Tuesday, March 31, 2009

reflections on March

I had a really solid sprint workout today, extending my maximum recorded power to 749 watts on sprint #4. Then 737 watts on sprint #5 was my second best. So maybe my recent series of sprint workouts is paying off.

With this ride, I closed off what's been a very mixed month. On the surface, with the exception of today's ride, it's been discouraging. I was 73rd in the Menlo Grand Prix, never a factor. I've never once this month felt strong on climbs. I got dumped from my one and only Spectrum Ride (okay, I ran 12.3 miles the day before). I did manage to hang with the main group over Canada College on a Thursday Noon Ride, but on a Tue Nooner, I was overwhelmed before I dropped off to do my assigned sprints (I'd done a long run two days before). I've not done a single ride this month much over 45 miles.

So when things don't seem to be going so well, it's nice to look at some data. Here's my maximal power curves for the month (before and after today's sprint workout), compared with Feb and Mar of last year.


maximal power curve comparison. PDF version also available.


What do I see here? Well, not so bad, really. Starting from the left, clearly sprint power is up. Moving onward, there's a gaping hole in the range of VOMax intervals, 3 to 8 minutes (180 to 480 seconds). I was doing those last March, but I'm not yet there this year. Towards 40 minutes, the Menlo Grand Prix yielded a sizable plateau. Not shown are my soporific data from climbing Old La Honda or West Alpine, which fell well below the Menlo numbers. Further to the right I see this year's data falls off below the 2008 curves. I was doing more long, hard rides last year. To a certain degree these have suffered due to the addition of running to my scheduler: I've done several runs approaching or exceeding the half-marathon distance these past few months, and am still recovering from straining my hamstring during the Woodside Half Marathon.

Also shown in the plot is an estimated maximal power curve assuming FTP = 280 watts (5.02 W/kg on a Powertap ≈ 5.17 W/kg on an SRM) with a ratio of anaerobic work capacity to critical power of 60 seconds. These would be fairly good numbers for me, and I may not be there yet, especially the AWC component. The Menlo data fall close to this curve, which is impressive given how non-uniform the efforts were there (the "normalized power" curve, which is claimed to better predict FTP in this time range, actually exceeds this "optimistic" curve, but I prefer to stick with average power). Last year my VOMax intervals come close to this curve. It will be interesting if I can match last year's numbers when I add these to my menu this year, no doubt soon.

So for April: I absolutely must get in some longer rides. With sickness, moving, and fatigue from running, my training progression substantially lags last year. And there's that uncertain benefit from all of those running miles.

Monday, March 30, 2009

Phinney's amazing kilometer

Last Friday, I needed a break. I decided to work from home, but I honestly wasn't getting much work done. Burning it at three ends. I was tired.

Tyler Phinney was going to be riding the kilometer at Worlds. I checked CyclingFans for on-line coverage, and found a link to Justin.tv. Bingo. I tuned in to some French coverage just a few riders before Phinney clipped in. The best time so far: an insane 1:00.666 by German Stefan Nimke.


Mini-Phinney riding the kilometer like a mini-pursuit (VeloNews)


Phinney was behind his pace at 200 meters. These guys generally rip out killer standing 200's at the start of a kilo, but Tyler's just not that sort of a rider, especially at 18 years old. But once he was up to speed.... he lost a small bit of time next check, then actually began to gain some back. He crossed the finish in an amazing 1:01:611. Wow! Second place!

One rider after another came to the start, bettered Nimke's opening 200 meter time, but then faded badly in second half of the event. Nobody could match Nimke's kilometer time, but more significantly, every single one was coming in behind Phinney. Malaysian Mohd Tisin caught me by surprise, coming within 60 milliseconds, but nobody else broke 1:02. Phinney, in an event his mother said he was riding "for fun", had silver.

Analysis of official splits, taken at 250 meter increments, reveals an interesting story. The following plots compare on the x-axis the opening 250 meter time to, on the y-axis, either the total time (top), the final 750 meter time (middle), or the final 500 meter time (bottom). Click on the image for a PDF version.


Analysis of split times.
Phinney is circled yellow, Nimke green, Tisin in red. Phinney demolished the second 500 meters.


The remarkable thing is each of the top 3 riders were at best mediocre, or in the case of Phinney, DFL by a wide margin for the opening 250. Phinney and Nimke each averaged over 64 kph over the final three-quarters of the race: 40 miles per hour in round numbers. That's a nice downhill sprint for me. I can't come close to touching it on a flat road, let alone sustain it for 42 seconds after a kilowatt standing start.

On Wattage I wondered why riders appeared to go out too hard. After all, in these data, the correlation coefficient of each of the middle two splits with the final result was substantially stronger than that of either the opening or closing 250 meter splits, even after eliminating the top three. Andrew Coggan pointed out riders go out at a schedule required to medal, rather than the schedule they estimate they can sustain, as they are gambling on a super-day, rather than a typical day. It's all or nothing at this level. It makes sense. But the top three guys, it seems, didn't need to gamble. This is particularly true for Phinney, who already had his rainbow jersey from the previous day's pursuit. Each of the three rode his race, and scored.

Were Phinney to stick with this thing much longer, it's hard to believe he can't find that extra 1.7 seconds to take himself below the minute.

Friday afternoon, I was so excited about Phinney's ride I went out and did a 12.2 mile run. Not the best recovery aid: I had nothing left in the tank for Saturday's Spectrum Ride. But that's another topic.

Friday, March 27, 2009

Cadence

Today's workout was a recovery ride with two high cadence drills, each consisting of the following:
  1. 5 min @ 100 rpm
  2. 4 min @ 105 rpm
  3. 3 min @ 110 rpm
  4. 2 min @ 115 rpm
  5. 1 min @ 120 rpm

Obviously this suggests the need for a cadence monitor. As I mentioned in my last post, the Powertap does make an attempt at cadence by monitoring the periodicity of the torque measured at the rear hub, expecting to see torque fluctuate as first one foot then the other foot moves between the downstroke. However this doesn't work as nicely as a good old-fashioned magnetic reed switch mounted to a chainstay to pick up the field from a magnet taped to a crank arm, and Saris will even sell you one for $50. For moderate cadence, the Powertap's scheme works okay, but for spinning drills, it tends to get confused and underestimate. But even at moderate cadence, the number fluctuates, such that trying to maintain cadence by watching the number isn't very productive. Smoothing helps reduce the fluctuations, but it also delays the feedback received from changes in cadence. So the best thing here is to mostly go by feel: for each interval, sustain a cadence I feel I can sustain for that interval.

I started my ride meeting the Noon Rider crew, providing a convenient deadline. Then I "warmed up" by riding with them over Arastradero before letting them go as they charged up Alpine. Okay, I admit I didn't need to warm up for a recovery ride, but the Noon Ride is addictive, and even a small fix helps. Then I began the business at hand.

My first set took be to the Portola Road - Sand Hill Road intersection. After a break, I started my second set on Tripp Road, finishing it where Whisky Hill T's into Sand Hill Road. I couldn't sustain continuous high cadence: turns or cars at times got in the way. But when I was pedaling, I kept the rpm's high.

Here's some analysis of the data. After pruning out the points for cadence below 45 (coasting, or contaminated by diluted by coasting during the sampling period), I average my measured cadence for each period, along with the total pedaling time for each period. I do that for each of the two sets.

RPM vs interval time
RPM for each interval vs. pedaling time

Since I am confident my last (and shortest) intervals were the highest cadence, and well above the listed values, the Powertap appears to be having problems with cadences over 120. This is consistent with my prior experience, and in fact with the Saris FAQ, which claims an upper limit of 130 (since cadence naturally fluctuates, this would show up as a reduction in average cadence for reported average cadences modestly below 130).

I was glad when the second set was done. High-cadence work is hard. On that final one-minute interval, I try to keep it on the verge of bounding on the saddle. On group rides, Coach Dan likes to show his spin, typical of those who've spent productive time at Hellyer. The ability to sustain power at cadence is non-trivial. Not only does it help in quick accelerations, but for descents and/or tailwinds it allows one to generate power to higher speeds. I was dropped at the Patterson Pass Road race a few years ago when I found myself gapped off the back on a gentle descent with a raging tailwind. I was giving my 53/12 everything I had, but I just couldn't close that gap. Better neuromuscular coordination at high rpm's could well have made the difference (so could have not getting gapped in the first place! But that's another matter.)

Rather than simply spin back to work, I finished my ride with one-legged pedaling. "Coach Troy" Jacobson's voice rang through my brain: "Pull up on the top of your shoe! Pull up on the top of your shoe!" One-legged pedaling is a prominent component of the Spinervals 23.0 technique workout.

So another virtuous day on the bike, but a hard day on the bike, one which had me sometimes tempted to return to the seductively simple suffering of the Noon Ride.

P.S. You'd think the Powertap would report half cadence for one-legged pedaling. But clearly it wasn't doing so yesterday: reported numbers during my single-legged drills were in the 60 range, right where I'd expect. It would be interesting to see the algorithm it uses.

Thursday, March 26, 2009

SportVelo ride

This week, an embellishment on last week's Z3 climbing: Z3 with 30 seconds at Z4 every two minutes, to get my body accustomed to being in Z4. A better approach to my noon ride "Z4 and blow" paradigm of a few weeks ago.

Here's the plot. The intervals on the climbs are fairly obvious. The bottom of Alpine is a gradual grade to flat, so for the first 8 minutes of that climb I tried to just keep it Z3, introducing the Z4 component when the climb proper kicked in.


Power over time for the ride.


I focus on on the climbs in the next plot. Here instead of ride time, the x-axis is the time within the interval. I did a fairly good job, it seems, of boosting the power up into Z4 during the final 30 seconds of each. There's no smoothing on these values, and point-to-point powertap samples jump around a lot due to differences in pedal stroke phase.


Power over interval time for the major climbs.


At Old La Honda Road, Coach Dan recommended doing that climb big gear, low cadence (high torque). Since low cadence Z4 is a bit much, I decided to do the 90 second Z3 phases in a slightly oversized gear, then shift down for the Z4 30 second phases, which I targeted at a more optimal cadence. On Alpine, I just rode a comfortable cadence in each zone. The powertap estimates cadence based on the periodicity of its torque signal, so I can plot torque versus power during these intervals.


Power versus cadence during the major climbs.


I can see that I tended to climb within a band from 60 to 80 rpm, but during the big-gear climbing on Old La Honda, I was in a band from 40 to 60 rpm. This resulted in higher force on the pedals during the Z3 segments than during the Z4 segments, despite the lower power.

Overall, a very solid workout. It always feels virtuous and productive to do something structured, rather than just go out and thrash myself in a chaos of masochism. Training is a balance between stimulating physiological adaptation and avoiding long-term fatigue. By carefully rationing the intensity, long-term progress should be improved, even if the short-term training stimulus might be a bit less.

Monday, March 23, 2009

Yoga Tree

forward bendA friend was visiting town and wanted to do yoga yesterday in lieu of our planned hike in the Marin Headlands. Too windy, she said, which I'd thought was a bit strange from someone who lives in Seattle. But yoga was fine. It'd been months since I'd attended my previous class, at World Gym at 16th and De Haro. The classes at World Gym are surprisingly good considering its reputation for a muscle-bound clientèle, but I wanted to take her to the a place with more of that distinctive yoga ambiance. I had a long-neglected gift certificate to Yoga Tree, so it was time to check it out.

As our 11 am Vinyasa session was about to begin, the instructor declared we had "the perfect number" of students. "Plenty of room to walk up and down the aisles," he gleefully observed. I glanced around; the place was packed -- it had to be close to its 60-student capacity. Hopefully I wouldn't topple over on any of the balancings, initiating a domino-like chain reaction.

And so we began. First, some gentle openers, then too quickly to more "flow" to "build heat".... and build it did. I soon regretted my long-sleeve cotton T-shirt I'd chosen against the blustery day outside. Sweat was dropping onto my mat with a frequency which reminded me of my now long-past 40C-regulated Bikram days ("Bikram is the complete opposite of what we teach here", I'd earlier heard a Yoga Tree employee comment).

A 90-minute class full of enthusiastic practitioners is a long jump from my usual 10-minute home sessions. I especially struggle with the one-legged balancings. Not only am I severely balance-challenged (I learned this at a young age during my ill-fated attempts at ice skating on New Jersey ponds, back when ponds still froze in New Jersey), but my cycling-fatigued legs, accustomed to 70 to 110 rpms, don't take isometric loading happily.

But other than my wobbly attempts at the balancings, I somehow survived. I especially appreciated the help from the several assistants who provided gentle corrections, or in some cases major corrections ("you want to be on the other leg!", somehow delivered without laughter), to my postures.

After the 3rd "final" asana, we sunk into the deep relaxation of a well-earned savasana. Eyes gently closed, I even received a brief, unseen neck massage from one of the assistants, immediately identifying tension I'd been unable to isolate on my own.

Overall a nicely intense experience. There's plenty of yoga options in the Mission, and certainly I won't discourage anyone from the underutilized classes at World Gym. But Yoga Tree is definitely worth the visit. Even my Seattle friend was impressed.

P.S. One thing I'd worried about was how my left leg would feel. I'd tweaked something during the Woodside Half-Marathon trail run through Huddart Park, I think in my hamstring, and it was still bothering me. It's been showing itself during hamstring stretches, not so much in reduced flexibility but in a harder time reaching full flexibility, and during the class, I wasn't quite as loose in the left-sided Pigeon and had to be supplied with a supporting towel to ground my hip. Things continue to improve there, however, and Daryl told me during my massage the day before I just need to give it time.

added: I just remembered my favorite "new" asana in the class, The Bird of Paradise. I was sort of able to do it.

Sunday, March 22, 2009

Tale of Two Fridays

Friday's Noon Ride seemed fast to me. The nice thing about power data is they allow one to put a quantitative validation on such qualitative assessments. One of my favorite ways of looking at ride data are maximal power curves. These are the maximum average power recorded during the ride over intervals of various duration. Here I compare average powers over four rides: three noon rides and the Menlo Park Grand Prix criterium, for durations from 0.05 minutes (3 seconds) up to an hour:

maximal power curves
Maximal power curves for two Friday Noon Rides, a Thursday Noon Ride, and the Menlo Park Grand Prix. FTP estimated from the Menlo Park Grand Prix normalized power is also shown, along with a curve using the Critical Power model for AWC/CP = 40 seconds with this FTP value.


On the plot, I point out a key feature of that Thursday ride: the hard effort of the Canadã College loop, which is a pair of climbs separated by the descent. Note the peak average power increases with increasing time for a bit here. This is a characteristic feature of two hard efforts separated by a brief recovery.

I also "sketch in" a "theoretical maximum" curve generated with the elegantly simple Critical Power model. This model assumes there is a fixed "Anaerobic Work Capacity" (AWC) which can be spent over a certain interval of time, then is slowly regenerated. The ratio of AWC to Critical Power (CP) is units of time. It takes a certain minimum amount of time to utilize AWC, so I sketch the curve only for times of one minute and higher. The effort of Canada College fails to reach this curve, because the curve can only be reached for efforts of approximately constant power, while the Canadã College interval was far from constant power. The same is true of Menlo Park, which had hard accelerations separated by drafting and braking (way too much braking due to my poor pack position).

The comparison of interest here is the two Friday Noon Rides. For each power interval, I worked harder during this week's Noon Ride than the one in February. Warm weather, later in the season, and folks are feeling friskier. I put in a little attack myself, just to make life harder on myself.

P.S. Nice discussion here of the Roaster Ride Sting last weekend. What a total joke. My money is on these tickets getting tossed out upon challenge.

Sunday, March 15, 2009

1998 Low-Key Hillclimbs

Two good things today....

One, I got in an 11.6 mile run, my first serious run since the trail run two weeks ago. The descent through Huddart Park took its toll, and I spent two weeks focused on cycling, stretching, and core strength work. It's important to me to maintain some running fitness through the summer. This will give me a big head start in my goal of qualifying for Boston, maybe at Sacramento in December, right before IEDM, which I often attend.

But additionally I think running is good cross-training for riding, improving muscular balance, bone strength, and core integrity. Running also increases my focus on relaxation and posture, focuses which can be beneficially translated to riding. For example, relaxing my shoulders on the bike drops my head, which lowers wind resistance. And I think the body just works better when it's relaxed than when it's tense. Running has helped me realize just how much I tend to tense up my shoulders and neck. As if that wasn't already obvious from the fact Daryl gets out his hammer and chisel every time I show up for a massage at World Gym...

The other good thing today is I finished "Project 1998" of the Low-Key Hillclimbs. Dean Larson had some old printouts of Giorgio Cosentino's web pages of the 1998 series, pages which were lost when hooked.net shut down. Ron Brunner scanned the pages, and I, Pat Parseghian, and Fred Butts entered the data into spreadsheets. Then I modified my circa 2009 scoring and web page generating Perl scripts to some of the scoring conventions used by 1998 scorekeeper Tracy Colwell. After a bit of debugging, viola, 1998 data are back up.

1995 and 1996 were already complete, salvaged from the WayBack machine. I have most of the results for 1997 on-line, but a few weeks are missing. Maybe someone else out there has the data. Crazy idea? Well, I'd thought the same about 1998.

Saturday, March 14, 2009

Z3, part 2

Wed I blew it. But Thursday I was serious. Coach Dan again said Z3, so it'd be Z3. The goal: ride segments of Z3 approximately 10 to 20 minutes long.

But first, I'd just "warm up" with the start of the Noon Ride. Nothing hard. Just "sit in". Nice and easy. Trust me. Really.

Fortunately, I finished patching my GoldenCheetah command line utilities, so I'm awash in data once again:

Thursday ride data
Ride data
from Thursday, with power zones (estimated from Menlo Park normalized power) indicated on right

Okay, so much for "nice and easy". I tried, really I did. When the group got strung out turning onto Albion due to a pesky car (who's the guy responsible for getting the road closure?), I cashed in what position chips I had to drift to the back of the pack, reducing the compression interval. A bit later, the crew turned off for some Z6 action on the Canada College loop. I split off, going out and back on Canada Road instead.

power histogram
Histogram
of data, showing a much tighter, bimodal distribution when riding solo. Note the logarithmic axis.

Nevertheless, you can see the Nooner involved plenty of time well above the Z3 target zone. And, for that matter, plenty of time in Z1 and Z2. Not even close to the "sweet spot training" which followed. At least the best approximation I could achieve of steady Z3 on rolling Canada Road (Z3 is a small target).

The Noon Ride is a great catalyst. Without that very hard deadline I have difficulty getting myself out of the office. Deadlines can be quite useful. And, of course, the social aspect of the group ride is a wonderful thing. Everyone united by their love of riding, a love not necessarily appreciated by others.

But I probably should have split off on my own sooner.

Friday, March 13, 2009

Menlo Grand Prix power data

I managed to port Ned Harding's Golden Cheetah patch for the new Powertap raw data format over to the command line utilities I prefer using, and finally got the chance to look at the numbers from my recent rides. A comparison: the Menlo Park Grand Prix to the Wednesday Noon Ride....

click on me for PDF file
Powertap data, Menlo GP and OLH. PDF version also available.


Interesting results. First of all, I simply didn't think I could produce that sort of power this year. Average power of 271 watts is bigger than what I've managed on any of my climbs of Old La Honda so far. And average power doesn't tell the whole story: normalized power gives higher weight to power above the mean than to power below the mean. It attempts to describe an "effective power": the constant power which could be produced with comparable effort. The number for the criterium was 289 watts for close to 38 minutes. That's almost as good as my best normalized power last year (291 watts) up Old La Honda, an effort less than half as long. So perhaps that is encouraging.

Menlo Park GP finish
That's me back in 73rd place (Shaun Baesman photo)


But another message is gained by comparing the numbers to last year's Menlo Grand Prix. That race was on a simpler 4-corner course. Power numbers there were substantially less for a similar average speed: 212 watt average power and 234 watt normalized power.

So how to explain this? All I can suggest is the addition of the "S-turn" resulted in more surging on this year's race. My relatively poor pack position, near the back the whole race, meant I spent more time on the brakes, more time accelerating than the riders in the front who could get through the turns with minimal wasted energy. The moral? Ride near the front, you dummy!

The comparison with Old La Honda is also interesting. After Menlo on Sunday, I did a rather mediocre sprint workout on Tuesday. Then on Wed, I was supposed to climb in Z3, according to Coach Dan Smith. Instead I succumbed to Noon Ride Syndrome (NRS). This is quite similar to ARS, "A-Ride Syndrome". Basically it means every ride becomes a race. Two guys go out hard.... I follow until I realize I can't possibly sustain that pace, and instead settle into a 290 watt target. Not crazy, right? Basically my NP for Menlo. But I just didn't have anything in the tank, and began fading badly. I was caught by Greg McQuaid and James Porter, who were chatting effortlessly as they rode besides me. Finally Greg put in a stealth attach (keeping constant speed when the grade increases) and dropped me like a bad habit. James eventually realized I was going nowhere fast and also left me behind as well. I finally arrived at the summit in 19:49, 2:46 slower than my best time, yet ready to be airlifted to the nearest emergency care facility.

So much for "sweet spot training".

Anyway, this experience reinforced my determination to stick with the program. If Coach Dan says Z3, not Z5, he means Z3, not Z5. Higher power is better training, right? Well, of course. Except it compromises recovery, which in the long term results in less fitness acquisition. It's important to keep an eye on the long-range goal, not go for that OLH PR every time out.

Wednesday, March 11, 2009

Powertap raw data format change

Powertap's gone ANT+ Sport. So when I sent Saris my previous head unit for replacement when it magically stopped functioning, they send me one with ANT+ Sport enabled firmware. All good: the firmware is still compatible with old hubs. But when I cranked up my venerable GoldenCheetah ptdl and ptunpk command line utilities to download and decode the data, they spit it back out. Saris had changed the "raw" file format.

GoldenCheetahI reported this to the GoldenCheetah user's list. Curiously, my report was the same day Ned Harding reported the same problem (this time in the context of a Windows version of the code. Windows is a virus my system doesn't have: I run Ubuntu 8.10. One month until 9.04! Hold you're breath. There's always issues with "upgrades".). The same day, Ned was able to reverse engineer the new binary format, and posted a patch. Wow. I consider myself a coding hack, but when guys who really know what they're doing show their stuff, I am impressed.

Anyway, I should be able to examine my Menlo Park data within a few days. Yeah, there's Saris PowerAgent, but it won't run in the presence of USB drivers required by GoldenCheetah. And PowerAgent has some really poor features, like plots with severely decimated data and a data merging algorithm which splices active time segments as if inactive time segments had never occurred (two hard 5 minute efforts separated by a minute of rest is not the same as a single 10 minute effort). Anyway, I prefer my own Perl scripts for doing my number crunching. And the GoldenCheetah command line utilities allow me to script without any sticky GUIs.

GoldenCheetah also has a GUI, of course. Actually, my favored command line utilities have been omitted from releases since 210, so I have some work to do porting the patch back to that version. I'm at least good enough for that, though. But for my purpose, the GUI still has a way to go, especially its handling of multiple workouts in a single download. But it's definitely worth checking out if you're weighted down with PowerAgent, especially if you're on a non-Windoze system.

Tuesday, March 10, 2009

Crit Inferiority Complex

Last year's Menlo Park Grand Prix, 35+ 3-4: my goal was to finish in the pack. 3 laps to go, I flatted, and with the pit closed that put an end to achieving that goal. Close? Horseshoes and hand grenades. I then entered the 35+ 1-2-3 race as a second chance and got dropped when I made a bonehead error: thinking the "pssst-pssst-pssst" I heard was my tire losing air, when it was the guy next to me... I stopped, then when I realized my goof I couldn't regain the pack.

But even though I knew I'd come close to my goal in that first race, I also knew I'd ridden badly. Too many times I'd found myself tailgunning, riding at the back of the pack. Call it CIC: "Crit Inferiority Complex". I don't belong there. Pardon me for getting in your way. I'll just hide back here, out of trouble, where it's "safe".

Warming up
Warming up at Menlo Park (Jeff Remer)

This year, a new team, a new goal. Finishing in the pack wasn't enough. I had to be a factor. My criteria for factorhood were generous: get to the front at some point during the race, and never ride at the tail.

The importance of this was highlighted from my girlfriend Cara's experience in the W4/35+ W4 race: going into the S-turn, novice riders slowed down way too much, filed down into a single file. Gaps opened. Those caught behind these gaps were gone: nothing they can do. Those at the front weren't waiting for anyone. I've been there too many times.

No, the race happens at the front of the pack. Those at the back are spectators at best, there at the mercy of the riders ahead. I'd been practicing this philosophy on the Noon Ride.

But crits are a long way from the Noon Ride. When the race began, I'd move up a few spots, then settle in. "I can't move up further": no room, too fast, yaddayadda. It's easy to convince yourself of stuff you want to believe. CIC. So I hung out, never at the back, but too close to it. As long as I was close to my teammate, I was content. And I wasn't always close.

In the end, I finished in the pack. Tim was in there somewhere, as well. If this had been a stage race: s.t., mission accomplished. But it wasn't a stage race. I was never really in this race.

Racing isn't just about the result: it's about the path to the result. Sort of Zen, that way. I need to step up next time. Maybe not escalate my goals, but certainly I need to believe in them.

In road races, the hills have a lot to say. That's the essence of racing to me, even if I don't always like what they decide. But in crits, it's left up to us. Ride smart, near the front, or stay home. These skills, these confidences, are important, even for road racing.

At least I lived to fight another day, rode better than I had last year, and met last year's goal. Any crit where you walk away at the end isn't so bad. But it still left an empty feeling. I hadn't done what I'd planned to do.

P.S. Special thanks to Lorri and especially the whole VeloGirls team for putting on a really special and well-run event. Despite pressure from those wanting more of the usual, Lori really focuses the race on beginners and especially women. Yet she still managed to have such world-class riders as Brooke Miller, Kathryn Mattis, Daniel Holloway, and Andy Jacques-Maynes.

Saturday, March 7, 2009

Dog Day

I did the Passion Trail Bikes demo today.... a 2008 Santa Cruz Blur XC (not the 2009 carbon fiber version), with the lower-end component spec (for example, the Fox rear shock had no lockout).

Santa Cruz Blur XC
2008 Blur XC

The bike was fine, although I preferred the ride of the Specialized Epic Carbon I rode there last fall. The Specialized had more of a hardtail feel on the climbs, while the Blur XC had a softer feel. But the problem today had nothing to do with subtle differences between VPP and Horst-link rear suspensions. Nor was the problem really the engine. The problem was with the control systems.

When riding up a steep mountain trail in Belmont's Waterdog Park, on numerous occasions I'd face a sharp, rutted switchback turn.... and bail. Or a steep drop-off on a descent.... and bail. Or a tight downhill switchback with a steep slope immediately beyond the edge of the trail.... and bail. Then face clipping back into my Speedplay Frog pedals, which aren't the easiest pedal to securely get into.

Even when I was riding, I'd be spinning slowly up steep muddy singletrack, watching guys significantly heavier float up the trail ahead. And on the subsequent descent I was on the brakes way more than these guys. I wish I could tell you about their descending form, but they were too quickly beyond sight for me to observe it.

Mountain bike tech is fun. It's cool when a company releases a frame which shaves this many grams off the previous year's model. But in the end, a few grams here or there, a few subtle differences in sag, rebound, or actuation, none of it matters if you look the trail in the eye, and blink. I might as well have been on my ten-year-old heavy Gary Fisher hardtail, which is to its credit a very fine bike. Far finer than its rider.

For now, I'll stick to the Marin Headlands, China Camp, Russian Ridge, and similarly less technical venues... Today the Dog took me down.

Low-Speed Rail

Caltrain just shifted their southbound weekend schedule by 15 minutes to allow extra time for nocturnal track work.

Caltrain weekend service is a long, sorry tale. For two years weekend trains were canceled altogether for track upgrades, replaced by a limited-stop "RRX" bus which kept up with the previous local-stop schedule to the stops it served. This included a brutal round-trip crawl between 101 and Palo Alto transit center.

With track work completed, Caltrain patted itself heartily upon its back by returning weekend rail, presenting three schedule options to public comment. There were plans A, B, and C (PDF schedules linked to page no longer available). Under A, the train would stop at approximately 2/3 of the stations, every hour. Under B, they would stop at every 2/3 of the stations, but some of these stations would be served every hour, others every hour. The final plan, C, and the one I vocally opposed, was to stop at every station, every hour.

Obviously with slow-accelerating diesel locomotives and time spent at a station for passenger exchange, the more stops the slower the service. It's easy to see from comparing schedules of different trains that each stop increases travel time by two minutes. It doesn't seem like a huge deal, perhaps, but it adds up quickly: there's 23 stops available each direction, not counting the event-only stop at Stanford Stadium (actually Town and Country Village, a short walk from the stadium), so going from even 15 to 23 stops adds 16 minutes to a full trip each direction, 32 minutes to a round trip. That's a nice chunk of time, even if you're doing only half of the full run. How many people would like being stopped for 10 minutes, each direction, on 101 or 280?

Except that for those lucky few who live immediately proximate to one of the less-well-served stations, the "as-the-Google-flies" distance to the nearest two or even three stations typically differs by less than a mile. For example, some of the most vocal opposition to the "limited" weekend schedules came from Atherton. But Atherton station is within walking distance of the Menlo Park station, and a short bike ride from Redwood City. Move a mile perpendicular from the tracks, with more route options, and the delta between the stations is even less. Basic trigonometry, albeit modified by the available street layout. No problem covering the extra distance in significantly less than the time which would be saved by a limited train in its run to San Francisco or San Jose.

But despite logic, YIMBYism prevailed, and the Schedule C was adopted, even including the Atherton and Broadway stops which had been abandoned altogether for weekday service. Caltrain resisted, even offering a Schedule D which alternated local and limited trains, but to no avail. The result? Very few people with a choice seemed to take the weekend train. Why would you? It's much slower than driving, even if there's a train available to take, an "social responsibility" to reduce fuel consumption doesn't go very far in car-obsessed California, despite the "green" talk. Even worse: if you show up at a station at a random time, say after a nice bike ride, it's an average 30 minute wait for the next one. This is a stunning contrast with BART, where with a 20 minute schedule, waits average a third of this. Not surprisingly, BART does excellent business on weekends.

But at least with an early start in the morning I could manage. The 7am train from San Francisco pulled into San Antonio at 8:10am, allowing me just enough time to ride up to Los Altos to catch the Alto Velo A-ride, which met at 8:30. I'd bring some reading material, find a magazine rack to stash it before the ride, pick it up at the end of the ride, and continue to read it on the long, slow trip back north.

But then the 7am train was canceled due to budget. Even RRX had provided a 7am southbound. But now the earliest southbound was an hour later. But I could still catch en route the Spectrum Ride on Saturdays, and later when it was rescheduled for 9am, a north-bound A ride (with a bit of luck), in each case at Foothill and Page Mill in Palo Alto. That extra stop in Atherton seemed especially painful, as those extra two minutes would have increased my chances significantly of catching the A riders, but there was still a good chance.

But now it's all over. 8:15 is just too late. They pushed it too far.

Blur XC Carbon
My ride this morning

This morning, I'm going with Cara to Passion Trail Bikes in Belmont for their test-ride day....Santa Cruz Blur XC carbon for me, Specialized Epic S-Works for her. Wow! But our ride is scheduled to start at 9am, and we need to get set up with our bikes before then. That 8:15am southbound pulls into the Belmont station, a short walk from the bike shop, at 8:58. Too late. 8:43 would have just been early enough. So instead we drive a 2465 lb car 43 miles to transport less than 11% of its mass in human. All due to that 15 minute shift.

As the train passes through South San Francisco and Brisbane, it's within sight of 101. Car after car is visible from the mostly empty train. I find myself thinking that if no more than 1% of those drivers took the train instead there'd be demand for full trains running up and down the Peninsula every 15 minutes. Then there'd be plenty of time for limiteds to still cover every local stops, with reasonable local transfer times for those rare few who must travel between minor stations not served on the same route. Yet the demand isn't there because gas is too cheap and because the train service is just too poor. An the service is poor because the demand just isn't there.

An although I obviously take the weekend train more often than most (even a few times a year would put me in that category), many months I don't ride it at all. It's just too slow. Too infrequent. Too late. If you don't build it, they won't come.

Caltrain doesn't run a profit on any of its trains. It needs local support, local funding. It's not just a commuter rail, despite what it would like to claim. With public funding, it's tasked with providing public transit. It's time more people supported more local funding for Caltrain, but with that funding hold it to its public transit responsibility to competitive transportation service on weekends, weeknights, and mid-day, not just the "rush hours" so well served by the Baby Bullets.

Friday, March 6, 2009

High Speed Rail

High Speed Rail
Rendering of high speed rail car passing electrified Caltrain along Brisbane Lagoon

I just fired off a letter to the editor of the Palo Alto Daily News.

A loud vocal objection has been voiced to the impact high speed rail will have on quality of life. However, the proposed alternative is always that the would-be passengers will stay peacefully at home. To the contrary, if they aren't on the train, they'll be flying or driving. Flying requires valuable land for airport expansion, and more planes roaring overhead. More cars on the road? 101 and 280 are incredibly disruptive on communities, on local access, on noise, on air quality, on public safety, and on aesthetics. I experience this every day, living close to 101 and 280. I far, far prefer investment in rail to dumping more precious resources into inefficient automobile or air infrastructure.

Transportation investment is always a matter of damage mitigation, whether it is in walking or cycling trails or in projects of this magnitude. There are always those who claim to bear an excessive burden. But as long as people feel the need to move around, transportation decisions will be a question of "either-or", not the "yes or no" rail opponents prefer to engage. Given this, I urge readers to continue to support high speed rail for California: it's the best approach.


Will anyone be swayed by my small voice? I suspect not. But it's important to engage in the process.

Sunday, March 1, 2009

trail run #2

Brightroom PhotographyYesterday was my second trail run: the EnviroSports Huddart Park "half-marathon". It was my first EnviroSports event. I'd thought Pacific Coast Trail Runs, the promoter of my first run (also at Huddart, in December) was low-key.... When Dave of EnviroSports announced that if we took time to admire the view, we should record how long we did so and it would be deducted from our official time, I was impressed! And earlier, when I'd asked how we check-in at the turn-around, I was told it was "sort of on the honor system". This trail running thing is cool.

The profile was simple: up for 6 miles, down for 6 miles. Not quite a "half-marathon", after all, according to the official course profile.

New Balance 790 trail shoes
My new shoes, post-race

I had new shoes for this one: New Balance 790 lightweight neutral strike trail shoes. I love these shoes, and was lucky to grab them, as they've been discontinued: apparently lightweight trail runners are not in demand. Chi Running advocates lightweight shoes, relatively loose to not constrict the foot: typically a half to full size larger than otherwise, laced such that the shoe can be removed without unlacing. I've followed these guidelines, and my feet have been happy throughout my return to running this winter season. These shoes are 136 grams heavier than my Nikes, but if the improved tread saves only 5 mm of slippage per meter of uphill stride, that's 30 seconds saved per 50 minutes of climbing. If uphill speed is inversely proportional to total mass, that's 22 seconds saved after accounting for the weight difference. Add in an intangible increase in confidence on the descents, especially in yesterday's muddy conditions, and I strongly suspect they're the better choice.

My other new toy: a Nathan water bottle belt. There's two issues here:
  1. Carry or not?
  2. Belt or hand?
First, carry or not carry. Hmmm... 91 grams for my bottle, 152 grams for my belt, add in up to 400 grams for water. That's 640 extra grams carried up the hill versus relying on the rest stop at halfway for hydration. The cost of carrying this mass for 50 minutes, again assuming speed inversely proportional to mass, is a whopping 34 seconds for 50 minutes of climbing. The advantage? One is I didn't need to stop and drink at the turn-around. That would have been around 10 seconds there. And that drink would have been water, as I don't like the orange powdered Gatorade solution they provided (at the last run it was de-fizzed cola, which worked quite nicely). But I could have offset the loss of carbohydrate by taking a much lighter gel flask with CarbBoom or Hammer Gel. My gel flask is 155 grams full. Even without a way to carry the flask, I'm down to 15 seconds saved by skipping the water bottle.

So it comes down to the question: did regular hydration improve my climbing speed by at least 15 seconds? A nontrivial question, as so much depends on conditions. I was under 55 minutes to the turnaround, so the question is how does one compare the benefits of a hastily gulped drink 55 minutes into an effort to sips every approximately 7 minutes?

I don't know. For longer events, or in hotter weather though, portable hydration gets the nod. And once I had the elastic straps tight enough, this belt worked great. I was able to retrieve and replace my bottle in a full run without issue.

Okay, so the other issue is hand versus belt. Trail runners often carry bottles in their hand. To me, this is nuts. First, it imbalances the arm swing, which imbalances the stride. Second, it fatigues the arm holding the bottle. Sure, you can swap it from arm to arm to distribute the load, but still... Perhaps this is some sort of "old school" thing. Another factor is weight. Belts are heavier. But hand carrying just seems the wrong way to go.

Oh, yeah.... the run. Well, a brief summary. Approaching the turn-around, I stepped in some mud and immediately heard a "suck!-pop!" only to find myself continuing on in one shoe. I swung around and somehow managed to snag both of my light-weight gloves on the barbed wire fence bordering the trail. The amazing thing is my hands never touched the stuff: not a scratch. I managed to reattach my shoe and get my gloves off the barbs without much of a problem: I lost one place, from 14th to 15th. This almost reoccurred during my return, but by curling my toes I was able to keep my shoes from detaching.

On the descent, a repeat of my frustration from December. I just can't descend as fast as my competitors: I lost 7 places from the turn-around (15th) to the finish (22nd), crossing the "line" (actually "reaching the finishing table") in 1:34:58. I'd thought then it was flexibility and stride length. But I've been working ever since on my hip flexibility. Now I think differently: simply muscular adaptation. The only time I ever run on trails is during these races. I simply think the muscles stressed during down-trail running aren't up to the task, limiting how fast I'm willing to go. So the simplest answer: practice.

What's next? I don't know at this point. I should refocus on riding. But come fall, I'm ready to start again with running. Maybe the Sacramento Marathon in December. It's a Boston qualifier, and I should have more than enough time to prepare for it after the end of the cycling road season in September.

Oh, yeah -- and more trail runs! They're a blast.

added: Here's someone's GPS data from the race. GPS-derived distance is unreliable due to signal-drops on the highly circuitous route: the GPS interpolates straight lines between locations.