enough and feel it's time to move on. But I'm not quite there yet.

Running the Rocky Steps

2011-12-29T15:32:00.000-08:00

This weekend I was in Philadelphia for the first time since I was a small child, visiting Cara's family. The house was only a mile from the Philadelphia Art Museum, and that got me very excited. Sure, we went to the museum (or rather the Perelman Building extension: a fascinating exhibition of Zaha Hadid's architecture), which was nice, but arguably far more famous than the museum itself is the steps leading to the front door of the main building, for these were the steps Rocky Balboa used to prove his fitness in Rocky, the Best Picture Acadamy Award winner of 1976.Here's a link to a YouTube version of the inspirational scene. You know you like it!

As I began the short run to the museum, I readied myself for the ridicule of bystanders. Here would be an adult living out a scene of a 35-year-old film: Heh! Look at that bozo! "Go Rocky!" Heh.

First, the statue. I had to take a photo. There's a long story on the statue, which was commissioned for Rocky III for the top of the steps. Some objected this was inappropriate for a museum, and it was moved away. But popular demand brought it back to the museum, to the side of the bottom of the steps: less intrusive. However, it still attracts a steady stream of tourists taking photos: I was just one of many during my time there. I never saw anyone taking pictures, or even looking at, any of the several other statues on the museum grounds.

Rocky statue

It was time to run the stairs! I felt strangely nervous. What if I couldn't make it? What if I doubled over, out of breath, like the "before" version of Rocky in the film? The horror... And then there was the mockery I was sure to suffer from surrounding tourists. I readied myself to ignore it.

But as I walked over the the center of the bottom of the steps, I was amused to find that several others were already on the steps, all running. And not just running: it was pro-forma upon arriving to the top to raise fists in the air, jumping up and down.

I had no plans for that, so felt safe. And up I went: taking the stairs two at a time, the biggest challenge to me was to not become disoriented and miss-step. This happens to me on stairs: I'm fine at first, but as I proceed, I feel increasingly out-of-phase with the stairs, until I'm fighting just to keep running without tripping.

I hadn't reviewed the video before running, and so had been mistaken that Rocky runs all the way to the museum doors. After the first steps, there's a broad mezzanine around a fountain, then a final short flight of steps to the front doors. This made a better interval for me than to just do the initial set. Yet others around me stopped upon arriving at the mezzanine. But better safe than sorry: I ran the whole way.

After reaching the top, I had a nice view of the city.

view from the museum doors

Once simply wasn't enough! I had to do it again. I decided three would be a good number of reps, then continue with my run. But after three, I reset my goal to five... then eight... then finally ten. After my tenth repetition, which I did single-step instead of double as an experiment, I decided I'd had enough, decended, and then began running along Kelly Drive of Cycling Classic fame. During the entire time I was on the stairs I was never the only one running there: there were a constant stream of runners, most slower, one faster, running the steps. It was incredible.

Strava times up full steps. It's unclear from data whether slow time on last iteration is due to running steps one at a time or due to fatigue.

I was convinced that surely I'd outrun Rocky. To check, I uploaded my data to Strava for segment timing. To my shock nobody had defined a segment, so I defined two: one to where Rocky stopped, another to the museum door.

So then I timed the videos. I already linked to the original film. I timed him there at close to 13 seconds. I was crushed: my best time via the Strava segment, which starts and finishes on the steps to provide some room for error, is 18 seconds (note only a few of my attempts match the segment, since the stairs are wide relative to the stair length, and I started from various positions). But then there's a sequel where Rocky comes back, now a star, fitter and faster than ever. Here's a link to that sequence. His time there: 10 seconds by my timing. I wasn't even close.

The key is while I took the stairs two at a time, Rocky takes them 4 at a time. Impressive! Obviously I have my work cut out for me if I want to challenge Apollo Creed myself.

San Bruno Hillclimb: Jan 1

2011-12-19T08:37:00.000-08:00

It is traditional that the top 3 men, top 3 women, top 3 juniors, and the Endurance Award winner of the Low-Key Hillclimb series are all awarded a free spot in Pen Velo's long-running San Bruno Hill Climb, held every year on Jan 1. I finished 4th, just out of the "money". But yesterday I signed up anyway, paying the entry fee: it's not often I have fitness and opportunity this time of year to do the climb.

The USA Cycling page has it listed as a "time trial", but that's incorrect. It's a mass-start race, riders starting in waves from the base of Guadalupe Canyon Road near Bayshore, climbing to the "saddle point" marking the top of Guadalupe. Then from there it's a sharp right into the state park, down a short descent past the ranger kiosk, an immediate right turn over rough pavement a short straight, another right, pass back under Guadalupe Canyon, then the narrow, sometimes rough climb to the summit. Here's the profile:

The rating listed on the plot uses my algorithm, which is normalized to Old La Honda being 100. The rating is lower because of the relatively lower average grade. However, given the distance, the times aren't that much faster than Old La Honda road, the difference depending on conditions (San Bruno is much more exposed to the wind).

It opens with surprising steepness, then levels out for a traffic light. Past this light (annoying on non-race days when it's too often red) the road climbs somewhat steeply again until the grade lessons, then flattens completely at the Guadalupe Canyon "saddle". Radio Road passes underneath, visible to the right, while to the left is a clear view to the summit. Then comes the tricky bit: the net 270 degree turn onto Radio Road and back under Guadalupe Canyon. You absolutely don't want to fritter away any seconds here, but given the poor road condition, taking a corner too hot is a real danger.

Once under Guadalupe, Radio Road really comes into its own. This climb begins moderately, but it becomes progressively harder to hold onto a gear until at a sweeping left turn the grade hits its peak of almost 10%. Now it's the end game: first a right turn which appears it might be the end, but following that a second sweeping right leading to the finish line at the radio towers (watch out for the 6-inch gap in the metal plates on the road!)

Here's a Strava segment:

As I write this, I'm 18th, a result I got in a training ride on November 13th. I'd like to improve that time on Jan 1. Unfortunately a lot of other riders are also likely to post good times on Jan 1, so if I can just hold my position in the rankings, that wouldn't be too bad! But I think I can reasonably hope to just crack the top 10 when the day is done, conditions permitting.

It's a fun race to watch as well as well as ride. 2007 was an off year for me: I'd ridden well at the Low-Keys in fall 2006, but IEDM and other factors got in the way of me keeping enough fitness to make riding San Bruno worthwhile. Instead I took photos. Here's spectators near the summit watching riders pass below:

The view of Guadalupe Canyon Road below is nice:

Here come the leaders! It's newbie Chris Phipps leading hillclimbing Low-Keyer Tracy Colwell! San Francisco is in the background. Chris would go on to take the win.

The next year, after I had done the Youth Hostels International "Christmas" bike tour out of San Diego, returning the day before, and was tired but at least had some fitness. So I gave it a shot. I did not ride a smart race. Here's the results.

On Guadalupe Canyon, I resisted the surge of riders near the front, preferring to ride a steady power the whole climb. This works for a climb which is a steady grade in consistent conditions, but San Bruno has neither. First there's the false flat where being in a pack is a benefit, so the faster lead group will gain time on slower trailing groups. Then there's a chance to recover on the descent and turn, so it pays to be slightly in the red going into this section. The worst part was the wind: it was a strong head wind on Radio Road that year and so whatever group you were in, you stayed in. Bridging gaps was extremely hard. The real race was to the park entrance, and I arrived there with way too much left in the tank.

In 2009 I was in Southeast Asia. But this video shows conditions aren't always ideal on San Bruno mountain:

San Bruno Hillclimb 2009 from Chris Stastny on Vimeo.

2010 I missed as well (more focused on running, and the forecast had been for rain), then in 2011 I was woefully out of shape due to 11 hours days at work at my then-new job.

But back to 2012....

As I noted, it's tough coming into Jan 1 ready to go. As is the case most years, this year I'm traveling to the east coast to visit family for Christmas week, during which time I won't have a bicycle. I'll probably run a bit and go to a YMCA or other local gym where I will ride a stationary bike and do some light lifting and core work. Hopefully things go well.

Fairwheel Bike's Project Right @ NAHBBS in March

2011-12-15T18:00:00.000-08:00

I'm very excited about the 2012 North American Handbuilt Bicycle Show in the first weekend in March. (I was tempted to ride the Death Valley Double Century, but the two conflict.... Death Valley can wait.)

Fairwheel Bikes, which has been stealing a lot of attention from the big boys at Interbike the past few years with their ultra-light project bikes, is delivering a new project to the show: "Project Right". I just came across this today, having seen their latest blog update. Here's the frame:

Check that out: the left side of the rear triangle is completely missing, and there's no seat stays. This is really only an incremental change from the trend taken by Cervelo and Pinarello. Cervelo reduced the seat stays to mere formalities to provide vertical compliance, while Pinarello has been a leader (in marketing at least) in focusing material on the right chainstay, since that's the side where force is transmitted via the drivetrain.

Here's the Cervelo R5 in its most expensive form (the R5-Ca). As an aside, I finally came across one of these "in the wild", on the bike path in Sausalito. It looked really nice. I've stopped obsessing over how outrageously expensive it is. If people want to squander money, bikes is a fairly harmless way to do it. Anyway, notice the super-skinny seatstays:

You may as well cut those out: they're hardly providing structural support.

So that's what Fairwheel is doing: no seatstays, only one chainstay. Seems like a progression of a trend. But it's hardly new.

Chris Boardman's Lotus Superbike had a similar concept:

Nothing of interest on the left side; all support for the rear wheel is on the right. This bike was based on the Windcheetah, which Mike Burrows designed in 1982:

But the idea goes way, way further back than that. Documented by Jan Heine in his book, The Competition Bicycle, here's a Labor bike from 1906:

The design was to allow for quick wheel changes.

I really look forward to seeing the Fairwheel Bikes project "in the flesh" at the Hand Built Bike Show.

Old La Honda Road: another PR

2011-12-12T20:31:00.000-08:00

I had a good year at the Low-Key Hillclimbs this year. Often by the time October rolls around fatigue from the year is starting to kick it. Last year I came into the series fairly fresh and fit, but then started a new job and my fitness went straight downhill from week 3 (my first) onward.

I finally started exercising at a reasonable rate again in April this year, a mixture of running and some cycling (mostly long commutes to work), and surprised myself with a sub-1:31 half marathon in August: I considered that good given my lack of formal running background. So I knew I had some fitness but wasn't sure about how I'd do on the bike. I did a few climbs of Diablo before the series, just to get some climbing legs, and was surprised my times weren't so bad. But for the Low-Keys, you've got to be better than "not so bad". Everyone seems to raise their game for the series.

But despite my worries I did pretty well. So as the series wound down, Tim Clark asked me if I was going to try for an Old La Honda time. Old La Honda is, to me, the most prestigious climb for times. More riders know their Old La Honda times than any other.

I've been adding some volume to rebuilt my aerobic base prior to the San Bruno Hillclimb on the New Year: during the series it was more about working hard Saturday to Tuesday then being fresh for Saturday.

So I got in some solid work Thanksgiving weekend, then the following weekend. I was pretty tired on Monday, so went super-light on my usual Monday weight room workout. Tue I felt a bit better, but a work meeting kept me from either riding in or riding at lunch. So I was starting to feel fresh again Tuesday night: a good change to give Old La Honda a shot on the Wednesday noon ride. It wasn't optimal: a few days of light work following my recovery would have been better. But I valued the chance to ride with the group rather than try a solo effort.

My last PR attempt on Old La Honda had been successful: in July 2009 I had set out to break 17 minutes and was delighted with my result, 16:49. It was a redemption against having failed miserably at the Diamond Valley Road Race the weekend before, where failing to follow through on my previous two strong races there, I'd been dropped on the first lap. For that attempt I'd ridden a very steady pace, turning my 36-18 gear until near the top, where I upshifted to my 17-cog.

It might seem surprising I'd not tried again. But a good run at Old La Honda requires that I be fit and rested, and that I'm willing to bring my Fuji with its climbing wheels on the train. The key thing here is fit and rested. Devoting a three o r four day block to a good Old La Honda time is a luxury in which I rarely indulge. So most of my rides up the hill are for training only, typically on my Ritchey Breakaway with clincher wheels, carrying water bottles, a heavy tool bag, and a pump. Almost always I've I'd done some sort of hard ride the day before, as well.

So I risked bringing my Fuji SL/1 with it's light wheels and 180 gram Vittoria Corsa time trial tubulars on the morning train and from there to work. At lunch, I met Mark Johnson and we rode along glass-strewn Central Expressway on our way to the ride start: a serious risk for the ultra-thin tires and no spare. But we made it, and I felt fairly good.

Doing good OLH's from the Wed Noon Ride used to be compromised by a fast pace over Arastradero and Alpine Road, then jamming to the sprint at Woodside. But no longer: the focus is now, thanks to Matt Allie's leadership, strictly on the climb. So the ride to the base is at a nice warm-up pace.

Conditions were fairly good for the attempt. The road is in historically good condition, most notably the upper portion which was repaved within the past few months. And winds were forecast to be light, from the north, a tailwind (winds are generally blocked by the trees on the climb, but every little bit helps). Cool temperatures had me carrying more weight in clothing than I'd prefer, but as we approached the climb and were stopped by construction crews on Portola Valley Road, Peter Tapscott offered to carry my jacket and empty water bottle for me. My excess mass was in my long-sleeve undershirt, my heavy long-fingered gloves, my compression tights, additional calf compression socks,and the wristwatch I'd forgotten to remove before leaving my office. All of this adds up, probably worth 2-3 seconds on the climb, the wind resistance from the gloves maybe an additional second. But I couldn't pick the weather and I wasn't going to help myself by under-dressing. Some recent emergency repairs to my bike due to a broken Power-Cordz added some additional weight: at least another second there. It all adds up.

As we hit the base of Old La Honda I made a tactical error. My plan had been to start a bit back, then move up to the early pace-setters, essentially shaving a second or two. But this back-fired, as nobody took control of the pace, and I was stuck in traffic. I went to the extreme left of the road, moved up, and went to the front myself. Not the best way to start: I should have hit the bridge marking the start of the climb at speed and carried momentum up the initial slope.

I was joined by Chris Zappala, who'd mentioned he'd commuted to Palo Alto from San Francisco earlier that morning. Despite the hard ride in the cold, he was ready to ride, so I got on his wheel for early pacing. As I climbed I looked down at my cassette to assess my gear. I had a 12-23 cassette, so tried to count the number of cogs between the chain and the end of the cassette and figure out which gear it was. Was it 17 or 18? It felt as if we were going fast, in either case, but I doubted if I was spinning the 17 that I would be able to sustain that pace.

I came across James Porter and Greg McQuaid climbing together. I'd expected them on the noon ride itself but they'd clearly decided to leave early on their own. Tim Clark was going to be with him but I didn't see him. Greg was recovering from a broken collar bone so I didn't expect him to be in top fitness right now, so I focused instead on riding with Chris.

Chris eventually faded, as expected given his morning commute, and I moved ahead. I hit the first mailboxes in 5:50-something, the fastest I'd ever done. That didn't seem to necessarily be a good thing. Then I noticed that I was indeed in my 17 cog, so downshifted to the 18. This put me back to where I wanted to be, gear-wise, but not where I wanted to be feel-wise. I was becoming distressed, struggling to maintain pace through the steeper corners, and there was still a long way to go. In 2009 I'd still be climbing seemingly effortless at this point.

I came upon Tim Clark, whom I was able to catch. For awhile he served as a nice rabbit, but eventually I caught him and he cheered me on as I passed. But I was feeling a growing sense of despair. I tried to suppress all negative thoughts, to simply push onward, but I couldn't deny that I was struggling.

I shifted down into my 19.

Now it was official: I'd gone from "get back onto pace" to "minimize losses". Sub-17 was still possible, I told myself, keep up the pressure, don't let up. Even if I were to miss 16:49, I still wanted to make Strava page 1, which required 17:08. This was no time to indulge in self-pity!

I looked down at my computer and saw 16-even with a few turns left. I knew that these turns always eat up way more time than expected: it always seems time accelerates here. But I kept pushing, all hope of smooth form now gone.

There it was: the mailboxes, then the stop sign. As I approached I looked again at my computer, the first time since I'd seen 16. 16:35 it said just as I went to hit the lap button at the stop sign. I'd done it: a new PR. Official time (from the Garmin) = 16:36.31.

It was with interest that I uploaded the data from the Garmin, and sure enough, the numbers tell me what I already knew from my gear selection: I faded.

To generate this plot I differentiated my altitude versus time, then applied a bi-exponential convolution with a 5-second time constant to smooth the data. To avoid issues with the smoothing at the beginning and end, I excluded the first and the final 500 meters. This generated the data with the red points. I then did a non-linear regression, fitting a decaying exponential to the result (the exponential versus distance, as opposed to time, but the two yield essentially equivalent results). This fitted function is plotted with the dashed curve. On the left of the plot is the rate of ascent represented as meters/hour ("VAM"). On the right I show the projected Old La Honda time for each VAM. The green background is sufficient to break my previous PR of 16:49. Red implies I'm losing ground versus that result.

You can see I obviously started way too fast: a 15-minute Old La Honda pace. Some people can do that (Ryan Sherlock beat 15 minutes the weekend before my ride, although Eric Wohlberg has the unofficial self-timed record at 13:50). I cannot. So as the climb progresses, the trend line has my climbing rate decrease at 3.4% per kilometer. Finally at around 4.5 km I crack fairly badly, my pace dropping even below this decaying trend. But I rallied toward the finish and regained some of that lost time.

Honestly: if I'd known my pace was going to be this far from uniform, I'd have said no way would I have been successful. But I was successful, which tells me that I am able to go out fairly hard, too hard, and still retain enough to do well for this interval. This is important for the San Bruno Hillclimb on 1 Jan, where there are considerable draft advantages for staying with a pack near the beginning, and given my experience here, it seems I can afford to go into the red to hold onto early wheels there. San Bruno is a relatively tactical climb, as there's an unavoidable recovery half-way as the route descends and turns under the main road on which it begins. Then from there the winds can make it very difficult to bridge gaps which may have formed on the first portion. So holding onto a good group is really important.

Anyway, it will be fun to see how it goes. I'm hoping for good weather on Jan 1.

Level of Service Analysis and Bus Rapid Transit

2011-12-05T06:17:00.001-08:00

Here's a good story: from the excellent StreetsBlog San Francisco, What's the Hold Up for Van Ness BRT?.

San Francisco has no meaningful intra-city subway, and most of the street car lines were ripped out early in the 20th century. Street cars were mostly privately owned, and they couldn't compete with the massive public support provided to car infrastructure under political pressure from the car companies. So en masse, street lines were paved over, except for a skeletal few: there's a few key MUNI lines, and a few cable car lines which cater to tourists. BART runs through downtown, but just makes a few stops in the city, designed primarily to connect San Francisco to the East Bay. It's original goal of surrounding the Bay was gutted by first Santa Clara County opting out in 1957, then San Mateo County in 1961, each preferring to focus on expressways; Marin dropped out soon after [Wikipedia] (what an unbelievable tragedy).

So San Francisco is stuck with buses for the vast majority of its area as public transit. Yet despite being carless, I take the bus only a few times a year. Why? Because it's increadibly slow and unreliable. If I decide now I want to get somewhere, more often than not I can actually run to a destination before I'd arrive by bus. It wouldn't be uncommon that I could arrive at my destination before I'd even stepped on the bus, since delays of up to 40 minutes for the next bus are not at all uncommon with buses running on a nominal 20 minute interval. Even with an even start I can often outpace the bus by foot. And by bike it's simply no contest.

So San Francisco proposes to give buses priority lanes. This is called "Bus Rapid Transit", the idea being to package bus lines like more expensive "light rail". You step on the bus, it zips down its own lane triggering traffic lights as it goes, and presto-magicko, you're at your destination in no time. Awesome! Public transit problem solved!

BRT Geary (Streetsblog San Francisco)

Solved, except that infrastructure projects must go through environmental impact analysis (CEQA) and a big part of environmental impact analysis is the effect on automobile traffic. This makes sense since automobiles are responsible for an enormous environmental impact: noise, pollution, resource depletion, and public safety. No brainer in this case, right? After all, an efficient bus line will take drivers off the roads by providing a timely, more efficient alternative, reducing the environmental impact of cars! No-brainer, let's go!

But wait! CEQA uses an analysis based on a fixed "level of service" (LOS). The assumption is a certain number of cars will use the road each day, and if you slow these cars down, they'll generate more congestion, their engines idling all the while. With this logic, the more car lanes, the better, as the fixed number of cars on the road will then zip to their destinations with a minimum of "impact".

This is obviously absurd. Build more lanes, more people drive, those who drive drive more often, and when they drive they drive further. This is seen time and time again in the data, in study after study. Cara was in Amsterdam recently, reporting on the enormous number of cyclists using the bike paths, cycling providing clearly superior local transport the cars, which are relegated to narrow roadways. There's a nice video here.

The LOS logic? Pave over the bike lanes: extend the roads by adding lanes. This will have the cars generate less exhaust during their trips. Sure, bike riders will be slowed as they struggle to find space on the public roads, but there's not much environmental impact from a slow-moving bike, while slow-moving cars are spitting out exhaust fumes and making noise every extra second of their trip.

I'm sure the Dutch would laugh with well-deserved contempt at this stupid demonstration of the pathetic state of U.S. public education were we to suggest this to them. And they'd have a point.

The problem with the LOS standard has been well known in San Francisco since the 2006 injunction against all new bike infrastructure in San Francisco, an injunction which was lifted only last year. The bike program, it was decided, needed to provide an environmental impact analysis, and analysis which was not able to claim that if you make a city better for cyclists maybe, just maybe people would replace car trips with bike trips, reducing the need for car lanes. The bike program was eventually able to survive even that silly standard, and is progressing well today due to the unrelenting political pressure of the San Francisco Bike Coalition. We can only hope BRT is able to get past the LOS hurdle as well, and even better that the LOS standard be quickly and definitively revised to recognize that driving a car a certain number of times each day is not an inevitable fate.

Strava power estimation: Cortland Hurl

2011-12-04T06:33:00.001-08:00

The Cortland Hurl is the only significant climb on the SF2G Bayway route. According to Strava, it gains 25 meters in 400 meters, an average grade of 6.3%, although the grade is non-uniform. It starts out fairly gradual, then steepens, then gets gradual again towards the top. I like to make a good effort here when I'm feeling good during morning commutes. Typically I'm behind at the top of the steep bit, but I tend to do fairly well on the final gradual portion. If I'm having a good day, depending on who's there and how they're riding, I have a chance to be first to the top.

I've not ridden with a power meter for a year now. I sort of lost interest: I just like riding my bike and I don't care what the power meter data are, so why carry around a heavy, expensive Powertap wheel? Strava gives me a fairly good idea how I'm doing with its segment timings.

However, in addition to speed numbers Strava also produces power estimates. In fact, it will use these estimates for reporting a rider's best effort over different time intervals. I have suggested to them this is a mistake: only power meter numbers should be used for this purpose, since their estimate is unreliable.

Since I had a lot of data for Cortland (28 rides + one I reject because I dropped my keys along the way and had to turn back to fetch them), I figured it would be interesting to compare Strava power to PowerTap power, plotted versus VAM which is considered a decent surrogate for power/mass ratio on climbs. I also compare these with "hand calculations" using the usual power-speed model. Here's the result:

The analytic calculations assume constant speed and constant grade and no net acceleration. Assuming start and finish speed are the same, they are thus a lower bound estimate given the assumptions used. For assumptions, I show three. One is using a "realistic" estimate for total mass and for CdA, the coefficient for wind resistance power. For this one I set my body mass to 57 kg, my bike mass to 8 kg, but then added in 4 kg for equipment and clothing and what I was carrying on my back. I assume a 0.5% coefficient of rolling resistance. CdA was set to 0.5 meters squared. For a second estimate, I eliminated the clothing + equipment mass and reduce CdA to 0.4 meters squared, assuming no back pack. In a final estimate I eliminate all wind resistance.

As can be seen in the plot, the Powertap measurements are almost always more than the Strava estimates. Strava estimates start out fairly well aligned with the zero-equipment-mass ("naked") estimate of power, except for my fastest runs where the Strava estimates become much lower, dropping as low as the zero-wind-resistance estimates. There's a few Strava estimates which are clearly anomalously low.

The Powertap data fall above even my highest analytic power estimate. This is as they should, since as I noted my analytic estimate assumed uniform speed and grade, and thus underestimate wind resistance near the beginning and end of the segment. Wind resistance is superlinear, so underestimates during faster than average portions are of a lower magnitude than overestimates during slower than average portions.

So what do I learn from this? Basically you shouldn't trust Strava power estimates. But if you do care about them, you should make sure clothing + equipment mass is included in bike mass. I had not done this. It all adds up: toolbag, pump, water bottles, clothing, shoes, helmet: you'd probably be surprised at the result if you bundled it all up and put it on a scale.

Low-Key scoring algorithm: addition of variance normalization

2011-11-28T18:00:00.000-08:00

As always happens in fall, the Low-Key Hillclimbs have taken up a large chunk of my time, leaving less time for blog posts. But it was worth it: the series was an unqualified success, with every climb coming off well, the last few finding valuable seams in the weather. At Hamilton riders experienced moderate rain on the descent, and for some towards the end of the climb, but it was warm enough that the long descent was still tolerable in the wet.

One aspect of the series worthy of revision, however, is the scoring system. Never before were artifacts in the median-time-normalized scoring more obvious. So for 2012, I am finally overcoming inertia and changing from the median-based scoring we've essentially used since 2006.

I've described in preceding posts a scheme to calculate a reference "effective" time for each climb. With this scheme, instead of taking a median each week, we take a geometric mean where effective times for riders (adjusted for male, female, hybrid-electric) are adjusted by the rider's "rating", which represents how the riders tend to do relative to the reference time. It's an iterative calculation which is repeated until rider ratings and reference times are self-consistent, weighting means by heuristic weighting factors to give higher priority to riders who do more climbs, and climbs with more riders, since these provide better statistics.

Here's a comparison of this approach with the median-based system used this year. I plot on the x-axis each rider's rating and on the y-axis that rider's score for each designated week. In this case I used weeks 5 (Palomares Road) and 6 (Mix Canyon Road). These climbs are at opposite ends of a spectrum: Palomares is short with plenty of low-grades, while Mix Canyon is relatively long with extended steep grades.

Here's the plots. I've omitted riders who did only one climb, as for them their rating from the one climb they did is equal to their rating.

With the 2011 scoring scheme, you can clearly see that there is a lack of low-weighted riders relative to Palomares. As a result, moderately-rated riders in particular were given low scores, since the median rider was, relative to the entire series, above average (rated over 100). In contrast at Palomares there were more low-weighted riders.

So then I replace the median time with a reference time, adjusting each rider's effective time by his/her rating. Now you can see the scores for Mixed Canyon have been boosted:

But there's an issue here: the curve for Mix Canyon is steeper. So relatively slower riders score lower, while relatively faster riders score higher, then they did or would at Palomares. So I added a bit of complexity: I compare the spread in scores with the spread in rider ratings and I make sure that the ratio of these spreads is the same week-after-week. I call the adjustment factor the "slope factor". The result is here:

Now the curves line up nicely! Sure, each rider may score in a given week more or less than his rating, but the overall trend is very similar.

I'll add in the other weeks. First, here's the 2011 formula:

You can see distinct curves for different weeks. Some weeks a rider of a given ability is more likely to score higher, some lower. This isn't what we're after, as we want riders to have the opportunity to excel on any week.

So I add in the adjusted effective reference time, and then the slope factor, and here's what we get:

All of the weeks have generally overlapping curves. No more fear of turning our for a tough climb or a climb in difficult conditions, and have your score buried in obscurity because there's a disproportionate number of fast riders. Or similarly, no more volunteering for a week only to have your volunteer score end up lower than riders you finish ahead of week after week, simply because the median times were relatively long due to rider turn-out.

To me, this system looks like it's working nicely.

week-to-week stability of proposed 2012 Low-Key scoring formula

2011-11-14T19:29:00.000-08:00

In two previous posts, I described an attempt to revise the scoring code for the Low-Key Hillclimbs. The scoring has placed a priority on simplicity. At first, we normalized times to the fastest man and woman each week. But then everyone's score was exceptionally sensitive to the fastest rider. Then I switched to using the median time for normalization, first separately for men and women, then using combing them with an empirically determined conversion factor for women. But while median is less sensitive to any single individual showing up, nevertheless the most challenging climbs tend to attract fewer beginner riders, deflating the scores for these weeks. So the alternative approach is to iteratively rate each climb using a reference time based on the rating of riders who show up, and assign each rider a rating based on the reference times (and their results) of the climbs they do.

A concern about this approach is that if I use all available information equally, I re-rate each rider and each climb after each week's results. This yields scores for previous weeks changing each time results for a new week become available. This is in principle an undesirable feature. It could be avoided by forcibly freezing ratings for climbs each week, rating only new climbs using results including those which preceed it. You might call this approach causal scoring (nothing is affected by future events). However, before taking such a compromize, I wanted to test whether pragmatically this is a problem. Obviously if relative scores from previous weeks are highly volatile then it makes tactical decisions difficult. For example, your previous scores might all be better than the scores of another rider, then you mark him and out-sprint him in this week's climb, but afterwards you've fallen behind in the standings because of a re-evaluation of the reference times for previous weeks. This is something of a pathological example, but it's in principle possible, so it needs to be tested using realistic data.

So I ran the scoring code for 2011 data, which exist for seven weeks of climbs. Two climbs, Kings Mountain Road and Mount Hamilton Road, have not yet occurred.

After week 1, Mountebello Road, there is only one climb on which to determine a reference time, so I revert back to using the median time. I could also use the geometric mean, which would be closer to what I do when there's multiple times, but the median works well so I stick with that. The climb's field is then by definition average. There is no adjustment for the strength of the field.

Then I add data from week 2, Sierra Road. Now we see that some riders did both weeks. On one or the other week, using median times, these riders would score lower (it turns out they generally would score lower on Sierra Road). I then assume that on the week they score lower the average rider was stronger, and adjust the climb reference time so riders who did both, on average, score the same (using geometric means). Then each week other riders are scored relative to these repeat riders. This causes a re-evaluation of the reference time for the first week: it's no longer the median time.

Now I add week 3, and I can use all data from riders who did at least two of the climbs to rate the fields of riders relative to each other. These riders are used to re-establish reference times for the first two weeks.

And the process continues until I have added data from all seven weeks.

Here's the test results. First I plot the ratio of each week's median time to its reference time. So if this number is more than 100%, that means that the reference time will be less than the median time, and riders will score lower than with the 2011 algorithm. This adjustment is because, according to the algorithm, on average there were more slower riders and fewer faster riders on that climb. The plot shows this ratio for each climb plotted at the end of each week. After one week, there is only one point: for week 1, Montebello, and of course since that climb uses the median time it is at 100%. After two weeks there are two points: one for Montebello and one for Sierra Road. That curve is orange. Here Montebello is around 102.5% and Sierra Road is around 97.5%, so there were stronger riders at Sierra Road. Week 3 was Page Mill and that came out between the first two climbs. You can see how each week the reference time for each climb is adjusted, generally upward since as the series has continued later climbs have attracted on average stronger riders, it seems. So each week scores from Montebello, week 1, would tend to drop a bit as the algorithm assigns relatively higher scores to riders with a similar relative placing at later climbs.

This seems like it might be a problem, having things change over time. And this is true for someone who has a particular score goal, like 100 points. They may have 100.1 points after Montebello only to find that has dropped to 95.1 points later in the series. But for the standings, all that matters is how points from one climb compare to points of another. For example, if after two weeks rider A, who climbed only Montebello, scored 101 points there while rider B, who climbed only Sierra Road, scored 100 points there than rider A is ahead of rider B. After week 3 perhaps rider A's score for week 1 drops to 99 points and rider B's score drops to 98 points, but that's okay as long as the gap between the two doesn't change much.

So next I plot the ratio of a climb's reference time to the reference time for Montebello. If the two change by the same proportion this ratio doesn't change, and a comparison of riders between the two climbs won't change much. As hoped, this ratio doesn't change much as new results are added to the analysis.

The resolution of that plot is limited, so in the next plot I how much each of these ratios changes after each week of results is added. Using the example before of riders A and B, for rider A to keep his 1-point gap over rider B, we want this ratio to be stable to around 1%. From the plot you can see that none of the comparisons between any of the weeks and week 1 changes by more than 0.5%. The biggest change is between week 2 and week 3, but still these change relative to each other by barely over 1%. So scores shifting relative to each other over the course of the series doesn't seem to be a big problem.So the scoring system seems to work pretty well, at least if you don't mind scores drifting a bit together.

testing 2012 Low-Key Hillclimbs scoring code

2011-11-12T05:07:00.001-08:00

I seem to have debugged the new Low-Key Hillclimbs scoring algorithm, so tested it on 2011 data for the completed first six weeks.

Recall the method is to calculate a rider's rating (not used for overall rankings) based on the natural logarithm of the ratio of his time each week to that climb's reference time. Meanwhile the climb's reference time is calculated as the average the natural logs of the times of the riders in the climb, subtracting their ratings. These "averages" are weighted by heuristic statistical weights which assign more importance to riders who did more climbs, and to a lesser extent to climbs with more riders. Each of these factors depends on the others, so the solution is done self-consistently until it converges, in this case until the sum of the squares of the reference times changes by less than 10^-6 seconds². This took 8 iterations in my test.

To avoid contaminating the results I check for annotations that a rider has experienced a malfunction or wrong turn during a climb, or that he was on a tandem, unicycle, or was running. These factors would generally invalidate week-to-week comparisons for these results, so I don't use them. So a rider whose wheel pops out of true during a climb and is forced to make time-consuming adjustments before continuing won't have his rating penalized by this, assuming that incident makes it into the results data.

All times here are adjusted for division (male, female, or hybrid-electric), as I've described.

week 1 median    = 2149.50
week 1 reference = 2054.26
week 1 ratio     = 104.636%
week 1 quality   = 0.0398
week 2 median    = 1760.50
week 2 reference = 1762.51
week 2 ratio     = 99.886%
week 2 quality   = 0.0096
week 3 median    = 2614.00
week 3 reference = 2559.27
week 3 ratio     = 102.139%
week 3 quality   = 0.0237
week 4 median    = 2057.50
week 4 reference = 2119.96
week 4 ratio     = 97.054%
week 4 quality   = -0.0140
week 5 median    = 1237.50
week 5 reference = 1246.35
week 5 ratio     = 99.290%
week 5 quality   = 0.0310
week 6 median    = 2191.00
week 6 reference = 2322.56
week 6 ratio     = 94.335%
week 6 quality   = -0.0254

Here the week "quality" is the average rating score of riders in the climb. You can see in general the ratio of the median to reference times tracks this quality score, although one is based on a weighted geometric mean, and the other is a population median.

In general less steep more popular climbs (1, 3, 5) have rider "qualities" which are positive, meaning times were somewhat slower, while steeper, more challenging climbs (4 and 6, but to a lesser extent 2) tended to have negative "qualities", indicating riders were generally faster. The exception here is week 2, Sierra Road. While this road is considered cruelly steep by the Tour of California, apparently Low-Keyers have a higher standard of intimidation, and it still managed a positive quality score with a ratio quite close to 100%. It essentially fell between the super-steep climbs and the more gradual climbs.

A side effect of this, even if I don't use this analysis for the overall scores (this year's score algorithm can't be changed mid-stream, obviously, although it's tempting, I admit...), is I get to add a new ranking to the overall result: rider "rating". This is a bit like the ratings that are sometimes published in papers for rating professional teams, not a statement of accomplishment, but a guide to betters on who is likely to beat whom. Don't take these results to Vegas, though, as they're biased towards riders who did steeper climbs, which produce a greater spread in scores. I could compensate for this with an additional rating for climbs (how spread the scores were), but I'll leave it as it is. I like "rewarding" riders for tackling the steep stuff, even if it's only in such an indirect fashion.

For the test, I posted the overall results with the official algorithm and with this test scoring algorithm so they can be compared. One thing to note is only this single page is available with the test algorithm, any linked results will be the official score:

Riders who did both Mix (week 6) and Bohlman (week 4) really benefit from this new approach. Coincidentally that includes me and my "team" for the series (Team Low-Key, even though my racing team is Team Roaring Mouse, which I strongly support).

proposed 2012 Low-Key Hillclimbs scoring algorithm description

2011-11-11T10:35:00.001-08:00

The whole key to comparing scores from week-to-week is to come up with a set of reference times for each week. Then the rider's score is 100 × this reference time / the rider's time, where times have first been adjusted if the rider is a woman or a hybrid-electric rider.Presently this reference time is the time of the median rider finishing the climb that week. But if riders who would normally finish in more than the median time don't show up one week, for example Mix Canyon Road, everyone there gets a lower than normal score. That's not fair.So instead we can do an iterative calculation. Iterative calculations are nice because you can simplify a complicated problem by converting it into a series of simpler problem. The solution of each depends on the solution of every other. But if you solve them in series, then solve them again, then again, eventually you approach the self-consistent solution which you would have gotten with a single solution of the full, unsimplified problem, except that problem might be too difficult to solve directly.So here's how we proceed:

For each climb, there is a reference time, similar to the median time now. The reference time is the average of the adjusted times for riders doing the climb.
For each rider, there is a time adjustment factor. The time adjustment factor is the average of the ratio of the rider's time for a week to that week's reference time. So if a rider always does a climb 10% over that climb's reference time, that rider's adjustment factor will be 1.1.

We have a problem here. The climb reference times depend on rider adjustment factors, and rider adjustment factors depend on climb reference times. We need to know the answer to get the answer.But this is where an iterative solution comes in. We begin by assuming each rider's adjustment factor is 1. Then we calculate the reference times for the climbs. Then we assume these reference times are correct and we calculate the rider adjustment factors. Then we assume these are correct and we recalculate the climb reference times. Repeat this process enough times and we get the results we're after.Once we have a reference time for each climb, we plug these into the present scoring algorithm where we now use median time, and we're done. The rest is the same.One minor tweak: not everyone's time should contribute equally to a climb's reference time, and not every climb should contribute equally to a rider's adjustment factor. This is in the realm of weighted statistics. Riders doing more climbs get a higher weighting factor, and climbs with more riders get a higher weighting factor. The climb weighting factor depends on the sums of the weighting factors of riders doing the climb, and the rider adjustment factor depends on the sum of the weights of the climbs the rider did. So this is another part of the iterative solution.But this tweak is unlikely to make a significant difference. The basic idea is as I described it.There's an alternative which was suggested by BikeTelemetry in comments on my last post on this topic. That would freeze scores for each week rather than re-evaluating them based on global ratings. That I haven't had time to test, but the code for the algorithm described here is basically done; just ironing out a few bugs.

San Francisco: City of Passive-Aggressive Losers

2011-11-09T08:16:00.000-08:00

The San Francisco mayor's election was yesterday, and it looks like Ed Lee won it with around 30% of eligible voters voting.

Quoting the San Francisco Examiner, referencing critics:

"...the career bureaucrat would be nothing more than a shill for powerful City Hall insiders. Lee also was dogged by accusations of voter manipulation by an independent expenditure committee that supported the mayor and other backers laundering campaign donations, which prompted a District Attorney’s Office investigation..."

He attracted a huge number of donations, driving up the amount the city needed to pay in public financing. His donations were largely from out-of-city donors, many laundered through low-income workers to circumvent the $500 donation limit. Then there were the nominally unaffiliated supporters, for example those who produced and distributed free copies of the book of his life story. Meanwhile, he violated the law by refusing the disclose details of public contacts within the required time limit. This was so obviously fee-for-service it couldn't have been any clearer.

There was massive fraud in the early voting. Housing managers in Chinatown were reportedly collecting the absentee ballots of tenants and filling them in en mass. Shady "voting booths" were set up where voters brought their absentee ballots for "assistance". "Helpers" were filmed filling ballots out for people, and in other cases stencils were handed out so voters could fill in Lee's slot without the risk of voting for any other candidate. No doubt these absentee ballots were procured with further "assistance"

Lee sat out most of the series of mayoral debates, instead choosing to go hang out at bars and get close to his people. His campaign slogan, "Gets it Done", couldn't have embraced mediocrity any more, mediocrity where the city desperately needs vision and leadership to get it back on track to fiscal responsibility. Indeed, even the deal he claims to have brokered himself to get high-income city employees to contribute to their pensions and health care, Proposition C, he himself is gutting by promising these same workers (SFFD and SFPD) a compensating pay raise. Not only does that cover the pension contribution, but actively increases the city's unfunded liability further by increasing pension payouts, which are proportional to salary. Oh pity the poor police officer, most of whom make well over $100k/year, and can retire on that with full health benefits at an age when public-sector workers are typically mid-way through their careers, wondering how they're going to be able to retire.

Lee supports Proposition B, which pays for road maintenance with debt. Why is debt needed? Because the budget for road work has been diverted to other things, like the massive city worker salary & benefits budget. Lee, "who gets it done", was the Public Works director from 2000 to 2005, the one most directly responsible for road infrastructure. He's part of the reason the roads are in such sorry condition, and a Proposition B is claimed to be needed.

Lee is a shill, a puppet, a tool of the money machine. It was clear as can be he had to go. Nobody I've interacted with admits to supporting him, and I even walked door-to-door in Mission Terrace as a David Chiu volunteer. Yet how can someone who is so clearly unpopular, so clearly corrupt, so clearly a tool of outside money, be elected?

Well, it doesn't help that the ten-thousand-member San Francisco Bike Coalition campaigned for him. SFBike endorses candidates based on a member survey on which members were explicitly asked to rank candidates based on responses to a lame curve-ball set of questions, and Lee finished 3rd. Yet surely the stories of corruption and fraud which followed that survey could have tempered the zeal with which SFBike daily bombarded the internet with messages encouraging a position for Lee on the 3-slot ranked-choice ballot? I even asked an officer of the coalition: had Lee made a racist comment, would you continue to support him? He admitted probably not. Yet naked corruption and fraud is okay?

And I'm also sure he got support from many of the 27 thousand city employees, 3% of the San Francisco population. The number of city employees has reportedly increased under Lee's brief tenure as mayor, and he's taken no steps toward making the efficiency improvements which are so desperately needed (efficiency means fewer workers and people working harder and learn new skills; none of these are popular with existing workers).

But the real culprit here is the lethargy of the voters. Turn-out was weak. It was reportedly very low even among protesters at Occupy San Francisco. Perhaps many of these protesters were from out-of-the-city, I don't know. But if you're going to smugly call for the downfall of the Man in the streets and then not exercise your legal obligation to vote for representative government, you're destroying the integrity of your own message. The reason the banks are able to get away with so much is voters aren't providing adequate oversight of their elected representatives. And there's no better example of that than yesterday's election. This is especially true because several candidates made an explicit point to support and defend the rights of free speech and public assembly of these same protesters.

San Franciscans will continue to whine and complain about how the city is in such a sorry state, how budgets are unsustainable, how public transit is little more than a fiscal tar pit, how there is no vision for how the city is to move forward. Yet many of those same felt they had more important things to do with their time than participate in yesterday's election. Those people are the definition of passive-aggressive, an affliction which is in epidemic proportions. These people deserve mediocrity, corruption, and insider deals. They deserve Ed Lee.

If you live in San Francisco and didn't vote, either yesterday or absentee, I extend my finger in your general direction. You, my friend, is what is wrong with this city.

Natural Selection Voting Theory

2011-11-08T06:18:00.000-08:00

In nature, if you can't do what it takes to survive, you die, your genes are eliminated from the pool, and someone else takes your place.

Maybe what takes your place is better, maybe not. But if not, it will also die, be eliminated, until eventually something able to do what it takes comes along and so, by this process, things generally improve over time.

This is my theory of voting. Rule #1: if the incumbent isn't doing a good job, vote them out.

So often in elections I hear about the "lesser of two evils". "I don't like the incumbent XXX, but he's better than YYY." Sorry: the rule of natural selection says I vote XXX out of office anyway. Maybe YYY is even worse. But then I vote YYY out at the first opportunity.

Eventually corrupt and unqualified candidates will stop running. Eventually you get someone good in office.

But if you vote "lesser of two evils", things will never change. You'll always have candidates who suck, just slightly less than the competition. We'll remain mired in the corrupt stagnation which we've had at all levels of government for as long as I remember.

So my first rule is if I don't like the way things are going, the incumbent doesn't get a vote. I pick from the alternatives.

New scoring scheme for Low-Key 2012?

2011-11-07T04:48:00.000-08:00

Low-Key scoring has gone through various phases.

In the 1990's, we scored based on fastest rider. The fastest man and the fastest women each week would score 100. Those slower would score based on the percentage of the fastest rider's score. This was super-simple, but when an exceptionally fast rider would show up, everyone else would score lower than normal. Additionally, this was frustrating for the fastest rider (typically Tracy Colwell among the men), since no matter how hard he or she pushed himself, the result would be the same 100 points.

So with Low-Key 2.0 in 2006, we switched to using the median rider (again treatng men and women separately). The median is much less sensitive to whether a particular individual shows up or not, so scores were now more stable. However, there was still an issue with women, and most especially with our hybrid-electric division, since smaller turnouts in these again made the score sensitive to who showed up.

So in 2010 I updated the system so now all riders were scored using a single median time, except instead of actual time, I used an "effective mens's time" using our history of Low-Key data to generate conversion factors from women's and hybrid electric's times to men's times. Mixed tandem's were scored by averaging a men's and a women's effective time.

This worked even better. Now if just a few women show, it's possible for them to all score over 100 points, as happened at Mix Canyon Road this past Saturday.

But the issue with Mix Canyon Road was because the climb is so challenging, and for many it was a longer than normal drive to reach, the turn-out among more endurance-oriented riders was relatively poor. The average rider at Mix would have scored over 100 points during, for example, Montebello (data here). It seems almost everyone who did both climbs had "a bad day" at Mix. That is far from the truth!

There is another scoring scheme I've been contemplating for many years. It's one which doesn't use a median time for each week, but rather compares the times of riders who did multiple weeks to come up with a relative time ratio for each climb. So if, for example, five riders did both Montebello and Mix, and if each one of them took exactly 10% longer to climb Mix, then a rider on Mix should score the same as a different rider on Montebello as long as the Mix rider's time was exactly 10% longer than the Montebello rider's time, once again after adjusting for whether the rider is a male, female, or hybrid-electric.

So why haven't I made this switch yet? It sounds good, right?

Well, for one it's more work for me. I'd need to code it. But that's not too bad because I know exactly what I need to do to make it work.

Another is it's harder to explain. It involves iterative solution, for example. I like things which are easy to explain. Median time is simple.

But another is it would mean scores for any week wouldn't be final until the entire series was complete. So a rider might celebrate scoring 100.01 points on Montebello, only to see that score drop to below 100 points later in the series. Why? Because the time conversion factor for a given climb would depend on how all riders did on that climb versus other climbs. And it's not as simple as I described: for example if rider A does climbs 1 and 2, and rider B does climbs 2 and 3, then that gives me valuable information about how climb 1 compares to climb 3. In effect I need to use every such connection to determine the conversion factor between these climbs.

But while scores might change for a climb, the ranking between riders during the climb would not. That's the most important thing. Finish faster than someone and you get a higher score. The conversion factor between men and women, for example, would stay the same. That's based on close to 10 years of data, so no need to continue to tweak that further.

I'll need to get to work on this and see if I can make progress. I'll describe my proposed algorithm next post.

Riding the Diabolical Duo at Mount Vaca

2011-11-06T07:01:00.000-08:00

approaching the Low-Key finish (Cara Coburn photo)

Yesterday I rode the Diabolical Duo at Mount Vaca.

First: Mix Canyon Road. Coordinator Barry Burr did an excellent job organizing this one, definitely the "road trip" ride for many in the 2011 Low-Key Hillclimb schedule. For my car pool it wasn't a big deal: one hour from San Francisco, even stopping for gas. Rides like Alba Road, Bonny Doon, Henry Coe, Jamison Creek Road, and Hicks Road we've done in the past are all substantially further, with plenty more of comparable distance. But most of our riders live closer to San Jose than to San Francisco, and for them the trip was further.

But even from San Jose this trip was worth it. A big part of it was our Strava event: The Diabolical Duo. The Low-Key Hillclimb covered just the first part of this: to complete the Duo, riders needed to also climb nearby Gates Canyon Road.

Inspiration for the Duo event came from The Toughest Ascent Blog. I won't even try to describe these roads: the blog already does an excellent job. All I'll say is they are seriously tough climbs. Take Mix: after a mellow start, it hits riders with a few steep pitches. Wow, that was tough, I thought, and then I hit the chalk "Hammer Gel Time" coordinator Barry had written on the road, marking the beginning of the "tough part". Tough? What had we just gone through?!?!

But what followed was truly impressive. When I hit the tight switchback from the Toughest Ascent Blog, I had to laugh. This was just insane!

But I got through that in my 34/27. Coming out of these switchbacks, you might expect some relief, but no luck: the road continues steeply until a following switchback. In the climb yesterday we had a photographer in that corner. Cool, I thought, the end of the steepness.

However, after turning that corner my hopes were crushed like a walnut shell under a cycling shoe. Not only was the steep stuff not over, but the road actually got even steeper still beyond this corner. This final part of the climb has been suppressed from lactic acid poisoning of the short-term memory centers of my brain. It wasn't even the raw grade numbers, but all that we had been through leading into those numbers.

Cara at the happy face marking the approaching finish (Lisa Penzel photo)

But like all climbs, this one finally ended. After the finish line for the Low-Key, Barry had set up a second line further up Blue Ridge Road, a gravel road which Mix Canyon intersects. This road was a bit wash-boarded but ridable even in my 20mm 182 gram sew-up tires. It took only a few minutes to hit this second line, the true summit of this climb. But I'd been wasted from my effort leading to the top of Mix, so I my pace here was far from impressive, especially given the distraction of the rough surface.

Returning to the Low-Key finish line at the top of Mix, I was a wasted shell of a human. I was still warm from climbing, but eventually the chilly 12C air penetrated, and I started to shiver. Fortunately I had warm clothes in the volunteer car, so was able to rectify this. I then got some fruit from the event refreshments. I was done.

Done, except that Ammon (with whom Cara and I had carpooled that morning) was waiting for me so we could do the second climb in the Duo, Gates Canyon Road. I couldn't disappoint him, I figured, I had to try.

So eventually we were off back down Mix. Ammon descends much faster than I can, especially with my flaky carbon Edge rims with their unsmooth braking surface which makes them prone to skidding, and my narrow tires pumped to 140 psi hardly inspired confidence. But eventually I made it down where Ammon and other riders were grouped. We set off together towards Gates Canyon, a few miles away. It was a nice ride along well-named Pleasants Valley Road.

Gates is interesting. It starts out almost flat, barely ascending from the valley. Then soon after the proper climbing begins, still not steep, one hits the road indicating the unpaved road. This is steep enough that grip is an issue. Ammon had zoomed ahead, and I was riding at that point with Low-Key regular James Porter, and while he was able to ride this, my tire skidded on the gravel. I could have let out some air pressure, but didn't want to risk damaging my rims if the pressure got too low, so just walked here. It was slow going. James was by now gone.

In sections, the road became just plain dirt. This had not been indicated on the Toughest Ascent Blog, so represented a recent change. I worried about clogging my Speedplay cleats, but they were fine.

But the dirt didn't last too long, and beyond it the pavement was very good, surprisingly good for what seemed to be a road to nowhere. But as soon as the pavement re-appeared, it bent into a disturbing, highly vertical angle. This road was even steeper than Mix had been! I was barely turning my 34/27, grinding away up the ferocious grade.

Many riders reported this had been tougher than Mix. Honestly I hadn't felt that way. Sure, it was steep, but I wasn't in nearly the same hurry, and I could focus on just getting up the thing rather than every second getting the most of out my legs. So here I was substantially less traumatized when I reached the end of the pavement, the end of the Strava segment. I then walked a bit on the gravel road which continued on to Mix Canyon Road, but just far enough to get a good view, then turned back.

As I got ready to descend, a local rider I'd passed arrived at the top, then soon after a group of Low-Keyers descended the dirt road I'd just walked down. The Low-Keyers had climbed to the high point of that road, almost to Mix, but had turned back. One had crashed and was bleeding. Just a flesh wound, though...

The local told us how he climbed Gates Canyon every Saturday, Mix Canyon every Sunday. He lived in Vacaville, he explained, and these were the local climbs. I was in awe. Deja-vu from Maui, where riders would climb 10 thousand foot Haleakala every week or two. No matter how extraordinary a climb, there are those for whom it becomes the ordinary.

We said goodbye to the local rider, descended together to Pleasants Valley Road. The descent wasn't bad. The fine gravel which I couldn't grip was fine descending, and the deeper dirt I could easily run. It was fun. At the bottom we then split up as we rode to our respective rides home.

Ammon, it turned out, had completed the root to Mix Canyon, and descended that instead of Gates. Very cool.

I didn't hear a single rider complain the longer-than normal drive hadn't been worth it. This day absolutely made the 2011 Low-Key series. If every one of the three remaining climbs is canceled from rain, I'd still say the series was a success. I'll never forget climbing these two roads.

Results of the Mix climb are posted here.

Low-Key Hillclimbs: over the hump

2011-10-31T20:55:00.000-07:00

The 2011 Low-Key Hillclimbs are over the hump, with 5 of the 9 scheduled events in the bag. Each one has had near-perfect weather, with warm sunshine without being hot. It's been supernatural, almost.

Week 1 is always stressful: after a long "off-seson", Low-Key returns to Montebello Road. I traditionally coordinate this one, more to take responsibility for the outcome rather than due to being qualified. Honestly, organization is not my strong point, and every year something gets overlooked. But I've had excellent assistance from Howard Kveck these past few years, and he helps keep things in shape when I stumble. Sometimes there's a bit of next-day revision needed on the results based on email feedback, but in the end we typically get them fairly good. This year things went even smoother than normal.

Week 2: a late swap with Barry Burr for week 6 (more on that) had me coordinating Sierra Road, as well. Biggest trick on Sierra Road is the start, which is in the suburbs. But nobody objected to our presence, and it was a great day on this road now made classic by the Tour of California stage race.

Week 3 and I was nervous again, not only for pulling off the climb of Page Mill Road in Palo Alto, but also because it was my first one I'd ride. I sort of recruited Janet LaFleur for this one after nobody volunteered from the Low-Key mailing list. And oh, my, did Janet come through! All the organizational skills I lack she has plenty. Everything came off with precision. It helped that we started riders in groups of around 15 rather than all together: the lower portion of Page Mill is too narrow, really, to send well over 100 riders at it in one pack. To top it off the day I really surprised myself with how well I rode. I knew my running had been going well: I was putting out good training runs after recovering from my August half-marathon. But would that translate to cycling fitness? For hillclimbing, apparently it did.

Week 4 and we tackled the first of the two brutally steep climbs in the series: Bohlman Road in Saratoga. Well, not really Bohlman, more precisely Bohlman-Nortan-Kittridge-Quickert-On Orbit-Bohlman. Again, we had a wonderful coordinator in James Porter, who like all excellent coordinators never gets flustered and never lets things get out of control. I was worried about the record turnout for this road: typically numbers fall off on the super-steep stuff. This time, that wasn't the case, and we maxed out our rider limit. Yet it turned out to be no problem, as residential density on these roads is low, and all of the drivers we encountered were amazingly patient. I think living on such steep roads they accept that high speeds aren't in their plan. So having to go around some cyclists isn't a big deal. Of course initially we filled the entire uphill lane, but soon after starting we turned onto the steep slopes of Norton Road and that strung things out very quickly. It was a great climb: I prefer it to alternate approaches up the mountain due to the steepness of the lower portion, the excellent pavement, and the very low car traffic. For now, at least, I think we'll stick with this version for future rides here.

Week 5 was a recovery week of sorts. This one had been Howard's suggestion: Palomares Road near Fremont and Union City. I had been sceptical: the climb wasn't steep enough long enough to break up groups, I feared, and so results would be challenging. But Howard is the results coordinator and it had his call to do this one, so that was his problem! To help, we used small groups (15 or so) of riders as we had at Page Mill. Unlike Page Mill, though, here the first group, of the self-assessed fastest riders, failed to break up. Pretty amazing: it's been too long since I've done a road race, and this had that feeling. Too afraid to pull, I was trapped in the vortex, at the mercy of the pace of those in the front: Tracy Colwell, a renewed and stronger Tim Clark, Nils Tikkanen, Jacob Berkman. Then I was freed from my trance by the 200 "paces" sign: the finish was near. Then it exploded, Tracy and Keith Szolusha off the front, the others scrambling for minor placings. I was fourth.

Amazingly the finish line crew managed to get most of us, and I went to Pat Parseghian's excellent finish video for the rest.

The other groups were less well matched, and the biggest sprint after ours was probably four riders. A few riders looked puzzled when we asked their numbers as they crossed the line (we avoid jersey numbers) until I realized we forgot to mention at the start that people shout their numbers at the finish, and people tend to be fairly brain dead at the end of a hard climb, even short ones.

It was another gorgeous day, and I was super-happy the series was going well. Three more weeks to go, then Thanksgiving @ Mount Hamilton, which is special.

But it's hard to look past this weekend: Mix Canyon Road. It stands to be by all accounts the hardest climb Low-Key has ever done. It's simply inhumane. But I'm the one who put it on the schedule, so no whining allowed from me...

I love it how this series magically comes together and works. Most people would say you couldn't do this sort of thing, but every year we do, and every week people have a good time.

Instant Runoff Election Simulation: Exhausted Ballots versus vote count in the 2011 San Francisco Mayor's race

2011-10-26T05:19:00.000-07:00

The 8 November election in San Francisco will have 16 candidates contesting the mayor's position. The city will, for the first time, use instant runoff voting for the city-wide mayor's election, avoiding the need for people to cast multiple ballots in the likely scenario that no candidate would get at least 50% of the votes.

Instant runoff works by creating virtual "rounds" of voting. Voters get a number of votes on their ballot. They list their first choice, then their second, then their third for the mayor's position. In principle there could be enough votes to rank all candidates (one less than the number of candidates), or more if you want room for write-in candidates. First, all candidates receiving at least one first-place vote are ranked, and the one (or multiple) candidates receiving the least number of votes are eliminated (assuming there's at least one candidate left). Ballots which had an eliminated candidate as a first choice have lower-ranked choices promoted until either the first-place vote is still in the race or the ballot contains no more candidates within the race. A ballot with no more candidates who have not yet been eliminated is considered "exhausted". The election continues until a single candidate has a majority of the votes from unexhausted ballots.

People have argued that exhausted ballots are a sign of the failure of the system. They claim that voters who submit exhausted ballots haven't had their voice heard. Of course this is false: the voters have had their voice heard but the candidates they supported failed to garner enough support. However, what an exhausted ballot implies is that among the final candidates in the race, that voter failed to have his preference counted. Not counting write-in candidates, for a voter to be guaranteed that his ballot will not become exhausted, he needs a number of votes equal to the number of candidates minus one. This allows specification of a preference between every pair of candidates, covering the every possible final round.

But San Francisco doesn't provide this many votes: not even close. Due, it is claimed, to limitations of the ballots used in city elections, only three votes are offered. I view this as a bit of a farce: of course it is possible to provide space for more than three choices, and there are plenty of examples of other nations which do so. Three is a long, long way from the fifteen choices which would be necessary to avoid ballot exhaustion.

I'll assume voters pick their honest top choices, no matter the perception of candidate viability. Of course this is an incorrect assumption, voters may prefer their ballot not get exhausted early, and so may want to pick at least one candidate considered likely to get a large number of votes. This "safety net" pick would likely be in the final spot on the ballot, since picking it earlier would likely render "long-shot" choices ranking below the pick irrelevant. But I'll assume here that there's no safety net strategy, and voters pick their preference every choice. I also assume voters vote for the full number of slots on their ballot: if they have six choices, for example, they don't vote for three and leave the final three slots blank.

In this case, exhausted ballots only make a potential difference if the winner, after all other candidates but one have been eliminated, fails to receive at least half of the total ballots submitted (not half of the remaining unexhausted ballots). If the winning candidate receives over 50% of all ballots, then even if all exhausted ballots would have had their next vote go to the second-place candidate, that second-place candidate still would have received fewer votes than the winner. And that's an extreme case: it would be virtually impossible that every one of those voters with exhausted ballots would have preferred the loser over the winner. So unless the number of exhausted ballots is sufficiently larger than the difference in votes received in the final round by two surviving candidates, it is very unlikely those exhausted ballots would have switched the result if the voters had had more picks.

So how many exhausted ballots is acceptable? I'll toss out a proposal that 1% is okay, assuming there is a "cost" associated with putting more choices on the ballot. But more than 1% and I think it's fair to say the reduced number of votes per ballot is seriously in danger of affecting the outcome.

If I want to estimate how many votes I need per ballot, with 16 mayoral candidates, to avoid at least a 1% exhaustion rate I need to make further assumptions. With 16 candidates, if one candidate is vastly more popular than the rest, then he'll get most of the first-place votes, and no ballot will be exhausted, since the election end on the first virtual round.

In the other extreme, if each candidate is equally popular, such that a voter chosen at random will have ranked candidates in essentially a random order (assuming candidate preferences are uncorrelated, which is clearly unrealistic), then the fraction of ballots which will be exhausted can be calculated fairly easily: suppose there are C candidates and V votes per ballot. Then the number of ways to vote or the C candidates with V votes = C! / (C ‒ V)! ("!" is the "factorial" operator), while the number of ways to vote without including the final two candidates = (C ‒ 2)! / (C ‒ V ‒ 2)!, assuming C ‒ V > 1. Therefore the probability a random ballot will exhauste among many random ballots =

[ (C ‒ 2)! (C ‒ V)! ]/ [ C! (C ‒ V ‒ 2)! ].

This can be simplified to the following, eliminating the factorials:

(C ‒ V) (C ‒ V ‒ 1) / [ C (C ‒ 1) ]

But this case is unrealistic. Some candidates are more popular than others: the difference in votes isn't just due to randomness. So I'll make an assumption somewhere between the two cases of one super-popular candidate and all candidates equally popular. I'll assume the most popular candidate gets 20% of the first place votes. Then the second candidate gets 20% of the remaining votes. Then the third candidate gets 20% of the votes remaining after votes have been assigned to the first and second candidates. Etc. So the most popular candidate gets 20% of the first-place votes. The second-place candidate then gets 16% of the votes. Third place then gets 12.8% of the votes. This goes on to the least popular candidate, who gets 20% of the votes which haven't been assigned yet to that point, which is 0.7% of the total. If a voter has given his vote to none of the candidates after this round (2.8% of them), I try again starting with the most popular candidate. For second-preference votes and beyond, I do the same game, except a voter can only vote or each candidate once. So less popular candidates have a better chance of getting lower-ranked votes than they have of getting first-place votes.

This vote distristribution is simplistic, obviously. In a real election there will be pairs of candidates who will be close to each other: the gaps between canddidates won't be so uniform as they are in this model. But that's not so important. What's important is that the votes tend to be clumped toward the head of the field, rather than distributed uniformly over all candidates.

So I ran 100 thousand randomized votes using this approach, and I compared it to the "worst-case" where each of the 16 candidates is equally popular. Here's a plot of the percentage of exhausted votes versus the number of votes per ballot.

The result is for the equally popular candidates you need 14 votes per ballot to keep the exhaustion rate down to 1%. This is only one away from the 15 needed to fully rank the 16 candidates. For the candidates with different popularity, you need 9 votes to get the exhaustion rate down to close to 1%. In each case allowing only 3 votes per ballot will result in high exhaustion rates: 65% and 25%. With only 3 votes per ballot, the number of votes is affecting the outcome: either a lot of ballots get exhausted or voters, to avoid ballot exhaustion, will deviate from their true preferences by engaging in the self-fulfilling prophecy of voting for whom they think are "viable" candidates.

And since everyone agrees interim Mayor Ed Lee is a viable candidate, I don't like where this leads.

So in summary: I really like ranked choice voting, but surely we can do better than this. With 16 candidates, I want at least 9 slots to rank candidates, and would prefer 15, but would even live with as small a number as 6. 3 is obviously way too few, however.

San Francisco Mayor's election votes

2011-10-20T18:55:00.000-07:00

I've been following the mayor's campaign as I've been able, and these look to be the candidates who will get my three votes:

First is David Chiu. I wrote about him yesterday, about his ride on SF2G. David's been on my virtual ballot all along, either first or second. I don't think we agree on much on the proposition ballot, to be honest. I take a fairly hard line on bonds, while he was a principal supporter behind Prop B (street maintenance bond). So I asked him about this directly at the Potrero Hill street festival, noting that my "undergraduate-level economics" tells me funding ongoing maintenance with debt is a bad idea. He agreed, but claimed our present situation is an exceptional emergency, and the bond is needed to avoid much higher costs down-stream. I still question the city's discipline to remedy the revenue imbalance when bonds are provided as a cop-out, but I respect his response. We clearly disagree on Prop D, public employees funding pensions. To me it's clearly superior to C, which Chiu supports. But as I noted I'll vote for both, because I don't want vote splitting to leave us with neither. But we agree the size of city government is a serious liability for San Francisco, especially given the city's virtual inability to prune out unproductive employees. What seals the deal with Chiu is our shared perspective on the critical role non-automotive transport plays in quality of life, that the best cities in the world are those that provide the best environment for pedestrians and cyclists as opposed to car drivers: New York, Boston, the great cities of Europe. To the contrary, cities which dedicate the most resources to cars (wide roads and large parking lots) are among the most unattractive: think Los Angeles, Atlanta, or Phoenix Arizona. And he was a serious proponent of bringing the 2013 America's Cup to San Francisco, despite the NIMBY forces who prefer that San Francisco be a bedroom community from where they can live their private lives as unencumbered by other people as possible.

Next is Herrera. Initially there was no way I was going to support him. As public attorney I felt he'd dragged his heels when the city's bike program was blocked by Bob Anderson's lawsuit claiming that the city should have filed an environmental impact report (EIR). I felt, as did at least one member of the Board of Supervisors, that a "safety case" could be presented for many if not all of the projects which would have exempted them from the EIR. But Herrera's view was that the EIR should have been filed from the beginning, that he'd advised one be filed, and since the bicycle program had neglected to do so there was no reason for his office to scramble to argue for projects one-by-one. Okay, so that's water under the bridge. I think Herrera clearly supports the bike program as it moves forward from here, the EIR done and approved. More importantly he seems an organized, businesslike guy who has what it takes to keep the city's business plan moving in the right direction. And he gets points for haven ridden with SF2G, which I unfortunately missed. Herrera is also endorsed by the Potrero Hill Boosters Club and the Potrero View newspapers, both representing my neighborhood.

And my #3 is Jeff Adachi. Like Herrera, Jeff also seems intelligent and business-like. He's the one behind Prop D, and has been a long-time proponent of getting the public employee pensions under control. I really admire his stand here, not out of some Republican-like vitriol against public workers, but because the public pensions are so clearly disproportionate to private sector and because they are so clearly unsustainable economically. Politicians are all too ready to push off liabilities onto the next generation and the public pensions are a clear example of this. Jeff also seems to take a reasonable stance on other issues and has handled himself very competently in the debates I've seen.

So that's it... on my top 3:

David Chiu
Dennis Herrera
Jeff Adachi

Filling out my top ten are Dufty, Avalos, Alioto-Pier, Rees, Ting, Hall, and Yee, roughly in that order.

Of these candidates Hall is an interesting case, clearly the most conservative of the bunch. For example, while most candidates are against Adachi's Prop D as either unenforceable or going to far, Hall argued it doesn't go nearly far enough. It could be argued, given our fiscal issues, he's the candidate we need even if he's not the candidate we want. But I'm not quite ready to go there, not just yet.

Another interesting candidate is Baum. I like a lot of her views, for example her stance against Prop B, but she's essentially a full-blown socialist running as a Green. Since the American Green party was an off-shoot of the Socialist Party, this isn't atypical. I'm a big supporter of the Green Party's environmental agenda but it's socialist agenda is naive. For example, Baum wants a city-run bank and a massive increase in public housing. I simply don't trust this city to efficiently and effectively provide high-value services. I can't think of any examples, none, where a major government in this country has done so in the past. I'm not sure if I'd vote for Baum or Lee in a head-to-head. Maybe I'd write in Tony Kelly instead. Lee's proven to be allegedly corrupt, sure, but Baum's programs would be a magnet for future corruption.

Avalos gets big points for his strong support of cycling in the city, and rode with SF2G. He also gets points for supporting the right of the Occupy San Francisco protesters to hold their protest (under Lee there have been open police raids against the protesters). But he also has a socialist tendancy which I don't think is consistent with the realities of city government.

So what's the forecast? This election has turned into an Ed Lee versus the world contest. Lee seems like a nice guy to most voters. He hasn't totally screwed up anything obvious. And he was even endorsed by the San Francisco Examiner as their #1 pick while also making it into the Chronicle's second-tier behind David Chiu. But I don't trust him. Soon after announcing his entry into a race he'd previously promised to not enter to avoid conflict of interest (he was appointed interim mayor when previous mayor Gavin Newsom moved up to state lieutenant governor), he quickly raised more money than all of the other candidates had, combined. Many of these donations were suspect, at the donation limit coming from lower-middle-class donors. For example employees of a shuttle bus company which had benefited from a questionable decision on shuttle parking at SFO which clearly benefited the company were later reported to have been coerced by management, an unveiled money laundering scheme. Lee eventually returned the donations but only well after it was obvious something was amiss. Additionally, Lee's office has failed to disclose city contracts to the ethics commission within the 5 day window required by law. According to Dennis Herrera, Lee had missed this deadline 67 times as of 4 October. So Lee stinks of allegid corruption. Despite this, he is strongly supported, including by the same San Francisco Examiner which is reporting these stores. In elections, especially local elections, the inertia of incumbency is very, very strong. Voter ignorance and and laziness is rampant. And while most potential voters sit out local elections, too many vote without adequate preparation, rubber-stamping incumbents based on a lack of obvious reason to vote against. That inertia, especially with such a diverse field splitting the remaining vote, will be very hard to beat.

However, I really hope it is beaten. Any one of the other 11 candidates (those invited to the KQED/League of Women Voters debate) are preferable to Lee's proven record of corruption.

Riding with David Chiu

2011-10-19T20:19:00.000-07:00

One or maybe two times per week I ride the 42 miles from home to work with SF2G, a group of long-distance bike commuters which began with Google employees but has branched out to employees of other high-tech companies on the San Francisco Bay peninsula. Really there's no way I would be able to tolerate my commute if I had to take the train back and forth five days per week (or drive even one day per week). And the nice thing about riding is after a good ride in I feel fantastic: full of energy and alert all day. It's great; if it wasn't that riding got me into work on the latish side, around 9:30 am, I'd do it more.

John Murphy's been an SF2G regular for longer than I have. This year he's had the audacity to ask three candidates for San Francisco mayor to join us at the start of the ride, typically 6:30 am in Ritual Roasters on Valencia Street. San Francisco is a city of around 800 thousand people; surely mayoral candidates have better things to do with their mornings than meet with 30 or so tech-heads. We're so lost in the noise it's a joke to even think they'd get up so early to show up in a Mission coffee house and ride their bikes directly opposite the direction to their work. John's obviously insane.

But the insane thing is that all three candidates he asked: John Avalos, Dennis Herrera, and finally David Chiu agreed. I unfortunately couldn't make Avalos and Herrera, but there was no way I was going to miss Chiu. And I didn't.

David Chiu is very cool because he doesn't own a car. He gets around the city by bus, by car share, and most of all by bike. That shared experience establishes a certain bond: a feeling of common perspective which gives the hope that things which are important to me will also be important to him. The car fosters a social view which is very self-obsessed, the "through the windshield" view where one is surrounded not by people but by machines, machines which are threatening your life, property, and mobility. They are machines which even hostile to you and should be feared, hated, and defeated. On the other hand when cyclists encounter each other there's no shell, no engine, There may still be competition but without hatred. There's still the threat from cars, but it's a shared threat, us against them, rather than us against each other. Cycling is a much, much more human mode of transport.

I got to Ritual right on time. As is customary with the ride, I brought my bike inside, greeted the early arrivals, and sat down. No David Chiu yet. Another David, a ride regular who works at Oracle near the half-way point of my commute, was commenting that there didn't seem to be any real issues in this campaign. No issues? I launched into a long speech about what I thought there were issues which were absolutely critical to the future of the city, namely the unsustainable increase in the size of city government (3% of those who live here work for the city, and once someone is hired it is very difficult to fire them), not to mention the transportation issues which affected each of us.

This rant was interrupted by the announcement that Mr. Chiu had arrived.

And there he was. He was clearly the best dressed one of those present, whose number had grown substantially since I'd sit down. The place was packed. There was some hesistation as Chiu wasn't sure what the plan was, but John told him there was no plan and David could just do what he wanted. So David presented a brief discussion of his philosophy, then invited questions.

Chiu addresses some of the crowd (more arrived later). Scott Crosby photo

Public transit was a popular target, in particular MUNI which runs the bus and intra-city rail service in San Francisco. MUNI is a mess, with a terrible on-time record, buses simply not going out when mechanical problems or driver absenteeism (around 15%, more or less) gets in the way, convoluted routes, and lumbering vehicles which are lucky to average faster than a moderate jogging pace across the city. San Francisco has clearly worse public transit than any city of its size or greater on the East coast, and probably worse than any city of its size or greater outside the United States. It's just terrible.

But Chiu understands that and while every candidate claims to put a priority on MUNI reform, Chiu seems to sincerely care about the matter. After all, he uses the bus himself. It's in his own self-interest.

At 7 sharp, a member of our crew announced it was time to leave. So we unstacked our bikes, headed outside, and rode off in the dim twilight.

Riding on his girlfriend's bike (his had been vandalized). Scott Crosby photo

John led the way. Normally SF2G takes one of two routes when riding along the flatter "Bayway" route. The traditional route is over Bernal Heights via the "Cortland Hurl", a short moderately steep climb typically taken at a caffeinated pace. The other route is the "express" route from Potrero Ave to Bayshore. This involves a wild ride through a pothole-strewn interchange with traffic merging from the right via a high-speed off-ramp. It's not for the timid.

But John wanted to take what should be the preferred route: Cesar Chavez to 3rd Street. It's also the most direct route to the 22nd Street Caltrain station. The sitting mayor, Ed Lee, had canned a community-agreed-upon plan to remove a vehicle lane from relatively uncongested Cesar Chaves to make room or bike lanes. Too much pressure from trucking companies seemed to be the reason, and given Lee's potentially shady record of alleged campaign contribution laundering, you can't help but wonder if there could have been something extra in the handshake which opened that meeting. Chavez is a mess. And John wanted to make sure Chiu saw that.

The ride was fun. Chiu rode well, and John and David Crosby, arguably the leader (if there's one) of the SF2G group, got in some quality discussion time. Most of the rest of us just hovered around, riding our bikes at the modest pace. David could have zipped along faster, I suspected. He seemed pretty comfortable.

Finally we reached 3rd Street, where our commute would take us south, while David's took him north. Everyone seemed in a good mood, as the ride had cracked our shells of cynicism about politicians, their motivations, and their sincerity. David certainly seemed like a good guy who wanted to do a good job for the city he lived in.

Position on San Francisco Ballot Propositions

2011-10-18T17:30:00.000-07:00

Here's how I plan to vote on the San Francisco ballot propositions:

Proposition A

This is a bond to repair schools. Sorry, school repairs should not be paid for out of bond debt. They're an ongoing maintenance cost, and for those you need to raise revenue. I'm against any bond measure which isn't an obvious short-term expense for a long-term benefit. Repairs don't meet that standard. A bond would just rob the school system of future funding as more of the budget goes to paying the costs of this bond.

Proposition B

This proposition would borrow $260M to pave the streets, among other things. Sure: road maintenance is an important investment, but this is absolutely the wrong way to fund it. The city simply needs to find funds to pave streets from its annual $6.6B budget. If you fund maintenance this year with debt, you just make it that much harder to balance the books in future years as you pay the cost of that debt, and I don't foresee it getting any easier to balance budgets in the future. This bond would make it harder for the city to come up with funds to pave roads in the future, as more revenue would be directed towards paying off the bond. This measure is extremely poor economics. So I vote no.

Propositions C and D

These are competing pension reform measures for city workers. Every endorsement I've seen has been for one, against the other. The problem is there's over 27 thousand city employees and I think it's safe to say a good number of these are going to be voting against both measures: of course they want to keep their generous pensions as they are. So if the rest of the voters split with anything approaching equal weight between these two measures, neither passes, and we're still in the ridiculously unsustainable morass we are at present. The only way to prevent this from happening is for enough voters to vote for both. So that's what I'll do, even if I prefer D, which requires those earning more to pay more towards their pensions than those earning less. So yes on both C and D.

Proposition E

This allows the board to over-rule voter initiatives. This is an important check, and I fully support it. The typical voter does not take the time to carefully analyze every proposition: it's way too easy for mistakes to get made. Way too often proposed legislation is misleadingly represented, fooling voters into supporting something which is very much against the interest both of the city and of the cause for which the voters thought they were supporting. It's important for a mechanism to be in place to correct these mistakes. I vote yes.

Proposition F

This proposition changes rules for political consultants, for example requiring them to file monthly instead of quarterly reports. Really this sort of micro-management doesn't belong on the public ballot. I vote no; we have representative government for a reason.

Proposition G

This proposition would increase the sales tax collected by the city by 0.5%. In principle I'm fine with this, except the tax is directed to "public safety", and I don't believe in directed taxes. For one thing they're misleading: directing one revenue source to a particular budget item seems like it should increase funding for that item, but in reality it allows the city to reduce the allocation of funds from other sources. But even if the targeted funding isn't circumvented, it prevents the city from optimizing its resource allocation based on the present need. So I'll be all for a 0.5% sales tax increase if it doesn't come with the extra targeting baggage. So there's two possibilities: either the funds indirectly and misleadingly end up in the general fund, or they're tied up to a particular cause and the city is no longer able to budget efficiently. Either case is a lose. So I vote no on G.

Proposition H

This proposition is simple: it would make proximity a priority in allocating students to schools. Opponents say some schools suck, and it's unfair that local students get stuck with sucky schools. Well, the reality is some students, whether they be local or not, are going to get stuck with those sucky schools. The solution is to make the schools not suck, not sustain the illusion of "school choice" while essentially holding an annual Russian roulette of randomized school assignments. Given an allocation of students, I'd rather students go to school closer to home to reduce transportation costs. If students can walk or bike instead of get driven or take the bus, that's a substantial benefit. Opponents claim the proposition is "flawed", but I think this is just a tactic, as it's very simply written. I vote yes on H.

Most elections I have at least a few propositions where the decision seems tough, but this year I've not seen any compelling cases made to challenge what were my early preferences on any of these.

2011 San Francisco's Mayor Election, Ranked Choice, and Exhausted Ballots

2011-10-16T07:28:00.000-07:00

In the 2011 San Francisco Mayor's Election (election day is this November, but early voting has already begun), there are 12 "major" candidates... from the LOWV debate:

Adachi
Alioto-Pier
Avalos
Baum
Chiu
Dufty
Hall
Herrera
Lee
Rees
Ting
Yee

There are four others running: Ascarrunz, Currier, Lawrence, and Pang, but I think it's safe to say none of these four candidates is in the running, given their lack of representation in the debates so far.

San Francisco switched this mayor's election to Instant Runoff Voting. The way this works is each virtual "round" the candidate receiving the least non-zero number of first-place votes is eliminated, along implicitly with those receiving no first-place votes. Votes lower than first place on the ballots of those who voted for this eliminated candidate are then promoted until either there are no votes left on the ballot (the ballot is "exhausted") or the new first place vote on the ballot is still in the race. The idea is to allow voters to vote their preference, rather than to worry about which candidates are most viable. I really like this system, as long as the number of exhausted ballots remains relative small.

The problem is if a voter ends up with an exhausted ballot, then in the final "round" that voter's preference is no longer being heard. If the number of votes per voter is no more than one less than the number of "legitimate" candidates, then no ballot becomes exhausted. However, fewer votes, and on some ballots the final two candidates remaining may be the two candidates omitted from a ballot and that ballot will be exhausted before this final selection. These voters tend to feel bitter that their preference among these candidates is being ignored.

In the San Francisco election, with 16 candidates on the ballot, there are three slots available for voters to select their preference. The election commission argues this is due to a technological limitation of the voting process here. Errr.... I simply refuse to accept this, especially in the presence of many counter-examples throughout the world, for example Ireland, where voters get more selections.

The result is in the San Francisco election, voters are forced to engage, at least a little, in game theory. Since they don't want their ballots to become exhausted before the final round, they feel compelled to vote for at least one candidate who is likely to make the final round. I think pretty much everyone agrees that Lee will likely make it there, as Lee is the sitting mayor; he was appointed by the Board of Supervisors to take over when Gavin Newsom took the seat of the state Attorney General. So suppose a lot of voters wouldn't place Lee in their top three, but they prefer Lee to some other candidate who might have a long-shot chance to win, so they would be tempted to put Lee in their #3 vote as a safety net in case their first two choices don't make it.

So maybe every single voter has Lee as their mid-rank choice, say 6th or 7th (of the 12 candidates with "a chance"). But a majority of these voters put Lee in the #3 slot because doing so essentially blocks out the 6 or 7 candidates they really don't want to see in office. It's a good strategic move, they may feel.

The issue is now there's only two slots left for their true preferences. These two spots will get split between the 5 or so candidates they'd all prefer to see than Lee. Many of these voters will thus have their top two votes eliminated, in which case Lee gets their top vote. This gives Lee the election. Add in those voters who out of inertia and laziness will just vote for the person in office at the time and it becomes exceptionally challenging for anyone to beat Lee. This puts Lee in a very comfortable position.

There have been major questions about Lee, most recently with campaign contributions getting laundered through low-income voters to avoid campaign contribution limits, but also because he oversaw a police crack-down on cyclists in which cyclists were ticketed for trivial offenses (or non-offenses) like not "putting a foot down" at stops or riding during the day without reflectors, neither of which is illegal. Previously he canned an agreed-upon lane reduction on Cesar Chavez to improve cyclist safety after lobbying from the teamsters (although an alternate plan removing parking is moving forward). I really don't trust him to be the city's mayor at a time when hard decisions about city budgets, decisions which are going to cost people money, need to be made. We can't have the mayor's office going to the highest bidder.

So I really hope the city changes its ranked-choice voter scheme. The number of options should depend on the number of candidates. For example, during the District 10 Board of Supervisor's election, with 21 candidates and no incumbent, I proposed the following:

choices = floor[sqrt(N)] + 1

This formula would yield 1 choice for no candidates, two for up to 3 candidates, 3 for up to 8 candidates, 4 for up to 15 candidates, and 5 for up to 24 candidates. In the San Francisco Mayor's race, as with that Board of Supervisor's election, there would thus be 5 choices available. I feel this formula is the absolute lower-limit of what I'd want to see given Lee's presence in the race. For example, the following might be even better:

choices = floor[N^2/3] + 1

This revised formula would yield one choice for no candidates, two for up to 2 candidates, three for up to 5 candidates, four for up to 8 candidates, five for up to 11 candidates. six for up to 14 candidates, seven for up to 18 candidates, and eight for up to 22 candidates. Thus we'd get seven votes for this election. I'd have no problem ranking seven candidates here.

The idea in these formulas is that all candidates aren't equally likely to get votes. In any election, some candidates are going to be more popular than others, so you don't need the number of spots on the ballot to be equal to the number of candidates or even one less, unless there is a very small number of candidates. In a 16-candidate field, the huge majority of voters will have one of the two final candidates among their seven top choices; the number of exhausted ballots need not be zero, just less than the number which could have swung the final result with reasonable probability. But fixing the number at three is clearly woefully inadequate to the democratic process. Voting should be about expressing real preferences, not engaging in game theory.

Climbing Haleakala

2011-10-12T20:53:00.000-07:00

When Cara said she wanted to go on a vacation I asked her where she wanted to go. Hawaii, she suggested.

Hawaii. I'd only been there once: to Oahu for a conference in Waikiki. Nice riding on Oahu, but I knew we'd not be going to Oahu this time. There was only one island which could be on this agenda.

Maui.

Why? One word: Haleakala.

Ever since reading the description in John Summerson's book, I knew I had to go. In the rankings, only three climbs in the United States have a rating exceeding 6: Mount Washington Auto Road in New Hampshire (6.45), Mauna Kea on the epinonymous island of Hawaii (6.33), and Haleakala (6.13). Of these, two have bike races. Of these two, only Haleakala is open to general bike traffic. Mauna Kea (and nearby Mauna Loa, rated 5.26) aren't paved to their summits. So of the four, the clear choice is Haleakala.

There are several ways to approach Haleakala. After all, the entire eastern half of Maui sits on the mountain, which rises from deep in the Pacific. At its historical height, Haleakala rose to 13 thousand feet above sea level, but with erosion it now just barely pokes it's head above 10 thousand: 10020 is the official peak elevation. I've ridden over 10 thousand feet before: in Colorado, in the Eastern Sierra. But Haleakala is different than these other climbs in that the climbing is essentially continuous from the beach. Paia, where August's Race to the Sun begins, is a small town abutting the ocean. Like the primordial fish, you could virtually ride your bike straight out of the Pacific and climb straight up the mountain.

I didn't do this. I was starting at the Paia Inn, which is next to Maui Cyclery, the excellent bike shop where the Race to the Sun begins. Elevation was probably 20 feet above the actual ocean, so if the ocean was at sea level (which it typically isn't), then that would be an even 10 thousand feet of climbing. My previous record climb was probably Horseshoe Meadow in the Eastern Sierra, which gains 6200 feet. My longest climb so far this year was the north side of Mount Diablo, which gains 3600 feet. So this climb would be longer than the two, combined, without a rest in between.

I was scared.

I asked for advice to Don, who runs the shop. "Keep pedaling," he said. That was it.

The locals typically do the climb with two bottles, refilling twice along the way. But I wasn't confident of my ability to quickly identify the water locations, and didn't want to have to negotiate with the sub-sized pockets on my Voler Team Roaring Mouse jersey over what precious few clothing items I'd bring for the potentially chilly descent. So I went conservative and in addition to my two bottles I brought my CamelBak Mule. This certainly wasn't the weight weenie's choice but it did eliminate any concerns over hydration or layering.

I really couldn't wrap my head around 10 thousand vertical feet, but I could handle the climb of Baldwin to Makewao. So after a protracted preparation of my Ritchey Breakaway, I set off. There was a block head wind up Baldwin, but I didn't care. This was all about settling in for the long haul. The grade barely touched 7% on this long, straight road, and was usually less.

Starting the climb (Cara Coburn)

Baldwin is somewhat spoiled by all the car traffic between Paia and Makewao. But there's no intersections, so no conflicts, and no hassles from the drivers. The miles passed one by one, and when seven of them did so, I reached the stop sign marking Makewao.

Car drivers would probably turn right here, but the preferred route for cyclists, and the route taken in the race, is to continue. I crossed this stop sign without issue. Here Baldwin turned into Olinda, and got steep, the steepest part of the whole climb. But with the climb barely 1/8th done, my thoughts were up the road. My Garmin grade indicator was hovering around 12%, but I just pedaled along.

Going the other direction I started seeing groups of riders descending the hill who were clearly part of the commercial operations which haul people up the hill in buses and let them ride back. Since they barely need to pedal the whole way, these riders were on comfy cruiser bikes. On their heads were full-face helmets. These groups used to descend from the summit, but after more than one fatality, the National Park Service put an end to all commercial bike operations in the park. The issue is the Park covers only the final 3 thousand feet of the climb, so instead these services take clients up to the summit, the clients take their quota of digital photos, then they get back into the buses and drive to just below the park boundary, descending from there.

There was a rest stop for the descenders at the one navigational challenge in the whole climb, the turn onto Hanamu Road. There's a sign mounted low on the street sign there which says "Haleakala", pointing the way to go, but since it's mounted so low it's easy to miss, and the road is relatively inconspicuous. But with the riders there, I had no problem identifying the road.

This is the one non-climbing portion of the route. Hanamu descends a bit, then rolls a bit to its end. I was a bit confused by the left at Kealaloa Ave, which I hadn't expected, before the second left shortly after onto Highway 377, which I had expected. But the way here was facilitated by the blue arrows painted on the opposite side of the road to help the descenders find their way.

377 is a wider, more heavily trafficked road which goes into the Maui Upcountry. Still, though, it's very nice riding, with a wide shoulder. Climbing steadily, I reached the 3000 foot elevation sign. Progress. But I'd decided I wasn't going to think too much about progress until I'd reached 4 thousand feet, which would mark the longest climb I'd done in 2011.

Soon after I passed paint on the road: "Feed Zone". This was clearly the first of likely two feed zones for Race to the Sun. It would be the perfect place for two fresh bottles did I not have the Camelbak.

The next and final turn of the route was the left onto Highway 378, Crater Road. As I expected, this was very well marked, with a sign identifying it as the route to the summit. There's no missing this left.

The next sign I saw was a bit sobering: "Mile 0". Crater Road is 22 miles of continuous climbing, and all 22 of those were ahead. This climb ranked up there in total climbing with anything I'd done, and I'd already been riding, and climbing, for well over an hour.

There's a few cattle grates on the road to the top. At the first, I saw some kids on skateboards hanging out at the side of the road. I wasn't sure what to make of that: were they hitch-hiking to the summit for a ride down the hill? That would be fairly extreme. Whatever they were doing there, they expressed approval at my crossing technique, which was to speed up a bit going into the cattle grate then move my weight back to get it off the front wheel, letting off the pedals just as my rear wheels were going to hit it. But these were fine grates, no issues.

I was looking forward to the 4000 foot sign because that would be confirmation this had been my longest climb this year. And soon enough there it was. Normally I'd feel better about having climbed 4000 vertical feet, but here it was just permission to begin thinking about how much I had left.

Next to the 8 mile sign I saw a man and a woman standing next to their tandem at the side of the road. "Only 14 miles to go!" I shouted enthusiastically and without irony. Close to 22 miles into the climb, I felt the end was near. "Only?!?" he responded.

The miles ticked by and, every 500 vertical feet, an altitude sign marked my climbing progress. These were occasionally augmented by road paint presumably for the racers. A few times there was even a message to "Breath" in blue paint. At 6000 vertical feet, I knew I was approaching not only my longest climb of the year, but my longest ever.... and at 6500 vertical that sealed the deal (okay, there is Climb to Kaiser, but that's really multiple climbs covering its over 8000 foot elevation difference from start to the turn-around).

Soon after I saw one of the "Descend Haleakala on a Bike" vans parked by the side of the road. A group of motorcycle-helmeted riders was lined up, straddling their cruiser bikes. Someone, presumably working for the company, was checking to make sure they could coast a few meters down the hill before launching them on their descent of close to 7000 vertical feet. I knew I must be close to the park entrance.

And there it was. Three cars were lined up at a ranger kiosk ready to pay their fees. I figured maybe I could cruise by, look as if I didn't need to pay, maybe try some of that Obi Wan Kenobi "weak minds" mind control, and they'd let me through. No such luck... they called me over.

I got off my bike and walked in my cleats to the kiosk, rudely cutting in front of the lead car. I hoped to get this done quickly. However, perhaps I could have done so if I'd just tossed $5 at them and gone on my way, but since I'd be coming up here again the next day with Cara, and since we wanted to visit the park again later in the week, a pass made more sense. So I instead paid $25 for the annual pass. This took several minutes, during which I couldn't resist expressing my feelings about cyclists being asked to pay 3 times as much, per person, as an SUV with 6 people. It just seemed yet another example of the park services catering to car drivers.

Finally I had my pass and, after apologizing to the hidden occupants of the vehicles I'd cut off, I continued on the hill. The short rest coupled with little adrenalin boost from the conversation with the rangers really helped me here, and the accumulated fatigue from the long climb to this point seemed to have disappeared. Less than Mount Diablo remained, and I knew I could climb Mount Diablo.

7000 feet ticked by soon after. No paint on the road for this one: the parks don't allow road markings by event organizers, as I learned the hard way from having missed a turn at Climb to Kaiser my first time there.

As I climbed, I admired not only the incredible view of the distant shore, but also how the vegetation had so drastically changed. The trees had disappeared, replaced with bushes. And as I climbed further even these became more sparse.

I saw another altitude sign. Which was that? I thought maybe 8000, but looking at my time, surely I'd gone further.

And indeed I had: not long after the observation building came into view, and past that the observatory. It was a decent distance, but not 2000 vertical feet. I was in the end game.

I tried to ramp up the effort a bit here but at close to 10 thousand feet this is fairly futile. There's just not enough O₂ available. So instead I continued on with my endurance mode.

The visitor's center appeared on the left, but the observation building was still ahead. Obviously the Strava climb didn't end at the visitor's center, so I continued. The road here became substantially steeper: I watched as my Garmin grade display went up to 12%... 13%... 14% before retreating again. It felt around 12%, but that wasn't a problem with my 36/26 low gear. But the clock was ticking.

As I hit the parking lot of the observation building I felt a brief wave of euphoria. I'd done it! But not quite yet, as a paved pedestrian path still went up to the building itself. I was going to take no chances with the Strava segment. I steered my way around the pedestrians and rode up the steep path. Now I was truly done. A glance at my lap timer showed I'd come in just under 4 hours: 3:53:11 by Strava's reckoning.

Not surprisingly, the tourists at the observation building were impressed with what I had done. A few even wanted to take my photo. I just passed on the advice I had been given: "keep pedaling". It works.

After admiring the view, and sending text messages to Cara, I began my descent. Not much to say about this, although the tailwind assist I got on Baldwin for the final miles into Paia and back to Maui Cyclery was fun.

Jerry Brown's Broken Logic on SB910

2011-10-09T06:18:00.000-07:00

I want to revisit Jerry Brown's pocket veto of SB910, Lowenthal's bill which would require a 3-foot passing margin when passing cyclists while driving more than 15 mph, and would further allow drivers to legally cross the double yellow line when doing so and when line-of-sight allowed:

Here's Brown's explanation:

On streets with a speed limit of 35 or 45 mph, slowing to 15 mph to pass a cyclist could cause rear-end collisions. On the other hand, a cyclist riding near 15 mph could cause a long line of vehicles behind the cyclist.

This conclusion is exactly correct if the bill required a 3-foot passing distance and that the driver go no more than 15 mph. For example, it was proposed cyclists on the Golden Gate Bridge be restricted to riding no more than 5 mph when passing pedestrians. It would then become illegal to pass a pedestrian going at or in excess of 5 mph.

But that's obviously not what the bill requires. In fact, nobody giving the bill more than the briefest reading, nobody who'd read the emails I sent on the matter, nobody who read any of the descriptions of this bill, nobody who'd seen any of the advocacy for this bill could be under such a mistaken conclusion. Here's the text:

(b) A driver of a motor vehicle shall not overtake or pass a bicycle proceeding in the same direction on a highway at a distance of less than three feet between any part of the motor vehicle and any part of the bicycle or its operator, except that the driver may pass the overtaken bicycle with due care at a distance of less than three feet at a speed not greater than 15 miles per hour,

Sorry, there's no way you get into Yale Law School, now or at any time in the distant pass, with such woefully poor reading comprehension. So I conclude if Brown misunderstood the bill he didn't read it. Nor did he speak with Senator Lowenthal. Nor did he speak with any representatives of the California Bicycle Coaltion. Nor did he speak with any educated advocate for the bill.

Perhaps he received all of his information of the bill from representatives of the CHP, an organization for which I have only contempt. The CHP represents the interest of local police on the matter and I have seen over and over and over again how police crack-downs of cyclists have misrepresented the law (I have had consecutive three tickets thrown out by the court), and how they almost always focus on the least reckless of offenders since they are easiest to catch (people slowly riding across T-intersections with stop signs, for example, or riding on short sections of sidewalk). The police willfully distort, misrepresent, and misapply the law every day. They can't be trusted to interpret either present law or proposed legislation. And we don't trust them: at least under the Constitutions of both the state and the nation there's all sorts of checks against police abuse of power.

Initially I concluded Brown must have meant something else. For example, that drivers who failed to have at least three feet of clearance would slow down to 15 mph in order to not require that three feet of clearance. But existing law already says:

21750. The driver of a vehicle overtaking another vehicle or a bicycle proceeding in the same direction shall pass to the left at a safe distance without interfering with the safe operation of the overtaken vehicle or bicycle, subject to the limitations and exceptions hereinafter stated.

Meanwhile the driver's handbook (big PDF) says:

When passing a bicyclist in the travel lane ensure enough width for the bicyclist, typically 3 feet.

This is pretty weak language in a lot of ways, but it at least establishes a precedent within state government that 3 feet is a threshold for a safe passing distance.

So consider a driver going 45 mph coming up on a cyclist. There's no room to pass with a 3-foot margin: either it's a one-lane road (rare) or it's a 2-lane road with narrow lanes and either oncoming traffic or poor line of sight which doesn't allow the driver to cross the center line. So he now has the following options:

If the cyclist is going faster than 15 mph, slow and wait for an opportunity to pass with a three-foot margin.
If the cyclist is going less than 15 mph, slow to 15 mph and pass the cyclist.
Don't slow, instead passing the cyclist with less than a 3-foot margin

For Brown's argument to hold water, the final of these options must already be legal and acceptable. Yet recall the old definition of a yard, which is three feet: the length of the King's arm. Three feet means the cyclist could reach out and almost or barely touch the passing car. That's already close. Passing with less than this, as implied by the driver's handbook, can hardly be considered safe by a reasonable person.

So the scenario of a car being rear-ended, if true, already exists under the law as written and interpreted by reasonable people. There's no additional hazard presented by Lowenthal's bill. Indeed, since it allows drivers to cross a double yellow when passing cyclists (something essentially all drivers already practice) it makes it easier, not harder for drivers to pass. What if there were an obstruction in the road? What if there were a pedestrian in the street? What if there's a slow-moving farm vehicle? On any road so narrow there isn't a 3-foot margin to pass a cyclist, with insufficient line of sight to avoid a rear-end collision with a slowing vehicle, it should be a violation of the fundamental speed law to be driving at 45 mph. That's simply an unsafe speed.

The governor's "long line of vehicles" argument is flawed in another way: it is already required by CVC 21656 that a slow moving traffic if there are five or more vehicles waiting to pass. So there is already a mechanism in the law to prevent this from occurring.

So either Brown didn't understand the bill, or he doesn't understand existing law (and he was attorney general from 2007 through 2011), or he's lying about his reasons for opposing this bill. None of these options do anything but destroy my previously high regard for his competence to govern.So what next? Since it appears he hasn't read the bill, no reason to change any language. Repass it, send it back to his desk, and demand he either sign it or veto it. Don't let this passive-agressive pocket veto stand.

Jerry Brown vetoes SB223, pockets SB910

2011-10-07T22:27:00.000-07:00

First Jerry Brown vetoed Sen. Mark Leno's SB 223, a bill which would have allowed cities to restore the vehicle license fee which Governor Schwarzenegger had eliminated during his administration. Whether or not you think cities tax too much or not, vehicle fees are fair and rational, because they help reimburse the city for infrastructure supporting motor vehicles. Further they have a negative impact on congestion, promote public safety by getting cars off the road, and promote demand for public transit which helps promote denser schedules. The bill didn't establish a fee, it merely gave cities the right to impose them.

But SB910, the Lowenthal's bill to mandate a 3-foot passing zone when a motor vehicle passes a cyclist while driving at least 15 mph, seemed different. There was no reasonable argument against the bill, which had early flaws but which was, after several iterations, pounded into excellent shape. Amazingly, despite the opposition of the Republican minority and even some Democrats, passed the legislature. Surely Brown, the popular Democratic governor, would pass it. But it sat on his desk. And sat. And sat.

And he didn't sign it. Instead of the bill, he signed a letter to the legislator explaining why he'd pocket the bill, allowing it to expire.

The letter is here.

I'm not going to waste any time transcribing the bitmapped letter here. The logic is so broken, so flawed, that it amazes me the man could have possibly gotten into Yale Law School. His letter follows twisted reasoning which ignores the evidence from over a dozen states already with such legislation, and of much of Europe (Germany, rural France, Spain) which doesn't have a 3-foot passing rule, but a 1.5 meter rule which is close to 5 feet. Despite receiving 1500 letters in support, including from Lance Armstrong who perhaps has some experience with riding a bike, he instead chose to follow the brilliant legislative analysis of the California Highway Patrol.

The Highway Patrol is paid to enforce the law, not create the law, not analyze the law. They're good at chasing down fugitives on the highway. That Brown, the attorney general of the state from 2007 until early this year, would follow the advice of the CHP leads me conclude one of the following:

He's senile
He's corrupt

In either case, while I was formerly a Jerry Brown opponent, I can no longer support his presence in political office. Seriously, could Meg Whitman have been much worse? Perhaps, but certainly not on the issue of cyclist rights.

The whole thing disgusts me. Maybe the bill can make a recovery next year, But it's unlikely. More likely we'll need to wait until a dozen or two dozen more states pass similar bills, by which point the tide will carry us in the right direction. But any notion of California being a national leader in cyclists rights is hopelessly lost at this point, despite the fact the state has the best conditions for cycling in the country, and the best roads to do it on.

San Francisco Proposition B

2011-10-06T11:06:00.000-07:00

The San Francisco Bike Coalition has been flooding Facebook and Twitter with posts supporting Proposaition B in San Francisco, a bill which would use a $260 million bond to pave roads. The population of San Franciso is 800 thousand, so that works out to a around $300 per person in the city.

Now if you got a bill in the mail for $300 so the city could upgrade its roads, on top of the city's rapidly growing budget presently at $6.6 billion (according to last night's mayor candidate debate at UCSF), you might be disturbed. It's extremely unlikely if there was a proposition asking each citizen of the city to pay $300 for roads that it would pass. You can even pro-rate it as you will, make those with higher income, more property, even bigger cars pay more and you'd rightly ask: "but I already pay taxes for a city spending money at an all-time record rate, even adjusting for inflation, why should I pay more for such a basic service?" The proposition would crumble and fail.

But bonds are different. People face them with glassy-eyed innumeracy, ignoring the number of digits, and instead evaluate them on a benefit-benefit basis. Paved roads? Sure! I'm for that!

Proponants argue that even with the interest associated with bond debt, the "interest" on potholes is higher. Leave a pothole in place for unit time dt, and there will be a cost of K dt. However, if you need to borrow L at an interest rate r, they argue, the payment on the debt, r L dt, is still less than K dt. So it's a good investment.

But the problem with this logic is it assumes you're in a position to pay back the debt. The interest rate is only the immediate cost of a bond. Bonds also contribute to the debt load of the city, and if the debt load gets too high, the bond rating for the city will drop and there will be a huge cost increment. This second-order effect can have an enormous impact on the true cost of a bond.

It would be as if I took out a loan to brush my teeth each morning. Sure, not brushing my teeth will result in much higher dental treatment costs, but the choice isn't between borrowing and not brushing. The solution is to get a job and pay for toothpaste out of my salary. If I were to rely on debt for my daily dental care, eventually the collection agency would come and reclaim the gold from my mouth (perhaps my best financical investment lately, but I digress...)

San Francisco needs the discipline to pay its bills. It's exponentially increasing budget, even discounted but whatever reasonable inflation rate you choose, is clearly unsustainable. Bonds are clearly not the answer. San Francisco Bike Coalition is dreadfully off-base on this. This is especially true because cyclists cause so much less road damage than cars and trucks, yet with a bond cyclists are footing as much, or in the case of commercial truck operators, more of the bill to fix it.

Improving on Strava's Segment Timing for Low-Key Hillclimb?

2011-10-04T20:21:00.000-07:00

Strava, as I've written many times, has been a real paradigm shift in cycling in the San Francisco Bay area. Riders compete for rankings on "segments", typically climbs, of various degrees of obscurity. As we've now entered the Low-Key Hillclimb season, I've gotten a striking metric of how much of a market share Strava is grabbing here. More than a third of the 121 riders in week one's Montebello Road climb uploaded their data to Strava.

This brings up the possibility of using Strava for timing. Well, as I've described before, Strava's timing is far less reliable than a good-old-fashioned stopwatch. However, I have been batting around the idea of doing something for the 2012 series.

The issue is some climbs require permits for events: in particular state parks and open space. "Event" is always clearly defined, and it's always a dance to have that definition include things like the Low-Key Hillclimbs but not include groups of friends riding a hill together. Permits aren't an issue if they are willingly issued. The problem is the state parks have a very low threshold for inconveniencing auto traffic, their primary constituency, while the open space just flat out denies permits to any event where riders are timed. Yet there's no restriction against informal groups of riders timing themselves. That's fine. Strava, for example, which reports times of riders who climb a hill is not an "event".

For the full history of the Low-Keys, it's been suggested we allow riders to self-time on these climbs and report their results. But self-timing is a burden. You need to have a watch, remember to start it at the precise time, then remember to stop it precisely at the finish. If riders are late to stop the timer, there's a tendency to overestimate how much time passed since the finish, and underreport actual time. We want Low-Key to be an accurate archive of times. So relying fully on self-reported times is something we've avoided.

But now GPS is becoming virtually ubiquitous. Using GPS to time riders is now becoming, for the first time, a feasible option for climbs for which organized timing is prohibited. Sure, we'd need to let riders without GPS self-time, but most riders can just mindlessly attack the hill and upload their data.

The problem is the timing: Strava's timing algorithm is far from optimal. The reason is Strava uses course points. There's a start point for a course, a finish point, and points along the way. How many actual races have a start point and finish point? None I know of. They all use lines for the start and finish.

Now a point of clarification: in geometry, lines are infinite, while "line segments" are generally finite. So when I say "line" here I actually mean "line segment". I avoid the term "segment" because with Strava, "segment" means something different: it's used to describe "courses". So I'll stick with "line".

So to improve Strava's timing, it's been proposed on the Strava support forums that Strava adapt a "start line" and "finish line" model for timing. A start line is defined by a center point and a vector to one of the end-points of the line. A finish point is similarly defined. The choice of the end-point for each implicitly defines the direction through which the line should be crossed (defined end point must be to the right). So timing is then the minimum time elapsed between an interpolated crossing of the start line in the correct direction and an interpolated crossing of the finish line in the correct direction. For example, suppose I have the following start line and finish line crossings, in the appropriate directions:

SSSFSFFFSSFSS

You can see there are three timing events here, three instances of a start crossing immediately followed by a finish crossing.

This may be all I need to define a course. Some routes, like Old La Honda Road, have no reasonable short-cuts and therefore if you cross the start line at the bridge and then the finish line at the stop sign, the best times will all be from riders who stayed on Old La Honda Road. But many potential climbs have potential short-cuts.

Here's where an option for a "gates" comes in. Gates can be defined by additional lines. So the ride must cross, in order, the start line, the first gate, the second gate, etc, until finally crossing the finish line. Each of these gate crossings would need to be in the appropriate direction.

It's possible a rider will cross a gate more than once, or even backtrack and repeat multiple gate crossings, but as long as the rider crosses the start line, the gates in order, then the finish the course can be considered to have been completed.

So if I am going to define a segment for a complex route like Lomas Cantadas via Alta Vista out of Orinda in the Berkeley Hills of California, I'd define multple gates to make sure the rider didn't take El Toyonal to bypass Alta Vista, for example. But for the purpose of writing timing code for Low-Keys, I'd use only enough gates to make sure no short-cuts were taken.

Fortunately Strava provides an API by which I can download raw rider data. So if I am provided a list or ride numbers for riders who wish to participate, I can download the data and do my course timing using a Perl script. Then I can use any timing algorithm I want rather than relying on Strava's native timing.

Low-Key Hillclimb #1 in the bag: 8 more to go

2011-10-02T06:33:00.000-07:00

Low-Key Hillclimb #1, Montebello Road, went exceptionally smoothly yesterday. It's always a huge relief to get #1 out of the way.

at the finish line with volunteers; Han Wen photo

Low-Key runs on a volunteer coordinator model. However, I always take the first climb. That's not because I'm the best coordinator (I'm not) but because I want to establish that I'm willing to do my share. This year I'm also coordinating week 2, Sierra Road. Week 3, Page Mill, is still open. I hope to get a volunteer for that so I can ride it myself, not losing too much fitness this month...

There's always a few issues. This time I tried a new, for me, timing method which is to use the Ultrachron app for the Android iPhone as our back-up timing. Primary is Howard Kveck on his manual stop-watch.

Problem is while in principle the Ultrachron app does what you need, the touch screen gets in the way, and too often I'd accidentally kick it into timer mode, losing lap times. But fortunately Howard was on top of his game, and we got a great set of results.

Another issue is not everyone crossing the finish line is in the Low-Key Hillclimb. The roads are open to anyone, cars or bikes, and Montebello Road is a popular target. So while the Low-Keyers all had stickers on their forks, when things got crowded, this still created confusion.

But still, results went relatively quickly, the weather was great, car traffic was very light, and everyone seemed to have a good time.

Next in the series in Sierra Road. Hopefully the rains forecast for this week blow through in time for clean and dry roads by next Saturday.

Then Page Mill.... as I noted, no volunteers yet for Page. Every other week has a very capable volunteer coordinator lined up. So if anyone's interested in Page Mill, please let me know!

SB910 sits on the desk

2011-09-30T06:24:00.000-07:00

After it passed both houses of the California state legislator, SB910 (see this web site) has been sitting on Governor Brown's desk for several weeks. A mere formality? Maybe not: word is that the AAA and the California Highway Patrol are both actively lobbying for the governor to veto the bill.

I've been a big fan so far of Governor Brown, and have supported some of his vetos so far. But vetoing this bill would be irrevocably sever my support for him. As I've described before, none of the arguments which have been made against this bill hold water. Indeed, CA is being left behind as Nevada's 3-foot passing law just this week went into effect. Passing such a law has ceased to be a statement of progressive advocacy of cycling as transportation and is more and more becoming simply following the herd.

So if the arguments proposed against the bill don't hold water, why the continued resistance? Simple: the AAA and CHP are both organizations of drivers. In the case of AAA, people who drive more than average are more likely than average to be members, and in the case of CHP, driving is what they do for a living. People who do something don't like feeling constrained in that activity, and members of both organizations would prefer to not have their right to endanger cyclists curtailed. Sure, the law already requires a "safe passing distance" but that's so fuzzy it's easily side-stepped. Without a firm number there, any pass can be claimed to have been safe, any collision due to the actions or lack of actions of a cyclist.

So Mister Governor, this is your chance. Sign the bill and do the right thing, or veto it and go on the perpetual black list of everyone who rides a bicycle on public roads in this state. And whatever you do, please get on with it. The arguments for, and those which claim to be against, this bill have been hashed and rehashed for decades now.

political vent: Federal transportation spending, "stimulus", bike infrastructure, and the Republican Party

2011-09-13T18:19:00.000-07:00

Okay, my chance to vent on some political issues right now.

There's been a lot of talk about the federal government, the Tea Party, and spending on pedestrian and cyclist infrastructure. "There's no federal interest" the Tea Partiers will say, and the democrats and their supporters foam at the mouth in return: "what? There's no federal interest in children being able to walk or bike safely to school?"

Yes, that's correct. What's meant here by "federal interest" is that the states have the obligation to take care of their own citizens. However, when there are issues in which a state would be obligated to look out for the citizens of another state, that's a federal interest. So for example if there's an interstate highway going from state A to B to C, and state B were to decide it didn't want to fund repairs, then that would affect states A and C. So the federal government funds that highway.

So the federal government should fund interstate bike paths. But since most bike and pedestrian infrastructure is local, there's no "federal interest" in this local infrastructure. It doesn't mean it doesn't matter, it simply means the states have the responsibility to look after it.

So what about roads? Really what should happen is the federal government should fund only interstate highways, and then charge users of these roads to cover the costs. That's obviously not what happens. What instead happens is the federal government doles out transportation funding to states who find all sorts of justifications to spend it. "Job creation" is a big buzzword. It's while the federal government funded so much road building during the first stimulus program in the Obama administration. "Shovel ready" projects create jobs, right?

Well, not really. Consider the 1930's... Roosevelt... the New Deal. The federal government pumps money into a road. The road builder hires American workers to build the road. He orders equipment like tractors build in American factories. Those factories hire more American workers. Then people use the roads and with more roads there's more demand for cars. The cars are built in American factories, and those factories hire more workers. The gas burned by those cars is pumped from American wells, and refined in US refineries, and these facilities hire more American workers. There's a cascade of jobs, and that's just the first round. There's a further cascade from each of these factories, wells, and refineries, as the equipment they use is built in still more American factories. This is called the "money multiplier effect". If you insert a dollar into the economy, it bounces around many times, and the increase in debt associated with that initial investment is paid back with interest due to all the economic activity which results.

But it's not the 1930's any more. You put $1 into the economy and it's likely across the border pretty quickly. None of those factories are in the US any more. The oil isn't pumped in US. The cars likely aren't built in the US. And workers aren't spending their new income on US-built goods.

Here's a chart related to that I pulled from The Google:

So stimulus doesn't work. Sure, things look good in the short term. If you borrow money you spend more. But the cascade is very limited. There's no exponential expansion of economic activity to pay off the interest on that debt.

Well, increasing spending is good anyway, right? Well, not really. Our problem in the US isn't lack of spending. There's plenty of income. The problem is with how that income is divided. Too many people aren't getting the benefit:

Another random plot from the web:

So there's plenty of income out there. The problem is that income isn't circulating through the economy. It's not creating jobs. Put more money into the economy and it's just going to congregate where the rest does: at the top of the income pyramid.

There's another big problem with auto infrastructure. An analogy I read is it's like getting a puppy for a gift. Only part of the cost is up-front; the real costs are down-stream. You've now got to maintain that roadway or bridge or overpass in perpetuum. That's a future obligation in addition to the obligation associated with the initial loan to build the thing. Building roads and associated infrastructure, because it is maintenance-heavy, is digging the budget into a deeper and deeper hole.

So really the solution is we need to stop focusing on spending, especially on transportation infrastructure consistent with our present infrastructure model. The solution is to change the rules of the game to make it more favorable to create jobs. That means redirecting the tax burden from hiring and business to consumption and spending. You've got to provide a favorable environment for business to create. And you've got to have the human infrastructure in place for business to operate. That means a healthy, well-educated populace which has adequate incentives to go in the sort of professions we need to be internationally competitive: primarily science and engineering. Presently these professions are underpaid relative to finance, law, medicine, and business. Even suburban cops make more, with better benefits, early retirement, than PhD level engineers and scientists in high-tech industries.

So more and more federal funding isn't the answer to anything. To improve the transportation landscape, the solution is simple: encourage people to drive less. And to do that we need to make driving, cost-wise and convenience-wise, less attractive. If we do that, there will naturally be more demand for local investment in bike, pedestrian, and public transportation infrastructure. The federal government need not be involved.

Cobo power estimation on Angliru in 2011 Vuelta

2011-09-12T06:59:00.000-07:00

A few posts ago I posted a plot of VAM numbers for Cobo in the Vuelta stage up the horrific Angliru climb. Of course, Cobo went on to win the race.

VAMs are difficult to assess. More relevant is power: in particular, watts/kg. Making liberal assumptions, I tried to come up with a W/kg estimate of Cobo on the climb (or in the early bit, the leader of the group containing Cobo). Here's the result (power at the crank, not hub, assuming 2.5% drivetrain losses):

Impressive for sure, but still generally below the Armstrong-era "6.7 W/kg" attributed by Coyle to Ferrari.

Fremont Peak VAM comparison, 2010 versus 2011

2011-09-11T18:49:00.000-07:00

Fremont Peak last year versus this year were two very different races. Last year, it was a steady tempo on the lower slopes. I stayed with Carl Nielson as long as I could, until I felt myself going too far into the red, and had to back off.

This year, a moderate pace leading up to the steeper climbing was shattered when Kieran Sherlock launched an attack at the front which strung out the group. Most riders responded to this acceleration, but I chose instead to settle back into what I felt was more sustainable. Kieran had done the same thing on Wednesday's Noon Ride up Old La Honda: a hard acceleration right from the beginning. He's faster than me, that's all there is too it. My PR up OLH is 16:49, and that's a mediocre time for him these days.

What I'd hoped is I'd be able to pull back some of the other riders. And I did, a few, but not as many as I'd thought I would. I crested the top of the climb just behind a Dolce Vita rider, having passed another guy not long before. The guy behind me powered past on the descent, and I followed the other two across the finish on the narrow, flat finishing road.

So that was the tactical view. But with climbs, the numbers always tell an interesting story.

Here's my local VAM for the climb. I used a truncated cosine-squared smoothing function with time constant 60 seconds to generate a rolling VAM number describing, at each point of the course, my rate of gaining altitude. Here's a comparison last year versus this year for me:

The 9.5% portion of the climb begins around 9 km into the race (the race starts somewhat before the profile I posted yesterday). So for the undulating, generally shallow grades before this VAM numbers are naturally lower. But then you see a similarity between the two years. Last year I started the climb harder: I was following Carl. But then I couldn't hold that pace and faded until the crest at around 15.2 km.

This year I also faded. The fade was less than last year due to my more conservative start, but the fade is still clear in the data. Arguably for a bit I was actually faster this year. I caught two other riders and it got slightly tactical, perhaps slightly reducing my pace toward the top of the climb. My descents last year and this year were very similar (negative VAM). You can see I stopped at the finish line last year (the sharp dip in the red curve) while this year I kept rolling up the following climb in the park.

Honestly I'm a bit disappointed. I understand why I faded last year but I should have been able to sustain a higher VAM this year. Explanations: older and slower, sure, but also my preparation has been different. Last year I was at the tail end of training with a coach I really liked, Dan Smith. I was doing his Wednesday morning ride fairly regularly, and if not was doing the Noon Ride at lunch on Wednesdays. This year I've been mixing running and cycling, and have done only two Wednesday Noon Rides all year. I've also done Diablo three times. That's basically it for my extended climbing all year. And I've always had issues mixing running and cycling: this year it's gone better than in the past, but last year I could feel my cycling tangibly improved when I stopped running. Running not only leads to lost opportunities for specific training, but racing especially digs me in a hole from which it takes at least a week to crawl out of.

So that's that. I was slower this year, and excuses, excuses, excuses. I'll try to think more about the 1:30:56 half-marathon I ran two weeks ago. I felt good about that and still do. Hard to do well at everything all the time, even if I'd prefer to.

Camino Alto

2011-09-10T20:44:00.000-07:00

Since Tam was canceled for today, I decided to make a run at my Strava PR for Camino Alto. Virtually all of my existing Strava times for the climb are on my Ritchey Breakaway, a relatively light bike in the low-17 lb range with clincher wheels, but with the Fuji I place weight at a priority. For Mt Tam I had my Reynolds 35-T carbon "road race" wheels (with Veloflex Record tires), not my lightest configuration but with the Garmin 500 checking in at 12.5 lb -- still pretty good. So I decided to risk taking these tires across San Francisco and into Marin for the climb.

Ironically the worst road I ride glass-wise, not even close, is Central Expressway in Mountain View which is typically part of my commute. The streets in San Francisco, despite a serious problem with cars getting broken into, are fairly glass free. It's because the streets here get swept once or twice per week, while once glass is on the Central Expressway shoulder, it's there for months. So I've had very few flats riding in the city, and my tires were fine today.

The Golden Gate Bridge just reopened the dedicated bike path after around 4 months of seismic reinforcement (or something; it's not obvious that anything was actually done). This was a huge relief after months of battling for space with distracted tourists on the pedestrian side. But it was hardly a comfortable crossing: winds were as strong today as I ever remember them, and crossing the bridge on the Fuji was a challenge. The Reynolds rims aren't the best in a cross-wind, and the Fuji's exceptionally large trail results in a proportionately larger steering moment when the front wheel experiences a lateral force. So it was dicey going.

When I reached the base of Camino Alto I stopped to place my vest and mostly empty water bottle near the side of the road. Why bring these up the hill when I was going to just return immediately after climbing? I forgot to also drop off my cleat covers and multi-tool. Every gram counts, and maybe the weight from these objects would have made a difference of a second in my Strava time.

I started hard. Here those strong winds provided a tailwind which was an obvious favor. I felt fast riding up the hill. The grade is modest, so I was able to stay in my 46 tooth big ring rather than go down to my 36. In the rear I had a 12-27 making it easier to avoid the downshift. It felt good riding the carbon bike with the carbon wheels instead of my usual steel bike. There's a hollowness to carbon frames that provides the feel that you're riding a high-tech race machine.

I love Camino Alto. The grade is never steep, but there's enough variation that you need to shift often enough, and you need to keep concentration up to avoid letting power slacken on the lesser-grade sections, and need restraint to avoid going into the red on the steepest portion around a third of the way up. It's short enough that if you start too slowly you don't get to make up for that mistake later, but it's long enough if you start to hard you'll fade at the end. There's some car traffic, but not too bad, and there's usually room for cars to pass without too much delay. And the cars expect cyclists on the road, because there's always cyclists on the road: it's an extremely popular route for individuals and groups riding between the city and either Fairfax or the Paradise Loop.

Indeed I was having such a good time that when I looked down at my lap time I noticed I was already at 4 minutes. My goal was sub-4:30 so this meant I was almost at the end, and sure enough I realized I was going into the final curve. I'd not been paying enough attention and should have been riding harder at this point. So I tried to ramp it up as much as I could, going for an imaginary finish line whose exact location was determined by the whims of GPS error.

My time: 4:25. This was 10 seconds off my previous Strava best, although I'm not sure it's my best ever including untimed efforts (I should check my Powertap data). This tied me for 13th on the climb; to get 10th I would have had to have been two seconds faster. Ah, well: the stuff in my pockets wasn't worth two seconds for sure.

VAM wasn't so impressive. Part of that is the low average grade: only 4.7%. But part is also that I've never been able to generate my best average power for the duration on this hill (when I was using a power meter). All of the changes in grade make it hard to stay on top of a good gear, and to really nail average power for a time interval, I personally need to be able to keep my power fairly constant. For example, I've gotten higher average powers down on the Peninsula on Kings Mountain Road to Huddart Park or Page Mill Road to Altamont. But for sheer fun, Camino Alto is my favorite.

Mt Tam curse

2011-09-09T20:22:00.000-07:00

Two years ago, Mt Tam Hillclimb canceled due to the fabricated state budget crisis. Last year: I was traveling in Italy. Tomorrow: Ready to go! But then Mt Tam canceled again (this time due to forest fire risk from forecast dry lightning).

I'm stunned.

I'd updated my racing license ($60), paid my $35 entry fee, refrained from running this week, done a shake-down OLH on the Fuji on Wed, and even eaten pasta tonight.... all for naught. Except perhaps for Fremont Peak on Sunday.

Last year I posted a profile for the Fremont Peak Hillclimb, which this year is the day after Mt Tam's scheduled date... a new back-to-back schedule for two rare hillclimbs on the NCNCA schedule. I don't know if I'll make it, but a comment on that profile:

Fremont Peak from San Juan Baptista (Motionbased:mooseman)

And again here's the grade extracted from that profile, which I've convolved with a Gaussian of standard deviation 100 meters, to keep the numbers significant:

Fremont Peak from San Juan Baptista (Motionbased:mooseman)

Sometimes too much knowledge is a bad thing, and last year I was waiting for that little sting in the tail at the end of the profile only to discover since the Park Service had denied access to the race (Park Service is generally uncooperative to cycling events: ask the promoters of the recent professional stage race in Colorado, who were denied access to park land, as was the Tour of California) it wasn't part of the course. So I was ready to jam it there when I hit the finish line just behind the rider ahead of me. Whoops...

I defined a Strava segment for that race, avoiding the start line which contaminated the segments of me and some other riders:

It's a fun race and hopefully I can make it down there on Sunday.

But I'd really been looking forward to Mt Tam tomorrow. No insult to the Fremont Peak crew, but Tam is a classic.

Vuelta 2011: Cobo on the Angliru

2011-09-08T21:22:00.000-07:00

Arguably the most epic stage of this year's Vuelta was Sunday's race to the mighty Alta d'Angliru, arguably the toughest steep climbs in professional bike racing.

Here's a profile of the climb, from Climb by Bike:

In the stage, Geox's Juan José Cobo climbed away from the group of favorites, including Team Sky's 1-2 in overall GC, Bradley Wiggins and Chris Froome. Some argued that it was a sign Cobo was dipping into the special sauce, a result too good to be true.

While I realize VAM is subject to many sources of variability and is hardly a valid test of doping, I can't resist running the numbers here. Fortunately I managed to take time splits for each of the final 10 km for Cabo or the group containing Cabo, although I had to guess a bit at 6 km to go, where Cabo was trailing Anton. I guessed that Cabo was around 5 seconds behind here, although I didn't see this explicitly on the Eurosport coverage which I was watching on the web. These times are on the crude side and each of them probably has error bars of up to 3 seconds either way, not even counting the possibility of variable delay in the video on the web channel.

Then to calculate VAMs I needed altitude data. Unfortunately the altitude data provided by ClimbByBike.com or Altimetrias are provided only each km measured from a start to the climb, and since I was recording at points which were even kilometers from the race finish, I needed to interpolate. My first approach was then to tap into some data from Garmin Connect. However, the Angliru is a twisting climb, with plenty of temptation and opportunity to take wide lines through corners and even traverse (as Bradley Wiggins was seen to do). These moves result in a lengthening of the rider trajectory which causes a reduction in the apparent grade. The Vuelta distance markers are likely set under the assumption of the path a car would take. So I was back to ClimbByBike. Fortunately the Climb by bike profile seems to match up well with the Garmin data until the steep bit where, predictably, the rider chose a path to mitigate the brutal grade:

The 17.5% grade shown in the plot is only for the steepest kilometer. The climb pops over 20% on numerous occasions but only briefly. Digression: interesting comparing this to Mix Canyon Road, often considered the toughest climb within driving distance of San Francisco:

Needless to say Angliru is tougher... Anyway, back to the topic.

The Vuelta web site claimed the KOM marking the top of the climb was 600 meters from the finish. This makes sense from the timing, corresponding to Cabo riding 44 kph along this portion. It was a twisting gradual descent so 44 kph seems no more than 10% low. So I used this to align the ClimbByBike data with the course. I assumed the KOM was at what ClimbByBike considers the end of the climb. This was 550 meters after the previous km point there, so since this was 400 meters past the 1 km to go banner, I assumed an offset in even km points of 150 meters between the two data sets.

In 2008 Alberto Contador won the stage on this hill. Cozy Beehive assembled the videos of that stage, which show Contador climbed from 4 km to the finish in, to to second, the same 15:25 I calculated for Cobo this year.

Despite his amazing climbing this year, Cobo still used a 34×32 low gear, the smallest of the top riders on the climb. It's another nail in the coffin of the "tough it out" old school thinking on gearing for optimal speed on climbs.

Diablo North Gate: two Strava segments

2011-09-03T20:07:00.000-07:00

Today, encouraged by the temperatures in the 70F's being reported for the Diablo summit, in striking contrast to the bone-chilling fog of San Francisco, I put on a jacket and rode to BART. I heard the announcement for the Pittsburgh-Bay Point train I wanted as I reached the top of the stairs, so I ran down the stairs in my cycling shoes and, after a few contact fouls with people leisurely walking the opposite way, got onto the train only 5 seconds before the door shut.

Ironically BART to Pleasant Hill takes about the same time as it would take me to ride out to the Golden Gate Bridge. Then the suburban-timed traffic lights from there to the base of the climb are no more delay than the riding the pedestrian path across the Golden Gate Bridge. Despite this for a long time East Bay rides were all too rare for me. Instead I bundled up against San Francisco and Marin's summer chill and rode north. But I've been riding Diablo more recently, trying to work on my extended climbing, breaking out of the rut of efforts 5 minutes and less. Even Mt Tamalpais lacks continuity from the most popular access point in Fairfax: that climb is in segments separated by short descents. For a continuous effort, Diablo is the easiest weekend access, the climbs on the peninsula south of San Francisco less attractive due to Caltrain's poor weekend service.

So in part to help kick myself out of my post-half-marathon tired legs slump, today I rode Diablo.

The Low-Key Hillclimbs last climbed Diablo in 2009. We started the climb just past the gate which sits on the road at the northern boundary of the park. This is also where C4 Racing starts their race, which goes 10 km up the hill, ending 500 meters from the junction between the North Gate, South Gate, and Summit Roads. So I think it's safe to say this is a popular "start" to the climb.

For the finish of the climb there's no question: the end of Diablo is a memorable experience. Strava has it at 15.2% for 200 meters. And at the end of 1100 vertical meters you feel every % of that... But it's the final sprint, so you do what it takes to make it to the parking lot at the top. A sign tells you where you are:

Brenda Brunner photo

So remembering my climb of Old La Honda a week and a half ago, I made sure to use the lap timer on my Edge 500. I'd upgraded my firmware so the Strava representation of my lap should be within 2 seconds or so of reality.

Here's the start of the lap as reported by Strava:

The sign is near that big tree, so the position here is fairly close to where I hit the lap start.

Then here's the finish:

There's a bit more error here; obviously I wasn't riding off the side of the road. I hit the lap finish close to where the paint is directing traffic, somewhat past the end of the indicated segment. But the error here is also fairly small.

So I was surprised when I compared my lap time to what Strava gave me for the segment Diablo NG to summit.

Here's what fitdump says about my lap:
temperature (13-1-SINT8): 37deg.C (37)
timestamp (253-1-UINT32): 2011-09-03T12:59:34 (684014374)
start_time (2-1-UINT32): 2011-09-03T12:01:12 (684010872)
start_position_lat (3-1-SINT32): 37.9017256deg (452185200)
start_position_long (4-1-SINT32): -121.9931266deg (-1455434692)
end_position_lat (5-1-SINT32): 37.8815485deg (451944477)
end_position_long (6-1-SINT32): -121.9149209deg (-1454501662)
total_elapsed_time (7-1-UINT32): 3501.660s (3501660)
total_timer_time (8-1-UINT32): 3501.660s (3501660)
total_distance (9-1-UINT32): 17607.62m (1760762)

Time was 58:21.7. This is only 0.7 seconds longer than Strava claimed for the lap, a smaller difference than I'd seen at Old La Honda due to my firmware upgrade and selection of one-second sampling.

But when I look at the segment, it says 58:35. So where'd that 14 seconds come from?

The listed start for the segment is the same, so no difference there:

But the finish is quite striking -- the finish was marked in the middle of the parking lot:

Bad match, perhaps? Here's the reference data:

Also the middle of the lot.

That's sort of strange, since many riders will reach the lot then stop, or even reach it then take a tight left to descend. On the plus side, moving the end of the segment away from the steep section guarantees riders will have completed the entire steep portion. But in this case it seems almost criminal to destroy people's results because they stopped in what at least in theory could be a congested lot.

So there's another segment: "Diablo North Gate to Summit". I'll not post images like I did from the last one, but if you follow the link you can see this one starts considerably later. South of the north gate itself, there's a bit of a rolling section which includes a descent before the road begins its monotonic ascent. This segment omits this beginning portion, which I understand. But the finish of this segment is what's most unfortunate. It's at the beginning of the 15% grade leading to the lot! I'm sorry, if you reach this section then faced with the sheer horror of what awaits, you turn back, then you, my friend, have not climbed Diablo. You've got to include that last portion.

So two segments, neither really doing the job. Of course I could define my own, but there's already too many segments flying around. Just remember when you climb Diablo to keep going once you reach the lot to make sure you trigger that first, more complete segment.

San Francisco Giants Half-Marathon

2011-08-29T08:29:00.000-07:00

Call me Fred.

At least that was my name as I lined up yesterday for the San Francisco Giants Half Marathon. We lined up on Terry Francois Boulevard, on the opposite side of Mission Creek from Pac Bell Park. When I looked down at my left shoe I saw a "D-Tag" plastic-encapsulated timing chip which clearly identified me as Fred.

An offer for the ticket had been posted to a mailing list I read on Thursday. I'd just gotten back from a 20.3 km lunch run, which I'd finished fairly briskly, so by my usual standards I'd not be ready to run again until Sunday. But this was too good to pass up, a "big" race happening less than 2 miles of home, and at a distance longer than I'd raced before on the road.

At the starting line they had three self-selected groups: under 7 minutes, 7 - 9 minutes, and over 9 minutes per mile. I expected to be around 7:10 - 7:15 per mile, or around 4:30/km. On my lunch runs, this had been a fairly good pace, and I'd only done it for a few km at a time, never sustaining for close to this race's 21.1 km distance. So it was fairly clear: I should be in group B.

But then I saw the pacers. There was a guy holding up a sign for "1:40", which is 7.62 minutes per mile (7:37/mile). I initially joined him. That pace I certainly knew I should be able to do, unless my still-tired legs let me down. Up the road a bit was a sign for 1:30. 10 minutes doesn't seem like a big jump, but that corresponds to 6.86 minutes / mile (6:51/mile). That's a huge difference in pacing. So I had the choice: go out a bit harder than I could sustain, slower than I could sustain, or self-pace? I knew self-pacing is tough, and I think the advantage of drafting in running is often underestimated, so I preferred to be in a group. It's very tough to come back from too slow a start, so I finally moved my way through the crowd to the 1:30 group, even if this risked me blowing up. Running with the group would allow me to run and think only about economy, not about speed, which I figured was a considerable advantage. I always had the option of dropping off the pace if it was digging me into a hole.

So there I was with some rather fast-looking people as ten minutes after the start the horn blew and we were off.

Immediately I could feel I was out of my comfort zone. I knew from previous road races, in particular the 10 miler I did with Dolphin, that a brisk start was typically closer to 6:30 than 7:00, and this felt similar. But these pacers know their stuff (I'd had a nice talk with the 1:40 pacer, who'd been meticulous in his preparation, before I promoted myself to the 1:30 group), so it must have been my tired legs.

The first water station came sooner than I anticipated. I guessed this was due to the 5 km runners. Despite having run for only a few minutes, I decided to take a drink; I hadn't drunk since leaving home. The first table was inaccessible due to paper cups littering a wide swath next to the table (who were all these riders up the road drinking this early?) but the second table was clear, so I grabbed my cup and actually managed to get most of the water down my throat. A bit went down my nose, more on my shirt, but overall a success. I was glad I'd watched DC Rainmaker's video tutorial on drinking at races. Despite having handled this fairly well, I still lost time on the pacer, and had to increase my pace yet more to close that gap.

The pace finally subdued a bit, but despite this we hit mile 3 I looked at my watch and saw 20:05. That's a 6:42 per mile pace, well above the 6:51 target which I'd considered a reach. "Is everyone feeling okay?" the pacer asked cheerfully. Err... that's not how it's supposed to work, I thought, but didn't say anything. This was a race, after all: no whining allowed.

"Pick up the pace!" a volunteer called from the side of the road. I laughed.

Water stops followed with remarkable frequency. I drank sports drink at the second: it was a bit too strong but I handled it. I decided to alternate water and sports drink. But for the rest of the race I drank at only two more stops. For a 90 minute effort, with drink cups the size they were using, even with the considerable loss of liquid onto my shirt and into my nose with each drink I was still hydrating ahead of the schedule I normally sustain at trail runs, let alone on training runs. Overdrinking can cause stomach problems, and in the cool morning air dehydration was a diminished concern.

There was still a decently dozen or so together with the pacer when we reached the "climb" on the course, the bump of Fort Mason. I run this hill regularly, so have it fairly dialed, and did my best shuffle-step climbing pace up the hill, letting the pacer, leading the group, gap me a bit. I was later surprised to see I PRed the climb on my Strava record: my previous PR I'd been trying to chase down a cyclist, not worried about maintaining a sustainable effort.

But here's where things went wrong. Following the climb there's a gradual descent back to base elevation, and apparently the pacer wanted to immediately make up for the time lost climbing. He immediately threw open throttle, and I simply couldn't respond to the acceleration without putting myself into the red zone. And the red zone is exactly where I didn't want to be less than halfway into the race.

So now I faced the western run along Marina alone, unfortunate since there was a stiff headwind. I was questioning my decision to let the pacer go, but clearly they were faster than pace at this point, and at some point you've simply got to pull the plug. Another runner soon caught me from behind, however, and after pacing him for a bit, I slowed slightly and pulled over to let him eat the wind for awhile. He complied.

Another runner caught and passed us, causing my partner to pick up his pace as well. This caused more discomfort. But I was able to hold on and get at least some shelter from the wind before the turn-around.

As we turned onto my preferred dirt surface here, and the headwind became a tail, I immediately felt better. Add in I was now running toward the finish rather than away from it, and my spirits clearly notched upward. I as running here with a very fit, fast looking woman who ran with a high cadence and smooth style which was encouraging. Honestly the shuffle of riders which led to this point is lost in the mental haze: my focus was distracted. On this section we needed dodge other runners, walkers, and mountain bikers using the popular path. I'm fairly obsessive about keeping the optimal line so prefer to apex turns if i can.

The woman and I caught and passed a guy whose style was in market contrast: a bounding, bouncing stride with notably slower foot turnover. After I checked to see he had a number on his shirt, I wondered how he'd gotten so far ahead of me running like that. I suppressed the temptation to give him advice: after all he'd been ahead of me for the previous 8 miles. But the woman and I managed to quickly leave him behind.

Fort Mason came and went quickly: in this direction the climb is easier. There was a slight confusion of the exact route due to the increasing morning crowds. As I've noted, I have a neurotic paranoia about going off course, but there were enough runners within sight that I assured myself I was unlikely to make a mistake.

Mile marker 9, the flag in mirror image due to the wind direction. My pace had not surprisingly slipped: I was still ahead of 7-minutes-per-mile but by less than before. I imagined the broom wagon was following at that pace, ready to disqualify and sweep up any unqualified runners who had placed themselves in the first wave.

But I was running well, and only 5 km to go. 5 km is around when it's time to ratchet up the discomfort threshold. Sustaining the pace was no longer much of an issue: it's only around 20 minutes of pain left, then the finish. So my protocol became regular: focus on relaxation to minimize the effort at my present rate of speed (relax the head, relax the shoulders, relax the torso, relax the legs), then increase the rate of speed, repeat. Speed doesn't actually increase, of course, but if I didn't try to keep it increasing it would otherwise start to drop.

Around a half-mile later we hit the 10 km turn-around (the course was longer out than back). This was a bit of a problem, as we were no overtaking more relaxed 10 km'ers, who'd started with us. Many if not most were walking, too many several abreast. Since I viewed every second as precious in this run, I wasn't as patient as I might have been with people taking up a majority of the path, and in one case I threaded a gap between two walk-runners who were cutting an exceptionally wide profile. "EXCUSE me" I heard indignantly as I ran by. For a second, just a second, I thought maybe I was taking myself too seriously, and should have swerved around them rather than taking the straight line between them. But I recovered from this thought quickly: they realized they were in a race on a shared course, and they were clearly giving zero consideration to the faster racers coming through. So I felt I had at worst only recipricated.

The route here was extremely familiar. The Bay Bridge loomed ahead, and I knew the stadium was less than a mile from the bridge. I kept my stride to the bridge, playing my game of relaxation alternating with acceleration. Soon, very soon, I would be on the baseball field. Cool.

Approaching the field, volunteers were everywhere, directing us towards the stadium entrance. I knew this entrance well: it's a spot at field level, separated from the outfield by a fence, where fans can stand and watch the game for free, a nice public relations move by the team. But now there was no fence. I ran among other runners onto the dirt margin of the field. There was a barrier blocking our access to the grass.

The finish wasn't far: between 3rd base and home plate. In contrast to our run, the field seemed puny, miniature. I sprinted as I could first over one tag reader, then over a second which was under the banner.

1:30:56. I could say I'm disappointed to not have broken 1:30, but that would be goal creep of the worst kind. Honestly, I never anticipated I would be able to run that fast, especially with tired legs from a long run two days prior. So that was very encouraging, not only for this day, but also for a Boston-qualifying marathon in my not-too-distant future, perhaps.

After crossing the line things got surrealistic. We were immediately handled medal, then a bottle of drinking water, then a yogurt and a spoon, then a back-pack-like bag. I walked a bit, admiring the hyper-groomed field, looking in the perfectly kept dugout. Looking up at the seats made them seem so very close. What a special experience it must be to play here, I thought.

I stopped to eat my yogurt: my hands were full otherwise. "Keep moving!" someone shouted. I was supposed to move into the stands and walk up some steps.

My shirt was soaked, not only with sweat but also a mixture of sports drink and water from fumbled drinking at aid stations (I'd managed to get most of the liquid down, but there was always significant spillage). I was starting to chill, and wanted my jacket, and I needed to pee.

But there was more. We were led onto the mezzanine where we were handed a Safeway bag (???), then a T-shirt, then a Giant bobble-head.... somewhere along the line I'd also been given a banana and an apple. I could barely hold it all.

After stopping in a men's room, I continued on. Clif bar samples, coconut juice, some sort of "healthy" chips, a photo with our medals, and booth after booth. There was a booth showing some shoes which looked interesting, but I ignored it: I just wanted my jacket.

Finally I was through the gauntlet. The FedEx trucks which had our drop bags was outside. I got my jacket and sweat pants, now quite cold in the chilly air which not so long ago felt warm, and began the walk back to the start area to pick up the bike I'd locked there. Hopefully it was still there...

When results were posted, I'd scored 6th in age group for Fred. I was glad I hadn't let him down. For the other runners in his age group, I was qualified, so there was no problem with contaminating the results. I'm not sure I'll enter this race again next year. The orgy of swag is a downer. But the support was fantastic, and I'm glad I did the race.

Garmin Edge 500 sample time and Strava lap time determination

2011-08-26T08:47:00.000-07:00

Last post I noted there had been a difference between the reported lap time up Old La Honda and the extracted time for the segment. The lap time had been reported by Strava as 18:12, while the segment time was quoted as 18:22, a ten second difference. But looking at the difference between the point at which the end of the segment was tagged and the end point of the lap I had a difficult time imagining that was worth ten seconds. After all, while I slowed there, not just coasting but actively braking to check for cross-traffic, I never came to a complete stop (law enforcement: my account here seems to have been hacked).

So it occurred to me I was assigning the full error on the segment time, but assuming the lap time was correct.

Back when I was making suffer score test files, I used a 3 second sampling time, the same as is used by my Android phone. Since each sample represents 3 seconds of ride data and since there are 3600 seconds in an hour, I generated 1200 samples. I figured I was calculating the suffer score associated with exactly a one-hour (3600 second) ride.

But to my surprise, the rides reported a duration of only 1197 seconds. Strava was taking the difference between the first and last point, not included the time included in the first point.

So with this experience, I went into the FIT file itself (using fitdump) and checked what the lap time was. The result: 18:14.55. Ah! So it turned out the difference between the segment time and the actual time I recorded with my lap timer wasn't quite ten seconds after all: it was only 7.45 seconds. That's still a big error, and the conclusion still applies you should use your lap timer, but you can't trust strava to report the proper lap time.

Suppose I have ten data points taken at 3 second intervals. Here's the points along with the lap numbers.

3 1
6 1
9 1
12 1
15 1
18 2
21 2
24 3
27 3
30 3

With Strava's apparent approach, the first lap was 15 - 3 seconds = 12 seconds, the second lap was 21 - 18 = 3 seconds, and the third lap is 30 - 24 seconds = 6 seconds. If I add these up I get 12 + 3 + 6 seconds = 21 seconds. Strava would then calculate the ride time as 30 seconds - 3 seconds = 27 seconds. So I've lost 6 seconds: one sampling time for each time I've hit the lap button.

Back to my ride: lap 1 = 56:13, lap 2 = 18:12, lap 3 = 68:54. That's, 2:23:19. But my ride time is reported as 2:23:28. So I lost 9 seconds there.

I checked and my Garmin was on "smart sampling". Whoops. That means while it typically samples at one-second intervals, if nothing interesting is happening according to its assessment then it goes to a longer sampling interval.

Here's the sample intervals found in my file:

Honestly I though I'd shut smart sampling off. One thing it's important to remember is Strava doesn't like smart sampling. So why have I been so stupid to use smart sampling?

Honestly I didn't try. I'n running version 2.60 of the Edge 500 firmware. At that time, sampling time wasn't an option: if you used a power meter you got one second sampling, but if no power meter was detected you got "smart" sampling. Garmin figured the only remotely sophisticated data analysis was being done to power data.

But that was pre-Strava: now position data is also very important. So it's time I upgrade my firmware... the option to force one-second sampling was put in in version 2.70, and they're already up to 2.80, so I'm upgrading my firmware as I type this.

So basically Strava is makes a lower-bound assumption on each lap time, and this will result in the sum of lap times being less than total ride time (assuming multiple laps). This is especially an issue if using a smart phone with a long sampling interval, or if you're using a Garmin Edge 500 with firmare version 2.60 or older. In either case the precise lap time is stored in the FIT file as a lap record. But Strava doesn't pay attention to this number, instead calculating lap times based on the difference between the first and last sample in the interval.

Wednesday Noon ride and Strava timing accuracy

2011-08-25T19:56:00.000-07:00

After working at my job for close to year, I decided it was finally time to indulge in the Wednesday Noon Ride this week.

The weather was perfect: warm but not hot, a so-very-welcome liberation from the tights and long sleeves which have been part of virtually every ride I've done this year thanks to San Francisco's persistent "marine layer". And Matt has done a great job of championing the often-neglected Wednesday ride which has climbed Old La Honda since the Egyptians were first domesticating cats.

I'd done two of the Friday rides so far and the experiment was generally a success: out the door by a drop-dead time of 11:30, back before the 1:30 pm time when the cafeteria shut down its main lunch service. Obviously it's not something to do every day but since my typical lunch break is a line-dependent 7-12 minutes it takes to go to the cafeteria, buy something, then get back to my terminal, it'd still be doable to go even once per week and still average less than a half-hour per day of lunch time. I figured adding the Old La Honda climb - 84 descent and cutting out a missing portion of the Friday loop would add around 20 minutes to this, assuming I didn't wait for the traditional regroup at the top. That's an extra 4 minutes per day.

So there I was, and what a turn-out! Once we'd picked up Greg, who'd ridden from San Francisco, with Kieran also in the group we had three riders who'd done sub-16:15 times. Matt in particular had been on a tear, getting sub-16:05 PR's each of the previous two weeks. So it was going to be an interesting climb.

No way I was going to PR, though. I've barely been climbing this year, a "long climb" for me being over two minutes. I'd done a few north-gate Diablos, which went surprisingly well (not great times, but not suffering my usual Juniper parking lot fade). But other than this my two Kings Mountains and three Old La Hondas had been on the north side of slow. Indeed, when I'd left home Wed morning I'd brought my steel Ritchey Breakaway rather than my much lighter, weenied-out Fuji SL/1. It wasn't worth risking the carbon on the Caltrain commute to work, or for that matter race tires on the glass-strewn Central Expressway which was the quickest route to and from the ride.

I was impressed by the pace to the climb: none of the attacks or surges which virtually always characterize the nooner, no leg-breaking suffer-fest up Alpine Road. It was clear under Matt's leadership this was all about Old La Honda. He was posting a list of times on Facebook, and there was no credit given for a good Woodside town line sprint.

So here was the main point of this post: times on the Facebook post were self-timed, old school stopwatch. Linda, a long-time noon-rider, had suggested that self-timing was obsolete thanks to Strava. I responded that Strava still has significant errors. As I documented here with data recorded by Alan Weatherall using two Garmin GPS units mounted to the same handlebars, times typically differ by up to 4 seconds, which is twice the default Garmin sampling time (Garmin also offers 1-second sampling time). But sometimes differences are larger.

As we hit the base of the climb I hit my lap timer right on schedule. Okay, so I did it at the trailing edge of the bridge, while some claim the leading edge is the official start. Public confession: I time from the trailing edge. I figure the leading edge convention some proclaim is sufficently obscure that many if not most don't follow it. "Bridge to stop sign" is all most people know.

I'll spare the details of the climb. I felt good, though: a good pace for me feels easy at the start slowly increasing in difficulty as the climb progresses. However, drinking from my mostly-empty bottle left me in gasping for air, a good sign I was still working close to my aerobic limit. And indeed the difficulty did increase. At the end, I tried to drop the rider I was with, but he repassed me, and I followed his wheel to 100 meters from the finish, when he gapped me. I tried to close, but couldn't quite recatch his wheel.

I hit the Garmin Edge 500 lap button again right at the stop sign. The last time I'd seen on the lap timer was 18:08, so add a few seconds to that. I then continued on, more slowly of course, up the remaining short climb on Skyline Boulevard before beginning the descent to the Hwy 84 junction where I stopped for water.

Despite my rush to get back to work I took an indirect route so I could chat with Greg (who'd joined me on the way to the store). I still made it back with just six minutes to spare before my self-imposed 2 pm deadline. Within two minutes of reaching my desk I'd uploaded my data and Strava had reported my segment times.

On the plus side, I'd gotten overall second places on two of my favorite segments associated with my normal commute: the Central Sprint and the Ringstorf to Moffitt Sprint. But the focus today was on Old La Honda. Sure enough, I had a Strava PR for the climb. But the claimed time was a bit shocking: 18:22. Whoa -- had really that much time passed from my last glance at the clock as I began my final sprint to the stop sign?

So then I checked my lap time: 18:14.55. Even that was longer than I had figured, but I always seem to underestimate the passage of time at the end of climbs. But that was 7.45 seconds faster than Strava estimated. As I previously noted, if a segment is defined too close to a congregation point, for example the Old La Honda - Skyline junction, then when a rider stops Strava may include waiting time in the segment timing. However, I'd never stopped here: I'd ridden on past the junction in my rush to get back to work. To see the problem you need to look at the data tracks.

Here's the start of the climb, comparing the lap to the segment. They agree fairly well:

But the top of the climb is a different story. Here I overlay the two plots in a GIF animation (reload or click on the image if the animation has stopped). The error in the GPS signal is obviously substantial. It's so substantial, in fact, that it's notable that Strava even gave me credit for the segment. The end of the lap, which is within a sampling time of the stop sign, and the identified end of the segment are way off:

So the result of this little "experiment": if you want a credible Old La Honda time, still time it yourself. The Strava-Garmin combo simply isn't up to the task.

From the work perspective I think the occassional Wed nooner can be worked into the schedule and still be well under the 5 hours per week "standard" for lunch. And on the cycling side that 18:14.55 wasn't bad at all given my lack of climbing this year and all the extra weight I had on the climb yesterday. So even if it did mean I was left eating the Cream of Rice I had at my desk as an emergency ration for lunch, it was worth it.

Strava segment creation tips

2011-08-23T22:57:00.000-07:00

By now, I think most cyclists reading this would be aware of what Strava is. Strava's exponential growth model is dependent on users defining segments. Segments are what provide the arenas for competition which provides the incentive for new users to join, then those users define segments, etc. There can be only one king of a hill, but if you have an unlimited supply of hills, then everyone can be a king (or queen).

But the weakness of allowing anyone to define a segment is there's limited quality control over the segments. Here Strava has a check: a "flag" option whereby users can tag a segment as flawed in various ways. But responding to tags requires human intervention, and human bandwidth is unlikely to keep up with the exponentially user growth model. So at least until they develop a way for automated tag resolution (for example implementing some sort of "thumbs up" versus "thumbs down" voting scheme where segments can be voted off the public database), users rely on segment definers to do a good job in their definition.

Here's some of my ideas in how to define a good segment:

Start with good data. This is by far the most important suggestion. If data are of poor quality, for example shifted versus the actual roadway ridden, then don't use it. Try again another day to collect segment-worthy data, or use another user's ride data (you can define segments from any ride, not just your own) to define the segment. It's likely your ride data will still match the segment if both you and this "reference" ride rode the same path, as Strava's segment matching algorithm is loose enough to allow for typical GPS errors. But it's important the original data be of the best available quality. I've noticed iPhone data tends to be particularly problematic. With me, the Garmin 500 is decent, but my HTC Incredible phone is regularly in better agreement with the Google map than my Garmin.
Avoid starting or finishing the segment too close to a congregation point. The starting point and finishing points should be far enough (at least 10 meters, preferably 20 meters or more) where riders would typically pause before starting a climb, or grouping after the climb. It's very frustrating to do a hard effort on a climb only to discover later that your time was padded by time spent at the bottom of the climb preparing to start. It's okay to click the start point a few sample points into the climb: that's unlikely to change the rankings much.
Avoid starting or finishing a segment on a steep section. Position errors translate into timing errors: GPS is remarkably good, but not good enough for second-resolution timing. The slower riders are going when they start or finish a segment, the greater the timing error resulting from an error in position. So if there's a nasty short climb you want to include, it's actually better to start a segment slightly before, and end it slightly after, the short steep bit. Unfortunately this often conflicts with the "congregation point" suggestion. So sometimes you can't avoid this one.
Try to clearly differentiate segments from other segments. A classic problem is if a dirt path runs next to a paved road, for example, a segment defined for the dirt path will be dominated by riders who used the road instead. The segment matching isn't picky enough, nor GPS accurate enough, to always determine which a rider was one. So the best segments are cleanly separated from faster nearby routes.
Avoid tight switchbacks at the start or finish of a segment. If the route passes close to the finish point, Strava may tag a rider as done (or just starting) at the wrong point on the course. It would be better to terminate the segment either well after or before the switchback.
Avoid segments with obvious short-cuts. Strava has to allow some variation in course along the way, and may well match data for a rider who "shortcutted" a segment. Preferably a segment would be over the shortest distance connecting a few well-differentiated points. For example, for a major climb, the road is likely the fastest way from the bottom point to the top.
Be more inclusive where it won't compromise the segment. For example, I recently did a good effort up Polhemus Road, a local climb. At the top it's common to cut through a short parking lot to a bike path. The official segment continued on the road to the end of the parking lot. As a result my timing included time spent waiting in the parking lot. Had the segment been terminated at the leading edge of the lot rather than the trailing edge, it would have been more likely to have given good timings to riders turning through the lot rather than just those continuing along the road.

Okay, so that's a lot of suggestions. Strava promised to improve segment matching, but there's no substitute for robust, accurate reference data. And since Strava doesn't use map data for segments, instead relying on user-recorded coordinates, it's especially important that users at least try to make that data as good as possible.

Mount Diablo "road plan": my proposal

2011-08-20T07:04:00.000-07:00

Ron Brunner photo of parking lot at Diablo summit
I just saw this on the official Mount Diablo Parks page:

What is a Road and Trail Plan?
A Road and Trail Plan is a recommended management plan for the roads and trails within Mount Diablo State Park. At this time Mount Diablo State Park is developing a Road and Trail Plan for the park. The staff at Mount Diablo State Park is very excited to be able to a part of this process. This plan will be used as a long term guiding document and takes into consideration all of the elements of the park's values, goals and mission.

Key components in the Road and Trail plan are:
Maximize visitor uses and experiences.

Reduce potential safety conflicts.

Minimize natural and cultural resource impacts.

Coordinate with local and regional planning efforts.

Provide access to surrounding public lands.

Public input An essential part of the Road and Trail Plan is creating an opportunity for the public to provide meaningful input. The park staff of Mount Diablo State Park will be setting up a series of public meetings during which time comments will be accepted. Times and dates of these meetings will be posted in the near future.

I'm glad they asked.

The problem with Mt Diablo is the summit road. It's a gorgeous 4.2 mile road from the junction to the summit. It carries moderate car traffic.

The mountain would be a much more serene place if it wasn't for all of these private cars driving up and down to the summit. It's simply silly. Instead I propose to the park service to close that road to cars. Run an electric shuttle bus every hour for those who can't hike or bike. The experience on the mountain for everyone would be improved. The only roar one should have to hear during a hike should be the roar of a swiftly flowing stream, not of an 8-cylinder engine, the only risk should be from mountain lions, not from getting hit by a speeding vehicle when crossing the road.

So how does this proposal compare with the priorities enumerated by the park?

"Maximize visitor uses and experiences": unless those "experiences" involve driving on a dangerous mountain road, clearly the experiences of hiking and admiring the natural serenity of the hill are maximized if there's fewer cars driving.
"Reduce potential safety conflicts": no brainer; big win.
"Minimize natural and cultural resource impacts": no brainer, even bigger win. Few things have a more negative natural and cultural impact than private automobiles on winding, hilly roads.
"Coordinate with local and regional planning efforts": neutral.

"provide access to surrounding public lands.": that is adequately addressed with a shuttle bus or even, I would argue, maintained trails.

Indeed, a few years ago the road was closed due to repaving. Toward the end of the project I went around the barriers and rode it on my bike: the few sections of gravel were trivially passable and not a problem at all, especially since I saw them during the climb and so couldn't possibly be surprised by them on the subsequent descent. I wasn't the only scofflaw: there were a few groups of hikers and even other cyclists I saw during my trip. It was wonderful: no worries about cars speeding around corners, no loud engines to break the peace.

It's a very 1950's mentality that people access = car access. I'd like to hope we're able to move beyond that very flawed vision. It's unfortunate our parks, which should be leading in this movement, are actually well behind the standard set by our more enlightened cities.

Vector launched!

2011-08-18T06:07:00.000-07:00

Garmin announced today the launch of the Vector pedal.

Pretty exciting stuff. I've written a lot here about the then-Metrigear Vector. This certainly looks a lot like the Metrigear version did. Of course the devil's in the details, and I know a lot of development has occurred since Garmin bought Metrigear.

It will be very interesting to see how this plays out. I know that when I'm riding sometimes I feel asymmetric, that my left leg maybe isn't working as hard as my right. But that's based on perceived exertion... now you'll be able to see real numbers.

And for tandems. I don't think there's another option (other than the Look/Polar) for independently measuring stoker and captain power. How many times have I seen a guy pulling his 8-year-old kid on a Trail-A-Bike and the kid (or kids in the case of the linked version) are slacking off? Slap some Vectors on that puppy and keep an eye on those little parasites.

For portability it's a bit of a wash. Swapping wheels (PowerTap), cranks (SRM and Quarq), or pedals (Polar/Look or Garmin) can all be done within a few minutes. But for travel, I think it's fair to say the pedal system is the most portable. You probably travel with your pedals anyway when you are going to rent a bike.

On Polar/Look versus Garmin: I think there's not much contest here in terms of ANT+ Sport or Polar's proprietary communications protocol. Big advantage Garmin. It will be interesting to see when each of the units get exposed to third-party scrutiny how they compare.

I am disappointed the pedals are Exustar Look-compatibles, as I'm a Speedplay fan, or particularly a fan of low-stack light double-sided pedals. Tyler Farrar was testing these pedals last December, so no surprise here.

Anyway, I'm sure there will a lot more to say on this...

Dolphin South End Runners Club Brisbane Scenic 12 km

2011-08-16T20:29:00.000-07:00

Last week I decided I wanted to do a short-ish running race, something close to home and easy to get to. Fitting the bill perfectly: Dolphin South End Runners' Club Brisbane "Scenic" 12 km race.

I'd proposed to Cara that I ride down there, do the race, then she could meet me at the finish and we'd do a little post-race ride together. We could do something flat, as she's just getting back to outdoor riding after back and knee issues. Instead, she said, she'd go down with me at the start and while I ran she'd do a little loop in Brisbane: a circuit with e little hill which has been used for the Brisbane Circuit Race.

The start (Cara Coburn).

So as I got ready to start, Cara left on her bike. I mostly did a few form drills and rolled my legs and shoulders with a massage stick. Then I did some shoulder stretches.

This race was a mixed 5 km/12 km start at a ball field in Brisbane, near the eastern base of Guadalupe Canyon Road. Both courses ran over a short climb on Mission Blue before merging with Guadalupe Canyon for the main climb of around 1000 vertical feet. Up Guadalupe, gradual in steps, then left around a closed gate to Radio Road which averages 7.2% to the summit. The turn-around was close to the summit finish of the New Years bike race. Then it was back to the start.

I noticed at the start line one guy in particular looked particularly ready to go. He had the chiseled look of super-fit guys like Leor Pantelot or Gary Gellin. Curiously he had two ribbons, one for the 5 km and the other for the 12 km race. I wasn’t sure what that was about…. Joining him on the front row were a bunch of other shirtless guys who seemed to all know each other. Hmm…. At least none of them had water, which seemed to validate my choice to leave the bottle behind. 12 km is close to my threshold for carrying water given there was none available on the course. Given the unusually warm sunshine at the start, the security blanket of water would have been comforting.

Soon enough the start, and these guys shot off in a sprint. What? Didn’t these guys realize we had a mountain to climb? I checked out quickly, and they were gone.

As I turned from Mission Blue to the curiously overbuilt Guadalupe Canyon Road (a sign of housing developments thankfully deterred?) I could see what was happening ahead. WAY off in the distance was the shirtless pack. Then there was a larger group including the lead woman. Then two solo guys, then me. I wasn’t worried, however: I could run the pace I could run and that’s all there was too it. I knew most of those following the early rabbits couldn't sustain that pace.

Sure enough I picked off people steadily on Guadalupe. Approaching the 5 km turnaround, I was surprised how many of the shirtless crew were coming back -- they were running the 5 km, obviously, including the double-tagged guy. It seemed wrong -- the quality on this run is all on the upper portion.

Soon I was at the gate, trailing one guy who I was clearly outclimbing at this point (“Alex”) and another on whom I seemed to be slowly making ground only because I was apexing corners he was taking at constant radius. I passed Alex soon enough, and tried to focus on catching the rider ahead of him. The sunshine was now gone as we’d entered the fog which enshrouded the top of the mountain.

Approaching the summit I saw a rider emerge from the mist wearing a Team Roaring Mouse vest. A teammate on a training ride? No -- it was Cara. I was shocked to see her here -- so much for her little Brisbane loop! I was very happy to see she had the confidence to try and the strength to succeed.

almost to the top (Cara Coburn)

The turn-around came a bit sooner than I’d expected. I still felt good, which I immediately tagged as a possible pacing error. I’ve been focusing on trail runs, where this would have been the first in a series of climbs. There you can’t go into the red or there will be a high price to pay later. But here, how much did I need to save for the descent? Maybe I should have hurt a bit more on the climb, given the descent would shift the load to different muscles, and I’d be limited by impact other than aerobic condition. If it had been 6 km into a 12 km flat race, fine, but the mixed terrain had me suspecting I should have pushed it a bit harder in the second 3 km.

Descents are my weakness. I tried to limit the losses to the runner I’d been chasing. But any thoughts of improvement here were shattered when Alex came blowing by. He literally seemed to be flying down the hill, making me look silly. Within seconds he’d passed the runner ahead of me and disappeared into the mist. I wasn’t even breathing hard: what was limiting my speed was instead the stride-by-stride impact of bounding down tarmac at grades up to 10%.

As the grade leveled out near the gate, I was able to actually run again instead of my inefficient downhill stomp. Passing through the gate I was surprised to see that fast 4 km runner heading back up. Later I heard he was doing both races today, the reason for his having both ribbons. Impressive: not so much running 17 km, but doing it with an opening 5 km done at an essentially maximal effort.

I tried to keep my Garmin at over 15 kph from here, but goals didn’t mean much at this point, it was just a matter of going as quickly as possible, of avoiding getting in the way of gravity. Back into sunshine, Alex was still out of view. The guy I’d been chasing was up the road but out of range. I glanced back and the closest chaser was too far back to be a concern. This was essentially a time trial.

My left leg in particular causes me issues on descents (although my right leg isn’t much better): probably some muscular imbalance. But it held out okay, and with the gravity assist I was soon enough back to Mission Blue. Back over the climb and to the finish… 55:04, I think. That was 4 seconds off my rough goal for the run. Ah, well: I couldn’t complain about that.

Approaching the finish (Cara Coburn).

I think I was 7th, but am not sure. Ribbons went 5 deep and I missed out on those. I think what would help my descending, in addition to practice, would be shoes with a bit more padding. I use 6 ounce Nike track shoes, pre-minimalist miminals, and I love them on the flats and especially on climbs but on the bounding downhills I think my muscles could use a break. Even the New Balance Minimalist road show looks beefy in comparison. Also I really need to work on the cycling-related muscle imbalances. A few minutes a day with a theraband might do wonders.

But no complaints: it was s fun day and I certainly can’t be disappointed with my result, especially given the fun, low-key nature of these races.

finisher's ribbon! (Cara Coburn)

Strava Suffer Score decoded

2011-08-09T20:52:00.000-07:00

Strava recently debuted its "Suffer Score", which attempts to quantify in ome way how hard an activity was. Suffer score, they claim, is based on heartrate: ride further or harder and suffer score is higher.

Suffer score is a good move for Strava. Strava's market segment is informal competition. Other web sites have logged how far people ride, and where, but Strava really locked into the demand for competition: competition via social networking. To date these rankings have been based primarily on speed over "segments": routes defined by users, for example on roads or trails, typically up climbs. They then added contests for volume: most miles ridden in a week and most feet climbed, for example. But with Suffer Score, they combine the two: a metric combining both quantity and quantity.

Why's this important? Of course not everyone can climb fast enough to compete for KOMs on popular climbs. But a heartrate-based metric levels the playing field to a large extent. Anyone who's been riding regularly for long enough can go out with the goal of generating an "epic" suffer score for the day.

Suffer score is an obvious correlary with the Coggan "training stress score" (TSS). TSS is an attempt to quantify the training stress of a ride: it's effect on fatigue and its stimulus for physiological adaptation. That both of these can be captured with the same simple metric is far from obvious, but TSS is a blunt instrument, and its popularity among hardcore power meter fans proves people find it useful.

But Strava's focus isn't on hard-core training junkies. It's on the typical Mission Cycling or SF2G rider: people who like to ride, and when they ride like to ride quickly, and as a consequence of this quick riding tend to acquire fitness. However, the goal of every ride is still the joy of the ride, not strictly to improve fitness. So "suffer score" is a "don't take me quite so seriously" number by which long-hard-rides can be compared. "Pain is good", it suggests, not only for it's "training stress".

The big question, however, is how Suffer Score is calculated? To test this, I used a Perl script to generate ride data at constant heartrate. My "rides" consoisted of perfect 1 km radius circules in the Bonneville Salt Flats ridden at a perfect 10 meters per second. Each ride had a different heart rate, perfectly constant, or a different duration.

First I tried using the GPX file format for this purpose, but Strava wouldn't digest the heartrate numbers. So I switched to TCX. This worked nicely.

Here's one of my rides:

From a few tests, I was able to guess the following formula was used for Suffer Score:

Suffer score = K1 t1 + K2 t2 + K3 t3 + K4 t4 + K5 t5

where:
t1 = time in hours in zone 1
t2 = time in hours in zone 2
t3 = time in hours in zone 3
t4 = time in hours in zone 4
t5 = time in hours in zone 5

With coefficient roughly equal to the following (may be off by 1/hr):
K1 = 12/hr
K2 = 24/hr
K3 = 45/hr
K4 = 100/hr
K5 = 120/hr

I like this approach, although I'd be tempted to increase K5, for example to 200. Time in Z5 is very, very painful. Of course this depends on having a good number for maximum heart rate, which may not be the case. In fact a geometric progression would have worked well here: 12, 24, 48, 96, 192. I think this would have helped the criticism I've been reading that it's too endurance-focused.

Aero mass-start frames and the Tour de France

2011-08-07T07:46:00.000-07:00

The trend continues that the big bike companies (manufacturers isn't quite the term, since so few of them actually manufacture anything... perhaps "branders" is a better term...) are delivering aerodynamically optimized road frames designed for use in mass-start road races. These bikes go way back, at least as far back as the Kestrel Talon. But the use of the bikes at the top pro level really began with the Cervelo Soloist (Bobby Julich won Paris-Nice on the bike). Then the Cervelo SLC and SLC-SL really got things rolling. Then Ridley came out with the Noah and Felt with the AR, and Fuji with the SST. And these are just the bikes of top-level pro teams. Kestrel updated the Talon (and Talon-SL), and Litespeed came out with its stunning Archon-CF; neither of these bikes got pro-level attention, however (Kestrel briefly sponsored third-tier Rock Racing, but Rock used their stiffer RT-800).

Yet except for the Cervelo, all of these frames seemed to encounter resistance from riders on the teams using the brands. What was up? The frames were heavier, typically around 250-300 grams heavier than a frame designed without consideration for aerodynamics other than fork shaping, but with the UCI mass limit at a relatively beefy 6.8 kg, with carbon-rimmed wheels there's still plenty of room for heavy frames for bike at the rule-limited mass. And even with Cervelo, many of the riders chose the R3 (or R3-SL) over the SLC-SL, despite Cervelo super-engineer Gerard Vroomen's claims the SLC-SL was the far faster bike on virtually all terrain, even without a weight limit to equalize mass. The epitome of this was the 2008 Tour de France, where team leader Carlos Sastre flaunted the wind tunnel by riding the R3. He then won the yellow jersey.

The issue, obviously, was the feel of the bikes. Aero frames have a reputation for a poor trade-off between ride quality and stiffness. The Kestrels have a reputation for smooth ride, but poor stiffness. On the other side, the Ridley was said to be super-stiff, but an uncomfortable ride.

In the 2008 Olympics Cervelo released the S-series, with the pro-level frame being the S3. Now even more of the riders on Cervelo-sponsored teams were using this over the R-bike. It was the first aero-optimized frame to get relatively wide acceptance among pro riders.

But this was just the start. Scott revealed the F1 (later called Foil) at the 2010 Tour. The entire HTC-High Road team road the bike, with the exception of Mark Cavendish who switched back to his customized (extra beef) Addict bike after the first few stages. After the switch, Mark went on to win five of the final 15 mass-start stages.

But this year saw some of the biggest brands in cycling deliver the goods. First Specialized debuted the Venge at Milan San Remo, and the bike promptly won that epic classic with a rider, Matt Goss, who wasn't a favorite. All about the bike, right? Then Cervelo came out with the impressive R5, which looks a lot like a Felt AR but with a skinnier head tube.

Cervelo S5 from its most flattering view
Both frames showed impressive success in the Tour this year. Mark Cavendish rode his Venge to five stage wins, while Thor Hoshovd rode his S5 to two stage wins and a remarkable seven-day run in the yellow jersey (he started the Tour on an S3, but switched to a yellow S5 when he got the jersey). Hoshovd's reign seemed to validate Vroomen's claim of the benefits of aerodynamics on hilly terrain, as he against hope managed to twice cling onto lead groups of GC favorites to defend his GC lead against eventual winner Cadel Evans by just a few seconds.

So it seems the revolution was finally being won by the aero frames. After years of wind tunnel data showing that only a fool would take a round-tube bike into anything more competitive than a friendly bike commute, the pros were finally coming to their senses and the aero road frames were rising to a position of permanent dominance.

So when the dust had settled on the final sprint into Paris, how did they do in the final GC? In the Tour, top 20 is generally considered a noteworthy result: below that and a large number of the riders were burdened with other priorities, like team duties or focusing on stage wins, so actual placing is far less meaningful. So I'll focus on top 20. How many of the top 20 were on aero-optimized frames? Here's the result (from CyclingNews):

1  Cadel Evans (Aus) BMC Racing Team                86:12:22
2  Andy Schleck (Lux) Leopard Trek                  0:01:34
3  Fränk Schleck (Lux) Leopard Trek                 0:02:30
4  Thomas Voeckler (Fra) Team Europcar              0:03:20
5  Alberto Contador Velasco (Spa) Saxo Bank Sungard 0:03:57
6  Samuel Sanchez Gonzalez (Spa) Euskaltel-Euskadi  0:04:55
7  Damiano Cunego (Ita) Lampre - ISD                0:06:05
8  Ivan Basso (Ita) Liquigas-Cannondale             0:07:23
9  Thomas Danielson (USA) Team Garmin-Cervelo       0:08:15
10 Jean-Christophe Peraud (Fra) AG2R La Mondiale    0:10:11
11 Pierre Rolland (Fra) Team Europcar               0:10:43
12 Rein Taaramae (Est) Cofidis, Le Credit En Ligne  0:11:29
13 Kevin De Weert (Bel) Quickstep Cycling Team      0:16:29
14 Jerome Coppel (Fra) Saur - Sojasun               0:18:36
15 Arnold Jeannesson (Fra) FDJ                      0:21:20
16 Haimar Zubeldia Agirre (Spa) Team RadioShack     0:26:23
17 Christian Vande Velde (USA) Team Garmin-Cervelo  0:27:12
18 Ryder Hesjedal (Can) Team Garmin-Cervelo         0:27:14
19 Peter Velits (Svk) HTC-Highroad                  0:28:54
20 Jelle Vanendert (Bel) Omega Pharma-Lotto         0:32:41

The answer: none of the top 20 riders were on aero-optimized road frames (reference). Sure, the Garmin-Cervelo riders had the option of the S3 or S5, but Tom Danielson, Ryder Hesjdal, and Christian Vande Velde were all on R5's. And taking these riders as the prime contributors to their team's win in the team general classification, none of the top 3 teams in GC got there aero road frames. The team GC podium teams, Leopard-Trek and Europcar (on Colnagos) were on round-tube frames.

So what do you conclude from this? Two choices:
1. The fools! Every one of the top twenty would have been even faster had they seen the light of science and ridden aerodynamically optimized bikes. Wind tunnel data do not lie!
2. There's a lot more to bike racing than wind tunnel data.

It's remarkable, really. Every single test shows these bikes a substantial (on order 5%) reduction in wind drag from riding an aerodynamically optimized frame. Even if this is less important during crunch time on the climbs, the advantage, even while drafting, should be enough that it leaves riders fresher to deliver power during these critical periods. But the results simple don't show it.

Word is Cannondale and Trek will both enter the game soon. Maybe the aero-optimized frames will finally show their advantage. Or maybe the advantage is less than simple analysis suggests.

San Francisco "summer"

2011-08-06T21:14:00.000-07:00

It's been regularly chilly in San Francisco this past month. The marine layer blows fog in from the coast, and while it can be toasty warm in east of the Bay, on San Francisco streets sweaters, jackets, and scarves are common.

Here's a photo I took from the train on Sunday as the evening fog engulfs San Bruno mountain:

I wrote a little Perl script to download and process data archived on Weather Underground for Mission Bay, one of the warmest neighborhoods in San Francisco. Data are recorded every 15 minutes. For each of the years 2006-2011 for which data are available at that weather station, I took daytime (6 am to 9 pm) data from the period 08 July to 07 August (the past month as I write this) and binned the data between 6 am in 9 pm in 1 F bins. I then generated cumulative distribution plots for each year. So for each temperature I indicate the percentage of daytime samples which were less than the given temperature.

There's a clear pattern here. For 2006-2008, temperatures in the 70's where fairly common, with even some numbers reaching into the 80's. However, these last three years, 2009-2011, there've been no data in the '80's, and last year the temperature even failed to get out of the 60's during this one-month period. The median temperature, even at warm Mission Bay, was only 61F. This year it's almost 63F.

While the rest of the northern hemisphere laments the dog days of August marking the end of the hot summer, here the warmest months are September and October, when the fog has the least influence. Ironic that these hottest months mark the end of the road cycling season and the beginning of cyclocross.

So my standard dress on rides continues to be knickers or full tights with long-sleeve undershirts under my jersey with a vest. I actually enjoy the heat, at least to a limit, so I look forward to those September and October days when I can ride with just a jersey and shorts.

San Francisco Marathon encounter

2011-08-05T06:02:00.001-07:00

On Sunday morning I forgot the San Francisco Marathon was happening. It's easy to do -- there's not much said about it. The primary publicity is of inconvenience more than anything: car travel across the race route would be delayed.

My first encounter with watching a marathon was when I lived in the Boston area and watching The Marathon was virtually mandatory on Patriot's Day, the day which marks Massachussetts' emergence from a long, grey, cold winter. The huge number of runners in the race were dwarfed by the number of spectators, with many sections stacked on both sides with cheering viewers.

Later I lived in Austin, Texas, and while the marathon there wasn't in the same class as Austin, it was still a popular event, and people would follow the race and cheer the runners all along the fast route.

San Francisco, on the other hand, despite providing probably the best terrain for a marathon in the country, treats the race as a necessary evil to be gotten out of the way as painlessly as possible. With an ungodly early start on Sunday morning, well before most potential spectators are awake, earlier even than many East coasters are awake despite their three-zone head start, there's very few accidental pedestrian encounters with the race. The winners finish the race before most first-tier marathons even begin.

Austin in Texas starts bright & early @ 7 am. Chicago starts at 7:30 am. Boston starts at 10 am (elite men; women start at 9:30 am). And at New York the first wave starts at 9:40 am. Each of these races embraces spectators. In the case of Boston it's a virtual obsession.

At San Francisco the elites heads out into the darkness at 5:30 am. Spectators? Hardly any.

So back to Sunday -- I headed out on my bike at 10:56. I heard whistles ahead -- what was that? Oh, no -- the marathon, I realized; I'd forgotten!

Fortunately runners were still out there. These guys had been on the road for a bit more than 5 hours: the slowest wave at started at 6:32 am. I saw a mix. Some were jogging along at a steady pace, no doubt close to their goal. Others, perhaps from optimistically earlier waves, were asymmetrically limping, barely walking. I asked a runner how far to the finish: 3.5 miles, I was told.

After stopping in Whole Foods for some bars for my upcoming hike, I rode my bike along the sidewalk, upstream on the route. At an intersection a cop was waving traffic through. A runner waited to cross, glancing nervously at his watch. First one car, then another, than another crossed the intersection at the cop's encouragement. Then a big rig slowly, carefully, entered the intersection as it began a turn, perhaps to deliver a load to the market I'd recently left. The runner was still waiting.

I moved on. The irony of this was the intersection had been unsignaled. Without a light, pedestrians have right of way. On a normal training run he'd have been able to head right across after a brief check to make sure drivers were aware of him. And even if there'd been a light, the signal phases in San Francisco are mercifully short: he'd have been waiting far less time than he was forced to by the cop. And this was supposed to be the signature marathon of a major U.S. city?

Next intersection and the cop was more sympathetic to the runners. A few cars were waved by, then a group of runners apprached. She put out her arm, stopping the next group of cars. Another gap, and more cars were allowed to pass, but then a solo runner approached. She put out her arm, blowing her whistle. But this time, an approaching car didn't stop, instead accelerating and veering around her. Fortunately the runner hadn't trusted her right to pass, and waited.

This all made me very sad. Marathons are great events, a celebration of the human body, of dedication to training and perseverance. Yet San Francisco seems to view its event as an unfortunate obligation to be executed with as little inconvenience to Sunday morning drivers as possible. For example, it would not do at all if they were late to church.

The "complete streets" movement is premised on the paradigm shift that roads are too valuable a resource to restrict only to car use. Yet people view it as a violation of their fundamental human rights if they can't drive their vehicles where they want, when they want, as rapidly as their engines, handling and reflexes allow. Once again San Francisco shows it lags behind east coast cities like New York and Boston in this realization.

Strava urban KOM matching

2011-07-30T06:05:00.000-07:00

On July 12 I rode up Kansas Street on Potrero Hill in San Francisco, as someone else asked me to generate data for a segment there since my HTC Incredible phone has relatively good accuracy here. So I did that, and after defining the segment, I had the KOM. Not too surprising, since typically if any other Strava users had made a hard effort between the same two end-points they probably would have created the segment earlier. Here's that ride. As you can see, I was lucky to have beaten out Steve Smith's effort from May. He'd done a San Francisco climb-fest and had done Kansas only one second slower.

The issue on that ride was in addition to the Kansas climb Strava also gave me credit for Rhode Island, one block over. I had two KOMs, one block apart, from the same ride.

So yesterday since I ended up at the 4th and King Caltrain with my small backpack and no laptop, I decided to try and validate my Rhode Island "KOM". I wasn't quite sure where the end-points were (the embedded maps don't render on my phone), but I got up to speed where the climb started mid-way between 17th and 18th. This was my first mistake: the climb was defined to start right at the corner of 17th. Then at the intersection of 18th I stopped to wave a waiting car past. This cost me a bit more time. But the rest of the way was smooth going. I went all the way to Southern Heights, beyond which Rhode Island steeply descends.

I felt it was a good effort, and had a good chance of taking some time off that faux-KOM I'd set earlier. So when I got home I uploaded the data from my phone, went to the laptop, and checked the result.

I was 3 seconds short. But what I had done is "broken" my standing KOM up Kansas, and in fact set a KOM on De Haro, one block in the other direction. So instead of one bogus KOM, on Rhode Island, and a legitimate KOM on Kansas I now had three bogus KOM's: Kansas, Rhode Island, and De Haro.

Strava obviously needs to tighten its segment matching criteria, especially for urban segments where streets are densely located. Some areas, like the top of Old La Honda Road, have signal reflection problems where relatively large errors can result. But you can't set the matching threshold everywhere based on these worst-case locations.

Anyway, now I need to go do a hard ride up De Haro, as well. Fortunately Wisconsin, one block further, doesn't go through so I know there isn't a segment there I can contaminate.

Of course, the only reason I have KOM's on these climbs, valid or otherwise, is they're sufficiently obscure that none of the significant number of faster riders around here have bothered. But that's part of the beauty of Strava: it taps into the animal instinct to mark territory. There's more territory than the top dog can pee on, leaving space for even lesser dogs like me.

appendix: after posting this I went out and rode hard up De Haro. This fixed my bogus De Haro KOM, but also registered a new KOM up Rhode Island. So I think for these segments to be distinctly meaningful Strava simply needs to fix its matching algorithm.

VAMs on L'Alpe d'Huez in 2011 Tour de France

2011-07-28T19:20:00.000-07:00

This year the Tour de France climbed L'Alpe d'Huez, the "Old La Honda" of the Tour.

The "official" climbs is 13.8 km long according to Wikipedia, and gains 1099 meters according to Strava, an average grade of 7.96% (the GPS recorded distance depends on trajectory through the switchbacks, for example, and in any case GPS isn't that accurate on distance along curvy routes). It was climbed during the Tour de France stage 19 this year.

The climb was complicated by riders arriving at the start in small groups. Timing thus needed to consider the finishing time, well documented by the race, but also the starting time. According to Wikipedia:

Since 1999 photo-finish has been used from 14 km. Other times have been taken 13.8 km from the summit, which is the start of the climb. Others have been taken from the junction 700m from the start.

The following times were posted as "official" by the Inner Ring on Twitter:

I assume these are from the camera Wikipedia says is 14 km from the finish.

The Science of Sports Blog claims the following:

In terms of the historical context of the 2011 performances, the overall time for Contador was 41:30. Sammy Sanchez was the fastest of the day in 41:21, while Pierre Rolland, first to summit, did it in 41:52 because he started the climb with a 51 second lead (bear in mind small errors in timing off the TV).

He initiated timing at "the start banner of the climb".

The "official" times had Sanchez 41'45", Contador 41'54", and Rolland 42'22". These times lag Science of Sports' by 24", 24", and 30". If as I assume the difference is due to a 200 meter difference in start location than this corresponds to Contador and Rolland riding at 30 kph and Rolland at 24 kph, to crude precision.

So to convert the "official" times to the times for 13.8 km "chrono", I assume all riders other than Rolland rode the 200 meters at the same rate as Contador and Sanchez, taking 24 seconds. Perhaps this is optimistic for the 200 meters in many cases, but that makes it somewhat pessimistic for the time over the climb itself. However, assigning pack times to individuals, as is done here for both the start and finish of the climb, adds several seconds of error, so I would put all times here with ±5 second error bars.

So here's the results: I show rider, time, and VAM. I also calculated an estimated W/kg, but as I've described before, there's plenty of potential errors in such estimates, especially when riders draft each other. Jonathan Vaughters tweeted 5.4 W/kg for Tom Danielson, very close to my still air estimate of 5.51 W/kg, but then I read a claim this was revised to 5.7 W/kg, so take all that for what it's worth. My calculation has Sanchez at 5.67 W/kg, with numerous assumptions.

rider                  min   VAM
Samuel_Sanchez         41.35 1594.7
Alberto_Contador       41.50 1588.9
Pierre_Rolland         41.87 1575.0
Cadel_Evans            42.07 1567.5
Damiano_Cunego         42.07 1567.5
Andy_Schleck           42.07 1567.5
Frank_Schleck          42.07 1567.5
Peter_Velits           42.07 1567.5
Thomas_de_Gendt        42.07 1567.5
Tom_Danielson          42.37 1556.4
Jean-Christophe_Peraud 42.57 1549.1
Hubert_Dupont          43.22 1525.8
Rein_Taaramae          43.22 1525.8
Ryder_Hesjedal         43.22 1525.8
Ivan_Basso             43.22 1525.8
Levi_Leipheimer        43.22 1525.8
Jerome_Coppel          43.63 1511.2
Kevin_De_Weert         43.88 1502.6
Rob_Ruijgh             44.18 1492.4
Thomas_Voeckler        44.48 1482.4
Christian_Vandevelde   44.48 1482.4
Yury_Trofimov          44.48 1482.4
Arnold_Jeannesson      44.50 1481.8
Robert_Gesink          44.63 1477.4
Carlos_Barredo         44.68 1475.7
Blel_Kadri             44.90 1468.6
Remy_Di_Gregorio       45.55 1447.6
J_Blazquez_Hernandez   45.98 1434.0
Haimar_Zubeldia        46.12 1429.9
Richie_Port            46.38 1421.6
Chirs_Anker_Sorensen   46.38 1421.6
Sebastien_Minard       46.43 1420.1
Bauke_Mollema          46.43 1420.1
Christian_Knees        46.43 1420.1
Niki_Terpstra          46.72 1411.5
David_Loosli           46.80 1409.0
Egoi_Martinez          46.93 1405.0
Sandy_Cesar            46.93 1405.0
Gianni_Meersman        46.93 1405.0
Jakob_Fugelsang        46.93 1405.0
Jonathan_Hivert        47.90 1376.6
Jeremy_Roy             47.95 1375.2

If you told me a few years ago you could win the Tour with a 1570 VAM up L'Alpe d'Huez I'd have laughed at you. Sure, maybe in the Lemond-Indurain era, but not since the age of Armstrong, who revolutionized pro cycling by previewing climbs, training in the rain, and drinking Michelob Lite.

Quick plot of these results:

For perspective, the Strava KOM is 48.75 minutes, a VAM of 1353. Even in this kindler and gentler era, the pros humble us poor weekend warriors even after having climbed the brutal Telegraph and Galabier in week 3 of a grand tour. But this is expected: it's what they do, they're probing the limits of human ability. And maybe, just maybe, they're no longer exceeding them.

Coastal Trail Runs: Golden Gate 30 km

2011-07-26T17:07:00.000-07:00

My first 30 km trail running race of the year was yesterday: the Coastal Trail Runs Golden Gate.

The Marin Headlands provided for a gorgeous course, with four big climbs and one little climb at the end, along with miscellaneous rollers liberally sprinkled between the major bumps.

On Tue I had gotten stung by a bee while riding to work through Palo Alto. My face swelled up within an hour, but I got it under control with ice. When I awoke on Wed, however, the swelling was back in force, and Wed and Thu the allergic reaction left me fatigued. I felt a bit better on Friday, still a bit tired but not 100%. My lip was still puffy when I awoke early on Saturday, but otherwise I felt much better. I'd not run at all in a week and a half, but at least I'd gotten in some decent cycling through that Tuesday commute.

In April I'd done 26 km and 27 km training runs on many of the same trails used in the race, then on July 5 I did a strong 29 km run there. With the improved hydration and calorie intake available on a supported race, I felt I would be able to put in a good effort. The goal was to run every one of the climbs at tempo, and come into the finish fast but tired.

Starting (Cara Coburn)
So after overenthusiastically jumping the "gun" by a half-second (whoops!) I was second into the start of the first climb, following a guy in a red sleevless shirt who was just flying. He disappeared out of sight and I settled into run-walking my own pace up the hill. I was then passed by a woman who I tried to follow for awhile, but finally let go as she was going too quickly for what I thought I could hold for 30 km. I decided then to not worry too much about who was around me but run my own pace.

First climb (Cara Coburn)
One constant goal is to improve my speed on downhills. Sometimes on trails there's a forbidden speed gap: you can descend at a speed which allows for stopping or go fast enough it's more of a controlled fall. In between is the forbidden zone where you're braking to limit your speed, and that braking exceeds the available traction. I have a hard time crossing this gap: I like having the out available of being able to stop if necessary. The good runners seem to be unafraid of falling occasionally, and are willing to take that risk of committing to running past the speed gap. Once that commitment is made the only way to slow gracefully is to reach the end of the steep grade. So for short downhills, those where I know I won't get into trouble, I try to let gravity take me a bit further than I'm comfortable.

I seemed to be doing okay with this, not getting passed on the downhills. I reached the first checkpoint at Tennessee Valley stables in good position. The runners ahead, none of whom had water, blew through this stop. I had a bottle in my Nathan belt, which I was drinking while I ran, but stopped long enough to drink two cups of Hammer carbohydrate solution, which Coastal mixes nicely dilute. Even though it was foggy and cool I wanted to stay on top of my hydration. I then fumbled grabbing three Clif Blok carbohydrate chews, eating one and storing the remaining two in my belt pouch.

All of this allowed the runners I'd been chasing to increase their gap, but I was no longer worried about them. However, I worried a bit more when soon after the stop, on the flat Tennessee Valley trail, a group of around six guys ran past me as if they were doing a 10k. I slotted in with them, wanting to take advantage of the draft, figuring if I'd been faster on the first climb I'd be faster on the next.

Soon enough we turned onto the Fox Trail. I'd not been here before: it falls between the Coastal Trail which goes by Pirates Cove to the west and, to the east, the Miwok trail. Coastal-Miwok provides a fantastic loop, so I'd never felt compelled to take the relatively broad Fox trail. I'd asked about it at the start, however, and had been told it got steep toward the end. So as the climb began I settled into a sustainable pace, even as others in the group surged ahead, saving something for when it steepened.

And sure enough as the climb progressed the runners towards the front slowed, and I regained ground on then. As we reached the top of the climb where Fox merged with the Coastal Fire Road, I was with a group of two with two more further up the trail. We then began the descent towards the turnoff to the single-track Coastal Trail, the highlight of the course.

I was pleased I was able to descend with these two guys over some of the steeper sections of the fire road. With the coastal fog, we didn't see anything of the water to the left, but the valley to the right was clear enough well below our height. Eventually I saw some ribbons to the left with a sign: what did that say? But the runners I was following kept going, so I figured it was just a minor trail I hadn't noticed before. In retrospect, I can barely believe I'm writing this: I've been on these trails many times. But I have such an inferiority complex about my navigational abilities that I always assume without question that others who appear more confident are correct. So I followed the other two.

The descent got steeper and less familiar. Muir Beach was getting awfully close ahead, the view clearing as we descended. We'd missed the turn.

"This is the wrong way", I said.... "when's the last time you saw a ribbon?"

"Back there", one of the two responded, "but this is the right way: we're following those guys up ahead."

Lemming line... It's tempting when you commit to a course of action to stick to it, long after it becomes obvious it was a mistake. So even at this point I took a few more strides, a few more precious meters of altitude squandered, before stopping. I stopped.

A runner was approaching from behind. Despite overwhelming evidence, I still needed to check. "The Pirates Cove trail is back up the hill, right?" I asked. She confirmed.

So back up the hill I went. All of the early pushing of my pace was now flushed down the proverbial toilet. Those tenths of a second saved following optimal trajectories through corners were tossed away as I'd just given away ten minutes or more in a massive exhibition of neural flatulence.

As I ran, I was amused to see more runners with numbers pinned on their shirts or shorts descending. "Wrong way!" I shouted as I climbed. Yet once again in my surrealistic haze I started to doubt if I was making a mistake. It was as if my existing trail knowledge was all a strange dream.

Anton near Pirates Cove in 2010

But there it was ahead: the turn-off to Pirates Cove, marked with a striped ribbon. The marking wasn't super obvious: a single ribbon on what is a tricky corner. Maybe Wendell could have invested a few more here. But navigation is part of the trail running game, after all.

Even though I saw a runner in the distance, one who had obviously been behind me earlier, I decided to not let this snafu take me down. This, after all, is a trail unmatched in its beauty: a dream-like run along the rugged Pacific Coast. I'd specifically trained here to improve my speed on the undulating, twisting route, and it had worked. I felt like I had this thing nailed, and made great progress even up the steep trail which extends above the strength-sapping stairs which rise above the Cove. That section had kicked my butt more than once.

The trail gave way to fire road, and I ran at a good pace on the steep descent to Tennessee Valley.

The Tennessee Valley trail is a relative low point in several ways on this run. It's low in altitude, but being flat and broad with a very gradual uphill slope, it is slow going without much scenery for distraction, and there's enough cyclists and hikers to require dodging. Here the two guys I'd followed off the course repassed me. They complemented me on my speed on the single track, but here they were clearly that little bit faster.

Then the rest stop... a volunteer filled my bottle. The weight weenie in me tends to over-analyze these moment. I knew I didn't need the full bottle to reach the next rest stop, but then if filling it would allow me to skip refilling it later, perhaps that compensate for the increased time taken to carry the bottle up the Marincello trail. In retrospect I probably should have also factored in the increased fatigue from carrying it up the Marincello trail, and stopped half-way... But either way I wanted to err on the side of too much water. The plan was to down a quick cup of Coke, and two of water, at the stop, but not spend any time refilling.

We merged here with runners doing other courses: the half-marathon and the full marathon (on their first of two laps). They'd started a bit later but since we had run the Pirates Cove loop runners here with me were slower. This was nice, as passing people on the climb gave me the opportunity to share greetings with other runners, and, I have to admit, helped my confidence. From my own race I don't think I was passing anyone. This wasn't a great sign, I decided, but I still felt I was running well.

At the summit we merged with Bobcat trail. I ran a briefly with a 50 km runner (sharing the same course to this point). He said he could run our pace all day. I told him I was getting tired... I wasn't slow at this point, but I was getting a bit worried with there still plenty of running left to go.

Alta Trail, which follows Bobcat, is up-and-down single track along the ridge separating the Headlands from Sausalito and Highway 101. Normally this is stunning, but here the wind-blown fog condensed into actual rain, and the trail in spots was even muddy. If I were hiking I would have frozen wearing only my tights and my event "technical" T-shirt, but under these circumstances I was fine.

For part of this way I found myself running next to the woman who'd passed me on that first climb. I wondered how she'd gotten behind me, especially given my wrong turn (try not to think about that, I told myself!) Again she slipped past me. I also ran with some people who seemed to be on the marathon/half marathon course. I should have been passing them, but was content running at their speed. I was becoming increasingly invested in my plan to drink a Coke at the final stop. I was banking on that caffeine + corn syrup boost.

The aid station approached. I checked my bottle: 1/3 left, which should be enough, so returned it to my belt and shouted "Coke!" as I reached the table. "We don't have any Coke" the happy volunteer responded.

My world froze.

No Coke? But I always drink Coke at the last stop. Invariability it gives me a much-needed kick to crank up the pace over the final kilometers. I considered abandoning right there. But instead, I went for two water cups and a quick Clif Blok. I declined a third water cup, thanked the volunteers, and began my descent.

Flash back to last year. This descent nearly took me down. My legs were in pain the whole way as I hobbled down the long, gradual way. Steeper and I would have been forced to take short steps and it would have been fine. Flat and the impact is less so I would have been fine. But gradual descents are the worst.

This year was better, but not by much. First my left leg, then my right let me know they'd rather be supine, thank you. But of course I continued on, trying my best to distribute the load to the less sore bits. My shoes, which normally feel cushy, felt ridiculously thin.

I reached the turn-off to the Coastal Trail climb to Conzelman. This has been burned into my memory from having missed it during a race last year. I saw runners ahead of me make the turn. I followed.

The climb, albeit short, was a relief. I knew I was going slower than I had any climb to this point, but I was still "running", and that was good. I feared the top, however: Conzelman Road would mark the return of gradual descending, and far worse than the Coastal Trail descent, on asphalt instead of hard dirt.

Before descending Conzelman climbed gradually as it passed the military installations along the coast. This alone was painful, and I knew the descent would be worse. I managed to trot onward, trying to ignore the growing pain in my legs.

Over the top, and I was relieved to see the pavement disappear. Road construction here was unfinished, and instead of pavement I had soft dirt to cushion my footfalls. Trotting continued, and the discomfort was bearable.

Field Road arrived and with it a return to pavement. Here's where I cracked. Once entering the pavement, my trot turned into more of a shuffle, but when the pain in my legs was joined by a sharp stitch on my right side, I stopped and bent over, defeated. I started to walk, slowly, focusing on my breathing.

"Are you okay?" a woman asked as she passed, "I can get help!"

"Not really, but if I walk, I'll make it to the finish. The finish is close," I responded, fully engrossed in my own misery.

"I'll wait for you at the finish!" she replied, and continued.

I walked a few seconds more, then as the grade leveled, decided to try running a bit. And that worked: slowly, barely running, but clearly better than walking.

We briefly entered the Rodeo Lagoon trail before turning off to the right to descend the staircase to Bunker Road.

Ouch... this hurt, as I took the steps slowly, one at a time. Then I heard behind me two women talking: "It's funny how my left leg feels unstable". I laughed and shouted back to them: "you, too?" That was exactly how I felt as I took each step, one by one.

I reached the road and this was the final stretch to the finish. No sprint for me, but I was able to run at a shuffle pace, and even caught the woman who'd said she was going to wait for me. I thanked her for the inspiration.

Cara was at the turn to the final meters before the finish... she's said she might leave after completing her hike, but she'd waited, and that made me happy. I turned the corner, ran to the finish line, and I was done... completely. I collapsed onto the ground and lay there, not wanting to move.

Finish (Cara Coburn)
Cara told me my time was 3:08, not much off my 3 hour target (official time was 3:08:30). I was amazed by this, as the wrong turn alone had clearly cost me more than 8 minutes, and my collapse toward the finish had been rather dramatic. When results were posted, I was 18th overall, not too bad at all, although I'd been hoping for a top 10.

But running's more about the feeling than it is about the hard numbers. That feeling of having delivered the best possible effort, to overcome discomfort and arrive at the finish strongly but with nothing more to give, is intoxicating. Here I'd failed in reaching that goal. Of course the extra distance from missing that corner didn't help: the extra km get tacked onto the end of a race, not the middle, and there was a very high rate of interest paid on that extra bit of running. Still, I finished, and placed fairly well. It just left me wishing I could try it all again: not go into the race is a partially weakened state, and not trust people to know the way just because of neurotic navigational inferiority complex. It's ironic that of the four Coastal Trail Runs I've done, on the three on courses I've "known", I've taken wrong turns (in two cases following people), while on the one on a course I didn't know, I navigated without error.

The run left me feeling totally wasted. My muscles sore, my energy spent. Here it is Tuesday, three days later, and I'm finally feeling a bit better. At least I can start walking from seated without a limp. Maybe I'll ride my bike to work tomorrow.

Time Gaps and Km in the Tour de France

2011-07-21T20:25:00.000-07:00

In its live coverage of stage 18 of the Tour de France today, CyclingNews posted the following update:

17:35:28 CEST
Andy Schleck powers under the red kite, with his grimace beginning to curl into a grim. Over 60km off the front for the Luxembourger.
17:35:56 CEST
Schleck has a shade under 3 minutes over Evans, but 3:30 over Contador.
17:36:34 CEST
Pierre Rolland sets the pace for Voeckler. The Frenchman is battling to hold on to his yellow jersey here.
17:36:58 CEST
Voeckler needs to close the gap to 2:35 to hold the yellow jersey.
17:37:24 CEST
Evans and Voeckler lead the yellow jersey group in pursuit of Schleck.
17:38:19 CEST
Schleck appears to have slowed in the final kilometre, his efforts have finally begun to tell in this steep upper section of the climb. He should take yellow by a handful of seconds, but it might be close.

When it was all done, despite being dropped by several of the riders in his group, Voeckler managed to close the gap to 2:21, saving his yellow jersey by 15 seconds.

So what was CyclingNews' mistake? How could they, watching the race on television, so overestimate how much time Andy had on the chasers?

The answer is likely the way time gaps are reported. Andy rides under the kite, the screen says "1.0 km" and next to it a time gap, for example "3:00" (indicated in the report). However obviously at this point the time gap does not apply to 1 km to go: that gap is impossible to determine since the chasers haven't reached that point yet. Instead the time gap is no closer to the finish than the chasers' current position, and with a 3 minute time gap, that's more than 2 km from the finish, not 1 km.

Science of Sport timed the actual gap at 1 km and it was 2:38. So between the point where Andy crossed the 1 km to go kite, when the gap was 3 minutes, and when the chasers crossed the same point the gap had fallen by 22 seconds. The gap was to drop another 21 seconds before Frank Schleck crossed the line in second place, 2:07 back.

So CyclingNews simply underestimated, by close to a factor of two, how much road Voeckler had to reduce Andy's hold on "la meilleur jeune virtuelle".

As an aside, a still air estimate for the power produced by Frank Schleck on the climb of the Galabier was 5.7 W/kg (Science of Sport blog comments). There was reported to be a headwind, but Frank was sitting in the group for most of the climb, protecting his brother's lead. Evans was probably a bit higher than this. But it's all further proof that riders blowing the top off 6 W/kg late in a stage race is, for now, apparently a thing of the past. And just in time for Voeckler, who finally gets to show his real talent. Of course the proof will come tomorrow, when the riders climb L'Alpe d'Huez, sort of the Old La Honda Road of the Tour as far as times are concerned.

Voeckler on Plateau de Beille: 2004 and 2011

2011-07-18T21:31:00.000-07:00

A lot has been said and written about Voeckler's remarkable climb to Plateau de Beille in stage 14 of this year's Tour. Voeckler had an advantage of 1:49 over second-place Fränk Schleck, 2:06 over Cadel Evans, and 2:18 over Fränk's brother and teammate Andy Schleck. Due to time and fitness lost in a crash, race favorite Alberto Contador was 4 minutes down. Any of these riders were considered a threat to overtake Voeckler, although the consensus most likely scenario was Voeckler would limit his time losses and hold onto the yellow jersey by maybe 30 seconds.

Strava KOM: around a half-hour slower than Voeckler's time.

Yet it's safe to say it was a shock to almost everyone when Voeckler not only held on to his lead over all other contendors for the yellow, but was a primary activist in chasing down attacks. I'd have expected Voeckler, by any reasonable measure outmatched in that group, to take a very conservative approach to the climb, following the smoothest wheel to avoid going into the red for even a few seconds. Cadel Evans and Alberto Contador are both exceptional in their ability to put on short bursts of amazing climbing speed. So if Andy Schleck attacked, Voeckler could have let one of these riders close the gap, perhaps following Basso who would steadily bring it back together afterwards. But instead Voeckler was sprinting back to Andy's wheel himself.

Voeckler's special motivation on the climb.

Meanwhile, Andy was tempering his attacks based on what was evident weakness in his brother. Fränk obviously wasn't feeling as frisky, and the risk was an attack by Andy could result in nothing more than Fränk getting dropped from the group. That would hardly serve his team's interest. Of course, best of all would be for Andy to attack and get away solo, something a lot of his fans obviously wanted to see, but Andy had a good view of how his brother and the others appeared, and if Fränk was a weak link, he may have realized the team would do better by staying united at least until the upcoming Alpine climbs.

All of this start-and-stop worked to Voeckler's favor. Voeckler is also a punchy rider with a strong top-end, but where he'd have most likely succumbed would have been a faster, steady pace. The lack of cooperation among the favorites also allowed first Jelle Vanendert and later Samuel Sanchez to ride away from the group -- Jelle is out of overall contention and Samuel, although now 6th overall, is a poor time trialist and so still needs to gain considerable time in the mountains to be considered a threat. So by cooperating the favorites obviously could have climbed faster, likely faster than Voeckler could have sustained. But dropping Voeckler wasn't the priority, since Voeckler also is considered weak in the time trial, and so I don't believe the favorites consider him a likely threat to still lead in Paris.

Science of Sport plot of great VAMs in Tour history. Voeckler's 1600 on Plateau de Beille is well off the chart.

This is all deja-vu for Voeckler, who also rode to defend his yellow jersey seven years ago on the same climb. He finished that stage 13th, 4:42 behind Lance Armstrong and Ivan Basso, to just hold on to his yellow.

So what is it that allowed Voeckler to go from a rider who got dumped by close to 5 minutes by the favorites to one who was in the mix, matching their accelerations to the end?

Michele Ferrari named and popularized the use of the VAM metric: the rate of altitude gain of a rider on a climb, typically measured in meters of altitude gained per hour. It's a decent way to objectively compare rider power/mass ratios, especially when comparing climbs of a similar grade and altitude difference.

The results: Vanendert finished the climb in 46:01, a VAM of 1627 m/h. Samuel Sanchez was 20 seconds slower, then the favorites finished Samuel Sanchez was not far behind (20") and the group of favorites did it at 1600m/h (5.75 w/kg), in 46:48.

In contrast, according to Wikipedia Lance Armstrong reportedly did the climb in 45:30 in 2004, and Marco Pantani climbed it in a remarkable 43:30 in 1998. In 2004, Voeckler finished 4:42 behind Armstrong, making his time 50:12.

So it appears Voeckler was able to take 3:24 out of his time from seven years ago. But the stage in 2004 was 206 km ridden at 33.8 kph, with seven rated climbs:

In contrast, this year's stage, while still extremely tough, was clearly easier: 168 km with six rated climbs (CyclingNews) ridden at 32.2 kph. That's 1.6 kph slower than the 2004 stage, which had an extra category 3 climb. So it's reasonable Voeckler would be faster this time.

It's no surprise the speed was higher in 2004. The finishing list looks likes a who's-whom in mid-naughts doping scandles. Of the twelve riders finishing ahead of Voeckler only the last, 12th-place Stéphane Goubert, has not been involved in a major doping scandal. I think it's safe at this point to include the stage winner in the "involved" category, given the consistency and number and specificity of charges against him. So it's really remarkable Voeckler was able to hang with that crew as long as he was without totally collapsing on that final beyond-category climb.

Anyway, Voeckler's weakness is his time trial, and thus I think he has little chance to win the yellow when the Tour arrives in Paris. And while a VAM of 1600 was enough on the tactical shadow-boxing on Plateau de Beille, it's unlikely he's going to be able to follow a more sustained attack during the back-to-back big stages in the Alpes.

added : Cyclocosm has an excellent comparison of times, which differ slightly from Ferrari's (it depends on if you time the group or the individual riders, since it takes a significant time for the pack to cross the threshold of the climb).