Monday, March 31, 2014

Low-Key kits arrive!

Working from home today due to a follow-up appointment with the oral surgeon, I was pleasantly surprised by FedEx which delivered my Low-Key Hillclimbs kit. And I was very pleased with the result!

Here, for reference, was my post on the design.

Here's a brief photograph, taken by Bryn Dole, who also ordered one:

Compare and contrast with the preview:

It came out pretty much exactly as expected, with one small detail: I implemented a drop shadow as a translucent black layer, which forms grey on white but darkens other colors. Somehow that came out as grey on colors as well. But it still works. It's a very minor detail. Also, the green on the front extends a bit higher than I'd anticipated. But the rest of the design came out extremely well.

And the fit: perfect. It's a lightweight jersey which fits me snugly, perfectly suited for hillclimbs (okay, perhaps a skin suit would be better for hillclimbs, but this is still excellent). And the pockets are a nice size, with an added pocket for credit cards or cash so they won't get accidentally lost when pulling a vest or bar out of one of the main pockets. And the shorts fit well, with the straps the correct length for me, the text and climbing guy on the back right where they're supposed to be.

I look forward to wearing it. For the next few climbs I do I'll probably be in the Team Roaring Mouse colors. But come Low-Keys in fall, this is what I'll have on.

Sunday, March 30, 2014

Crissy Ave climb: power data

Training for my upcoming 50 km race was going so very well. And then my tooth broke.

The result: I had to get a tooth extracted by an oral surgeon. No big deal, right? An hour of unpleasantness, and it's done. Business as usual.

But no such luck. I ended up exhausted. I did a few runs on the days following the procedure, very short runs of a few km each, but then something just dragged me down. Maybe it was allergies. But I think it was a cumulative thing, and the oral surgery was the straw that broke the adrenal's back.

So for the first four days last week I was work from home, with essentially no exercise. Then I commuted, by foot and train, to work on Friday. Saturday I rested again, then today (Sunday) I got in a nice little ride with Cara. Another chance to get on the Winter Allaban. It was fun, although I couldn't resist a little test riding over some scarified pavement on the shoulder of Great Highway. Obviously, the bike was a lot more confortable there with 32 mm tires than it would have been with 26 mm tires. But my rear fender came loose. Examining the bolts, I realized they didn't have washers, so after the ride I put in a washer on one (couldn't get the bolt to grab on the other one) and Loc-Tite on both. Hopefully that holds better.

On the ride I made just two little efforts. One was on a little climb of Crissy Field Ave, from Crissy Field up to Lincoln Ave. It's a nice short little power climb. It's funny how certain climbs just call out to be ridden hard.

Recently I looked at data from climbing the Cortland Hurl. Here's a similar plot for the Crissy Field climb:


Note the dependence is a bit steeper than speed proportional to power. That's curious from the standpoint of steady-state power equations, but if you throw in acceleration power (proportional to speed-cubed if I'd done a standing start) maybe it makes sense.

Despite my new torque wrench, the Vector still appears to be reading below the trend line set by Powertap, although some of the Powertap data also fall similarly below that trend line. I know -- it's terrible I've not done a direct comparison yet, and I'm long overdue for that.

Saturday, March 29, 2014

UCI rule 1.2.064 & E3 Harelbeke

I was going to do a post on UCI rule 1.2.064 bls, but Inrng already covered most of what I was going to write following Kuurne-Brussels-Kuurne.

Here's the rule in case you don't want to click on that link:

1.2.064 bis It is strictly prohibited to use sidewalks/pavements, paths or cycle paths alongside the roadway that do not form part of the course. Non-respect of this requirement is sanctioned in accordance with Article bis, without prejudice to any other sanctions that may apply.

The winning break at E3 Harelbeke yesterday. Photo by CyclingNews:

Cycling's a weird sport.

Actually, when you think about it a bit, this rule is just weird on its own. You can't ride on sidewalks, pavements, paths, or cycle paths along side the roadway that do not form part of the course.

This, apparently, is in contrast with roadways which are not part of the course, which are fine.

I thought it was already well established that in a race, any race, you must complete the entire course, which generally implies remaining on the course unless you plan on returning to where you left it. So it's not clear this rule changes anything, in principle. In practice, though, there's the written rules and there's the actual rules. That was, after all, a major theme of The Secret Race.

Friday, March 28, 2014

UCI revised calendar: who makes the 1A-1B race cut?

UCIThe UCI reforms appear to be moving forward, with a goal of professionalizing professional cycling. The reforms have been summarized in various stories (like this one) but the primary source is the UCI's Info Reform newsletter.

It's good stuff, and while San Francisco has gotten some recent insight into realpolitik, so far I've been really happy with Cookson's new term at the head of the UCI. Things honestly appear to be moving in the correct direction.

A big issue is the professionalism of professional cycling. Professional bike racing has tended to be ruled by chaos as much as any formal organization. Historically the race promoters have called the principal shots. They want to invite who they want to invite to their races. That typically meant balancing considerations for attracting the best teams, local teams, teams with charismatic riders, and notably in the case of the Tour of California, teams who bring big money race sponsors. It's all been a mess. If I go to a perspective sponsor and say "here's my business proposition: you pay X and your return is Y, where Y >> X", the sponsor will view it as a favorable deal. But a big part of Y is getting in big races, and if I don't know what races I'll be in, then the X the sponsor is willing to pay is much diminished. That means less money for teams, less money for cycling, and almost by definition bad for the sport. Virtually no other team sport is ruled by such caprice.

There's been various attempts to remedy the situation, including points systems where teams would add up points scored by riders. But this has had disastrous effects on the team aspect of racing. If there's a 240 km race, with 9 guys on a team, and 8 of the guys bury themselves so 1 can win, the one gets the points, while the 8 get Strava kudos but at the end of the season good luck with that contract thing: they'd be a liability to the teams which hire them because to win the Tour de France you need to be in the Tour de France and to be in the race you need individual rider points. It was corrosive.

To help address some of these major issues Jonathan Vaughters proposed a 10-point plan for professional racing. It all makes sense. But Vaughters was not well received by the previous two UCI heads, Hein Verbruggen and Pat McQuaid. Indeed, while Vaughters called for transparency, few have used the work "transparent" in conjunction with either of these two men. But then, just when it looked as if McQuaid's attempt to twist the rules and gain a third term as UCI head might just succeed, Brian Cookson appeared and, undeterred by McQuaid's political games to win the favor of third-world delegates in the UCI vote, won the job.

Now the UCI and Vaughters and his supporters seem much more aligned. The UCI is moving much more in the directions proposed in that 10-point plan. And the that UCI's Info Reform newsletter is the latest proof.<

I won't try to summarize it: no need, as it's completely clear. But there's a few key aspects worth comment.

One is how the entry to the top races is gained. Teams can be level 1A, 1B, 2, or 3. There's 16 teams at the 1A level, the top level in cycling. Eventually, you get to the IA level by being the top 1B team or else being one of the top 15 of the 16 1A teams. Retaining all but one of the top teams provides stability while still putting a check on team quality. The top 1B team can be promoted as a replacement. This generates competition among the 1B teams for this privilege. Of course, teams come and go, so I'm not sure what happens if 1A teams dissolve. Hopefully they dissolve a lot less, and instead change ownership, more consistent with how other professional sports teams evolve.

The 1A teams are ranked based on how they do in 120 1A-level race days, into each of which they are guaranteed entry. Meanwhile, 1B teams are ranked based on how they do in 50 1B-level race days, into which they are guaranteed entry. That's 170 days of racing at the top level. These days include the days allocated to grand tours. To avoid conflict, at most one 1A or 1B level race can occur on a given day. This allows 1A and 1B level teams to be competitive with their top riders in every race, never (in principle) forced to split teams between two races while still completing the entire schedule. And this creates the curious question of which races make the top level.

Of course, the Grand Tours make the cut: the Giro d'Italia, Tour de France, and Vuelta d'Espana, each 23 days or 21 days depending on if you count rest days, will be 1A. Additionally, the "monuments" and other top-level classics will be 1A: Milan-San Remo, Paris-Roubaix, Flanders, Paris-Roubaix, Liege-Bastogne-Liege, and Giro di Lombardia are the monuments of cycling and will of course be 1A, as will I suspect Gent-Wevelgem, Amstel Gold, San Sebastian Classic, and Paris-Tours. I'd also expect Fleche-Wallone to be there. That's 11 more days. We're up to maybe 77 1A races: still 43 days left to allocate to 1A, 50 to 1B.

The challenge comes when we get to the one-week stage races. The issue here is overlap. Paris Nice? Of course it must be there. But what about Tirreno Adriatico? These races are considered parallel paths to Milan San Remo fitness. Annually the game is which of the races will produce the Milan San Remo's winner. Paris-Nice starts and finishes a few days earlier, but they substantially coincide. Under these rules, if they are going to retain their roles as Milan San Remo preparation, then they will need to stay overlapped and that means at most one can be 1A or 1B, at least one of the two relegated to level 2. This would be hard to comprehend.

Then there's the Volta Catalunya, which is finishing now. This is World-Tour level stage race, providing key preparation for the Giro d'Italia, which starts in May. But it overlaps with the early races of the Belgian classics season, including Dwars door Vlaanderen, E3 Harelbeke, and even Gent-Wevelgem. How do you resolve this conflict?

Then there's the Tour of California and the Giro d'Italia. It's clear who wins this one.

The UCI Worldtour calendar is here. While it generally avoids such conflicts, the ones which remain are difficult to resolve, especially since the choice isn't 1A or 1B, but 1A or 2, since 1A and 1B conflicts aren't allowed.

It's not tragic if a race is class 2. 1A and 1B teams can still participate. It's just that these races will not count toward 1A and 1B standings, and so the races can't expect to be a focus for these teams. This is an issue for the Tour of California, which wants to attract a top squad from teams with strong US riders.

Actually, one correction. I'm calling the race classes 1A, 1B (combined to level 1), 2, and 3. That's not how the UCI labels them. They use classes: 1A, 1B, HC (beyond category), 1, and 2. This is obviously marketing spew. Come on: the third-ranked tier is "beyond category"? I'll stick with my enumeration for purposes of this blog post.

One more aspect worth mentioning: to determine initial entry into the 1A tier, rider points will be used, but only from the top 5 point-earning riders on each team. This is a compromise, leaving plenty of room for riders on teams whose focus isn't on scoring their own points but rather on aiding other members of the team in getting results. Cycling is a team sport, after all, despite the fact medals and prized are typically given to the top individuals more than the top teams. It isn't clear, however, how this point scheme will be applied in future years, or how new teams will be assigned. I suspect the same points scheme will be used for new teams.

Tuesday, March 25, 2014

Milan-San Remo and "breaking tradition" with Pompeiana

When I rode the finale of Milan San Remo I took a wrong turn approaching the Poggio, finding myself on an amazing climb winding into the hills. I eventually realized I'd gone the wrong way, but my thought at the time was "wow -- there's other climbs the organizers could add to this route if they wished".

And for 2014, they decided to do so But when it was announced that the Pompeiana climb was being added to the course, it was decried by traditionalists as breaking up what was the traditional opportunity for sprinters among the monuments. The only other "sprinter's classic" is Paris-Tours, and that's more of an end-of-season consolation prize, hardly on the level of Milan San Remo, Flanders, Paris-Roubaix, Liege-Bastogne-Liege, or the Giro di Lombardia.

I had to laugh a bit at this, because when Eric Zabel won in 1997 it was considered an enormous accomplishment that a sprinter won. He went on to win 4 out of 5 years. Then the following year Mario Cippollini won Milan-San Remo, in 2002. Until this time, the finish was considered too difficult for a sprint, with solo attacks typically launched from the Poggio or even the Cipressa. The question wasn't if an attack would succeed, but rather from which specific spot on the Poggio the wining attack would be launched, or possibility if the attack would be launched on the Cipressa. Sprint? Incredible.

But from then on, it did indeed become a sprinter's race. Teams became coherent, tactics more focused, and the training more dedicated to getting the fastest finishers to San Remo in a group. It's become the rare exception rather than the rule that a small number of riders, let alone one, are able to break free.

It hasn't helped that the finish was moved further from the base of the Poggio. A few seconds advantage from the bottom of the descent, while it may have previously led to a narrow victory, is now more likely to get absorbed in the final kilometer.

This plot was posted to twitter by PlataformaRC showing the size of the lead group arriving to the finish line @ this, what is now annually the longest single day of racing on the professional calendar:

It's clear: "tradition" is for solo moves, with new climbs added each time the sprinters start to take over. First the Poggio was added in 1960, then the Cipressa before Poggio in 1982, then Le Manie still earlier in the race in 2008. Le Manie was removed from the course this year to make room under the 300 km cap for Pompeiana, but then that was deferred to 2015 due to recent weather damage. In each case except for Le Mani which wasn't on the route long enough, when a climb was added the finishing group size generally diminished, but then tactics adjust, and the large-group finishes return.

So Pompeiana is not an end to tradition, it's a continuation of it, an attempt to restore Milan San Remo from a race of sprinter's teams to a race of solo conquests. Tradition suggests it will help, for awhile, until the sprinters take control again.

Monday, March 24, 2014

Milan San Remo 2014: active rider podiums on cycling monuments

Still suffering the after-effects of oral surgery, and mourning the loss of my missing tooth, I took solace in watching a pirate feed of the Eurosport coverage of 99.7 of the final 100 km of Milan-San Remo. Unfortunately I lost the feed during the final 300 meters...

So why don't I pay for Because I keep telling myself I'm not going to watch these races any more. Except for this one. Special circumstances. That's it...

The result: my man Cavendish didn't win Milan San Remo after all, his sprint tempered by the freezing rain, and once again I sold the remarkable Fabian Cancellara short as he finished an amazing second in a medium-group sprint. Katuscha's Alexander Kristoff took advantage of the work of his teammate, Luca Paolini, to win the sprint. Both of them pack some beef on their frames, something which likely worked to their benefits in staving off hypothermia.

Cancellara's 2nd place has continued an amazing streak of podium finishes for him. Here's his results in the last ten monuments he's finished, as he retweeted: 1, 1, 2, 3, 2, 2, 3, 1, 1, 2.

CyclingTips posted this excellent montage of Cancellara podiums to Facebook:
Cancellara podiums

Back to Milan-San Remo:
sprint photo

US riders had a relatively quiet race this year, with candidate-for-the-win Taylor Phinney missing out due to illness. United Health Care had two US riders make it to San Remo: Kiel Reijnen in 77th, and Christopher Jones in 109th. I simply cannot imagine riding 294 km in under 7:15, let alone in those conditions, as these riders did. It's truly mind-boggling. The lead group finished in 6:55, an average speed of 42.4 kph, which is a typical fast criterium by NCNCA cat 3 standards.

The following data were impressively compiled by jaylew on the cyclingnews forums, showing the record in cycling "monuments" by current riders as of Milan San Remo 2014. Cancellara's record is absolutely amazing. Even one monument podium is enough to make a career a success.

F Schleck4000031
S Sanchez4000004
A Schleck2100020

Here's a correlation matrix for the above podium statistics. A positive correlation implies a rider podiums in one race is likely to do podium in the other. A negative implies a podium in one race is less likely to podium in another. Paris-Roubaix and Tour of Flanders (RVV) have a strong correlation, while Giro di Lombardia, held at the end of the season and on a very different sort of course, is negatively correlated with these races, but positively correlated with Liege-Bastogne-Liege, which is a similar course.


Sunday, March 23, 2014

Tale of 3 videos: Hobbit, Bosch, and Sherlock

These past four months I've gotten to see three of my favorite book series I've every read get made into video. As a kid, two of my favorites were Tolkein and Conan Doyle, in particular the Hobbit and Lord of the Rings, and the Sherlock Holmes series. I don't need to describe why these both are absolutely brilliant.

As an adult, I got hooked my Michael Connelly's series, the most famous Harry Bosch detective novels. These books have a gritty darkness, along with a veracity which is attractive. There's always plot switches at the end, and it's fun trying to anticipate these. I wouldn't put the books on the level of the others, but they're definitely worth reading.

All three of these authors were made into video recently. The first I saw was The Hobbit, the Desolation of Smaug. I'd see the first Peter Jackson Hobbit film when that was released in 2011.

Peter Jackson apologists argue that this should not be judged based on fidelity with Tolkein, that this is an independent, fresh tale. Bullshit. It is obviously presented as the same name as the book, the same ovations as the book, and the same quest as the book. It's not too much to ask to show some respect for the book, one of the masterworks of the genre. Jackson basically guts Tolkein's work, replacing it with his far less refined set of shallow plots and even shallower characters.

These films have gotten criticized as dragging out a single, relatively short book into over seven hours of film. This isn't an issue, in my view: it took me longer than 7 hours to read the Hobbit, and that skims over a lot of action which would be longer in video. A main theme of the book was the long, long journey to Lonely Mountain, including long, difficult treks through the Misty Mountains and the Mirkwood Forest. Then there was an extended series of events when they finally reached Lakewood, and on to Lonely Mountain. There's plenty there for three films.

But Jackson chose instead to invest hundreds of millions in frenetic hyperreal action sequences and tie-ins with characters and even entire species from Lord of the Rings which didn't exist in print. There seems the implicit assumption the target audiences of today enjoy these super-charged over-the-top too-fast-to-follow chase-and-battle scenes. I'm not sure. To me, it causes detachment, any hope of a suspension of disbelief shattered, and I'm simply left to passively stare at flickering graphics without any perception of risk or uncertainty. I'd much rather see the company slowly hiking through the alien environment of the Misty Mountains, with a chance to enjoy the stunning computer graphics. Then the actual danger scenes, as they are in the book, would no longer be diluted into meaninglessness. As it is, it's an insult to viewer intelligence, no matter what the age.

So the Hobbit was a disaster. What I did like about it were the incredible computer graphics, Bilbo, and Smaug. Two of these three turned out to be somewhat ironic.

Next was Bosch, produced by Amazon Video. This pilot was a translation to near-present of some of Connelly's works from decades past, in particular Concrete Blond and City of Bones. The characters were relatively loyal to the book characters, at least on the surface. But the surface is about as deep as they went, and they came across relatively shallow, without any of the depth which makes the books so attractive. I was glad to have watched it, but it felt a bit like eating airline food. And I thought having the main characters in the story smoke, something characteristic of Connelly's books in the 1990's but phased out since, while loyal to the original stories was out of place now that the story was translated 20 years forward.

Then finally, at a coworkers encouragement, I started watching the BBC series Sherlock. I'm up to Season 3 episode 1. The pilot, A Study in Pink, was absolutely brilliant. It's perhaps gotten silly at times since, but for television, my standards are very low, and compared to these the work here is just amazing. I love the way references to the original Conan Doyle stories are subtly inserted, so if you're familiar with the story you catch them, but only if. For example there was a reference to The Valley of Fear, from Series 1 episode 2, where Sherlock is looking for a book which everyone would own to crack a cipher and he grabs but immediately discards as inappropriate a copy of the Bible. In the book the Bible was the book. There's many other such references.

Indeed in many ways the program is more in the spirit of the original stories than versions which place Sherlock in the 1890's, because when the early stories were first published, Holmes was a very modern person. Restoring him to modernity provides a context for viewers much closer to the context experienced by readers of Strand Magazine at the time.

Back to my smoking criticism of Bosch: in Sherlock, the chain-smoking Holmes has become the patch-wearing Sherlock. "A three-pipe problem" becomes a "three-patch problem". Well done, BBC. To lean on tobacco as a character prop is a crutch.

The irony is that the two main characters in this series, Holmes and Watson, are played by the actors who did Smaug and Bilbo in the Jackson Hobbit series. Elijah Wood for Series 4? Who knows...

Saturday, March 22, 2014

Milan San Remo prediction

Ah, well. My running streak ended at 7, as today I decided the best approach was simply to rest. I watched 3 episodes of Sherlock, season 2 episodes 2-3, and season 3 episodes 0 and 1 (0 doesn't count). Oral surgery just left me totally drained.

Tomorrow: Milan San Remo. My prediction: Mark Cavendish. Headwind finish makes it very hard for a break to stay away. Cavendish has a strong team, he has the experience, and he has the acceleration and timing needed to take it. He just needs to get over the Cipressa.

running every day

Each day this week I felt like not running. My right leg was bothering me a bit and I was even hobbling a bit at work from a sore something-or-other (foam rollering was "productive"). Lunchtime would come, my usual running time, and I couldn't go out: too tired, too much work to do, yadda yadda.

But then I'd take a not-too-late evening train, despite succumbing to temptation and eating first thing after riding my steep commute route from the train to home, put on my running clothes and trot down the hill.

And after a solid weekend running in the Marin Headlands, every day I did this: Mon, Tue, Wed, and Thu, my legs magically felt better and I got a wonderful run in, taking advantage of daylight time's later sunsets, and running into the night if necessary.

Today was going to be the day I for the first time ever made it to 7 consecutive days of running.

On Tuesday a 9.5-year-old crown came loose, taking a piece of 18 with it. There wasn't much left. The verdict: it had to go.

Tooth extraction is an ugly business:


So all through Thursday and into Friday morning it was with a sense of growing dread that I anticipated my appointment with doom,

I entertained thoughts of running to the dentist/butcher to get my Friday's ride in, but was still tired from my Thursday night 15 km, so I bicycled there instead. This was perhaps my mistake (although running in the morning after an evening run is much more difficult: every hour of recovery is precious). It was with the grimness of the condemned that I pedaled to the office, each pedal stroke one less remaining in the life of the tooth I'd grown to know and like all of these years. It will be replaced by something shiny and new, but not me.

The procedure was quicker and easier than I expected. Local anesthetic, some rather unpleasant manipulations, and then it was out. A few stitches and then I was ready to leave. Don't exert yourself, I was told, as elevated blood pressure is bad for healing. Even though I'd skipped the general sedation chosen by apparently 70% of the patients at the clinic, mostly to avoid the after-effects, anesthetic alone wears me out, and the trauma of the whole thing didn't help. I was exhausted.

I walked with my bike to a local Whole Foods market where I worked remotely for a few hours. Then I went home, partly walking, partly riding, and took a nap.

Post-nap, I took an ubuprofin to temper my throbbing head, ate some yogurt, and then did some more work. Eventually I started to feel better, and ran to Trader Joes for some more carrot juice: 3.2 km total. So the streak goes on.

Friday, March 21, 2014

more running metrics: how often, how far using ATS framework

The generalized form of the ATS and CTS metrics developed by Coggan & Allen for the analysis of power meter data is the following:

fn = sum { i from 0 to n } exp[ (ti - tn) / τ] si

where ti is a time associated with point i, τ is a time constant associated with the stress score f, and si is the stress associated with the activity at point i.

For my tracking of running training, for simplicity I've used distance run as the training stress (which I measure in kilometers). Then I've used the conventional values τ for two stress scores: 1 week for an "acute" stress ATS, and 6 weeks for a "chronic" stress CTS.

The advantage of this sort of approach versus a more conventional "miles per week" is "miles per week" is arbitrary. You could run 1 km today and if you rested 7 days prior, ATS goes up. Then you could do 30 km tomorrow and if you ran a marathon the week before, ATS drops. Obviously that's not representative of how your training stress changes. Recovery isn't nothing for 6 days then 100% for one day. It's a gradual process, and the exponential functions capture that. Additionally, fitness from an activity doesn't stick around for 6 weeks then disappear in a single day. It may well be more complex than a simple exponential but the exponential is a lot better than simply tracking distance run over some arbitrary time period.

These scores as I described them measure how much I've been running. But two other metrics of interest are (1) how often I've been running and (2) how long my runs have been. Recently here I plotted the number of runs in the preceding 7-days for each day. This provides insight but suffers from the same problem as plotting weekly distance. Another metric of interest is the average length of my runs. I can use the same 7-day window to do this, but again that would be arbitrary.

Fortunately, I can easily use the generalized formula above to get better metrics for "how often" and "how far".

For "how often", I simply use an si in the above formula which is 1 if I ran, 0 if I didn't. This is described by a step function u, where u is 1 if the distance is greater than 0 or some other threshold distance like 1 km, or 0 if the distance is zero or less than that threshold distance. I represent that for each point i as ui:

fn = sum { i from 0 to n } exp[ (ti - tn) / τ] ui

Then for "how far per run", I divide ATS by how often I ran.

So here's an update on my ATS, CTS plots (using km per day as my "daily stress"):


I've been moving the CTS curve up to the trend line target.

I resolved to run more often. Here's how I'm doing, looking at % of days run:


But in doing this, my runs are getting a bit shorter (shorter after-work runs instead of longer lunch runs, and long weekend runs are a smaller fraction of total):


The goal is to keep running every day. If I do that the CTS will surely stay on track.

An interesting thing about these two plots is it doesn't seem from a quick glance I've been ramping anything up for the past few months. It's only in the ATS/CTS plot that it's clear my training has been moving forward. So the primary value is still in ATS/CTS. These auxiliary metrics just provide supporting data.

Of course, these metrics could be applied to cycling as well, although a power-based metric like TSS would probably work better than distance for measuring each activity, or an estimate (such as GoldenCheetah provides) if power data is missing from a given ride.

Wednesday, March 19, 2014

INRNG Contador photo

It's really worthwhile following the INRNG Tumbler page. I'm glad I do. Because of it, I found this photo of Alberto Contador approaching the brutal finish of Stage 5 of this year's Tirreno Adriatico stage race. Contador won the stage and thus took the lead, for good, in the general classification.

Oh, the pain, the pain...

But I can't stop looking at it.

It helps I've always admired Contador. He rides with style, he pays meticulous attention to equipment details (experimenting with things such as large-diameter derailleur pulleys, lightweight wheels and saddles, and low-friction bearings), and he's always modest and polite in his interviews.

Of course, there's the inevitable doping questions: convicted doper Contador who dominated during the peak of the microdose-plus-transfusion era is back due to "a good winter". But training at altitude is no gimmick, and if he indeed went to a structured, focused off-season it shouldn't be too shocking he begins the season in shape. In previous years he's documented to have gone into the early season over-weight. This year, he's supposedly right where he wants to be.

For what it's worth, on the climb of Passo Lanciano, the final 9.10 @amattipyoraily (veetooo) extimated Contador's power at 5.75 W/kg assuming a CdA of 0.35 m2, which he averaged for 28 min 30 sec. That was the final long climb of the stage, but the above photo was taken on Muro di Guardiagrele, the far shorter but much steeper finishing climb.

Tuesday, March 18, 2014

progression of peak running distance approaching 50 km

Another day past, and another day less until the Woodside Ramble 50 km race.

I've been focusing on plots showing weighted average distances (which I refer to as ATS, CTS, after the cycling power metrics of the same names), and to number of runs per week. These plots have been guiding my training. An example is the following:

But long run distance is important as well, and that's been increasing. This hasn't been as much planned as it's been based on what I can do: I like to run long, and if I feel good, I try to extend my previous distance a bit. Here I plot the distance of my runs versus the date relative to 18 Mar, where I label training runs as crosses, and races as stars:

The long run distance has been steadily tending up, but not fast enough to approach 50 km by 13 April.

I've run 3 races so far this year. The first was just 4 km (listed as 4.5 km here because of warm-up). For that race, distance wasn't a factor; I'd already done runs twice the distance.

The next race was a half-marathon. I'd had 3 days approaching that distance already, so again the distance wasn't a concern. Indeed I had plenty of confidence I could complete the course if I didn't hurt myself.

Then came a 30 km race (actually 29 km). This was to be my longest run of the year, but I'd already had two days of 25 km, so making the 20% or so leap from 25 to 30 didn't worry me too much. The 29 km is a natural part of the distance progression in the plot.

My long run so far is a bit less than 33 km. That's only 2/3 the distance of the race. I've still got time to get one more long one in before I want to taper off. Maybe Saturday if I feel good. But it won't be over 35 km: no way. That'll still be no more than 70% of race distance.

I simply need to treat this race as what it is: a stretch. Goal: hit the half-way chomping at the bit, ready for the fun to begin. Then that leaves me 25 km to suffer. But I need to hit the half-way fueled, hydrated, and warmed up. Racing has advantages versus training: it has a taper, it has rest stops, and it has a willingness to push yourself harder.

Monday, March 17, 2014

Epic Bike Tour: the Marin Headlands

Cara wanted to do a bike tour, so when we were visiting Ely Rodriguez to talk about a handlebar bag, he suggested the Marin Headlands as an excellent destination. The Marin Headlands are just north of San Francisco, across the Golden Gate Bridge, but I immediately embraced the idea. It was a good chance to try out the whole bike touring process: pack stuff on bike, ride to destination, camp out, pack up and leave. The rides themselves would be short, less than 20 miles following the most direct route, but it would provide a good chance to learn what works and what doesn't. On a bike tour, the actual riding is just a part of it, maybe even a minority. This would test the rest.

First, I had to reserve the campground. This turned out to be surprisingly easy. Many campgrounds in the SF Bay area need to be booked months in advance, but in March, Haypress campground in the Headlands obviously isn't one of them. I initially tried to get Hawk, which is the most remote campground there, but it was full so Haypress was recommended instead. Haypress is down a 1 km fire road off Tennessee Valley trail, really Tennessee Valley Road at that point, just west of the stables. It's super-easy to reach on bike, with the only unpaved portion the 1 km fire road, and very little climbing. Hawk, on the other hand, would have been up an extended dirt climb, plus a 1 km access trail which was a total unknown to me. So we took Haypress.

The next step was the bike. Part of the reason I got the Winter Allaban was for bike touring. It has a randonneuring rack in front, an "optional" rear rack for panniers. I had bags for neither. Ely is making me a rando bag, but that's not ready yet due to a delay getting the fabric. And I emailed Chris @ Roaring Mouse Cycles to ask him to order some Arkel Dry-Lite panniers, but those hadn't arrived. So I sent out a note asking for a loaner, and a Roaring Mouse teammate offered to let me use his Bontragers, idle after an extended tour in France. Those worked out perfectly for the rear rack.

For pedals, I still had the Garmin Vector Look-compatible Exustars. These go much better than Speedplays when walking in dirt is involved. Unfortunately I'd removed and reinserted the batteries, since I'd re-torqued the pedals, and the left battery cover had gone in crooked, not closing all the way. As a result I didn't have power data until I realized the cause of the problem at the end of the trip. Ideally I'd have used my BeBop pedals, but I still need to mount the cleats, a non-trivial job of shoe surgery.

I was faced with the situation that my low-trail (3.9 cm) rando bike was going to be essentially 100% rear loaded. Jan Heine, who just published a report of a bike-camping trip he did with similar geometry where he was 100% front-loaded with low-hanging front rack but no rear rack (photo), would not approve, not at all.

My bike on right, Cara's on left. My front rack is supporting only my tool bag, which didn't fit on the rear. I packed it smarter for the return.

The packing list: my REI 10F down sleeping bag, a light weight pad, Cara's MSR Hubba Hubba tent, most of our food supply, a T-shirt, a sweatshirt, running shorts, running shoes, my Lake Chabot running shirt, a lightweight cycling windbreaker, a and a pair of underwear. I had my phone in my jersey pocket, and I had a large belt-pouch for storing food we bought along the way. This obviously isn't excessive, but neither is it minimalist. The tent had spikes and a ground pad we could have left home, or potentially even the tent could have been omitted completely. The sleeping bags could be replaced with quilts. But I don't have a quilt, and I like having a tent, so it's what we had. Actually, I regretted not bringing any reading material along, although Cara came through with an Outside magazine which I could read.

I was slightly nervous about starting, as from the first clip-in we were to descend two blocks peaking out at close to 20% grade. But the bike handled it well, and the brakes had no issues even with the heavy load.

From the bottom of the hill we road out to the Embarcadero, along the waterfront, past Fort Mason, then along Marina Green to the Golden Gate Bridge, At the first traffic light I learned the bike didn't handle quite like it does unladen. When I went to clip in I slightly lost balance. But once I came to terms with the substantially increased inertia, I was fine. It was a bit like starting a tandem.

But the handling was always slightly squirrely. There was a bit of a feeling of pulling a trailer. I'd steer, and the bike would delay in turning the corner. This is why Jan Heine advocates carrying weight in the front of the bike. I'll need to see what I can do next time. Maybe just the addition of a handlebar bag will help.

I was fine on the cobblestones near Pier 39, and had no issues on the climb of Fort Mason or on the following descent. Then came the bridge, with its cross-winds. Going around the supports, with heavy tourist-bike traffic, I unclipped going in, but then pedaled the rest of the way. I didn't trust the handling enough to go into the blind corners ready to dodge a misplaced tourist.

When we reached Tennessee Valley road, we stopped for fruit at the outstanding outdoor market at the intersection, open only on weekends. I love the dried parsimmons there, but the fresh fruit, preserves, and fig bars are all worth stopping for.

At the end of the Tennessee Valley Road rollers, we passed the jam-packed parking area (pun not intended) at the trailhead, then through the barrier. It's only a few hundred meters from there to the entrance to the camp access road. But Cara wanted to ride to the beach, so we kept going.

On a short roller, I went to downshift in the rear, but was in my big ring (48) front and that led to a 48-32 cross-chain. I threw my chain onto the small ring, which caused the bike to stall, and I hopped off. I walked the rest of the way up the short hill, went to get back on, but the loaded bike didn't react as I expected, and my foot slammed into the top tube. Oh, no... a scratch. The first. I felt terrible. I resolved to adjust the front cable tension as soon as we got into camp. It needs to be fine tuned with the wide-range gearing on this bike: 32-48 front, 12-32 rear.

The ride to the beach was fun, with a steep dirt descent which I knew would be even funner coming the opposite way. We at lunch at the beach, then turned for the return.

At that descent-now-climb, I made use of the low gearing on the bike, and headed up. Cara initially followed but decided, wisely, it was smarter to walk. There were some ruts which caused me to dodge left, then I dodged left again, then there was no more left. But losing momentum, I cut back to the right across the ruts, and made it up the hill, despite the load. I wondered how that would have gone with a front-loaded bike.

Back to the camp road, we turned left to go to camp. The road had recently been muddy but the lack of rain for the past few days had dried it out. It was rough in spots, with horseshoe hoof prints, and plenty of horse droppings. The trail led only to the camp-site, so I found the heavy horse traffic surprising.

At the camp site we took the furthest of five spots, thinking this would provide the least noise, but a large group moved into the adjacent spot. They weren't too loud, though.

setting up camp
Setting up camp

With camp set up, I was able to adjust the derailleur. One issue was the ramp adjust was too slack. It's best to have the pulley track relatively closely to the cogs, and the gap was too large. That generally makes everything work smoother I'm not sure if this contributed to the front derailleur issue. I also adjusted the front cable tension slightly. It was only problematic in the big-big cross-chain combo. Hopefully the tweaks solve the problem.

With camp set up, I got ready for my trail run. I was unsure how my legs would respond as I set off back down the camp access trail, but they were fine. I did a bit over 10 km up Fox Trail, down Coastal, out to the beach and up the stairs to the overlook, then back on the Tennessee Valley trail.

Potable water is hard to find in the Headlands, and neither of the two sources I know, the visitor center and Rodeo Beach, were anywhere close. So with my run finished, our bikes unloaded, I remounted a pannier and we rode back up Tennessee Valley Road into Sausalito with Cara's expandable water container in search of a supply, and also coffee for her Thermos for the morning. It was really nice being back on an unencumbered bike. We found water at the San Francisco Running Company, a fantastic trail-running-focused store. I resisted the temptation to immediately buy not just one but multiple pairs of trail shoes, for different conditions and run distances. My New Balance 970's seem to work fine, although on gravelly trails the traction could be better. I chatted with the co-owner awhile: he finished 17th in the Western States 100 in 2011, and is training for the 2014 race. While I was there, Cara got coffee at a nearby Starbucks. It was actually still warm the next morning.

Surly Pugsley in the nextdoor tent site

On the way back to camp, we stopped at the Tennessee Valley parking area, near the stables, where Inside Trail Running had the final rest stop for their Marin Ultra. A 50-miler runner (the long distance for the day) was recovering for the final 9 kms to the finish, which was near Fort Baker. After he was gone I chatted a bit with the volunteer, expressing my trepidation over the upcoming Woodside Ramble, on April 13, which would be my first ultra at 50 km. He said the trails in Huddart and Wunderlich, due to relatively recent resurfacing, were in excellent shape for running a first ultra (the resurfacing was a bit controversial, motivated by the local equestrian community). His advice: start conservatively. Obvious advice, really, but it's important to hear it over and over, from experienced 50 km'ers like him. It needs to be a mantra I drill into my brain during the beginning of the race. It can be incredibly tempting to follow the 30 k'ers, to run at a pace I know I can run.... for 30 km. Save the speed for the rollers, he said, at the top of the first climb. I'm not in the race "just to finish", but finishing well requires finishing, and if I lose a place or two for too conservative a start, that's way, way better than losing 10 or more spots for starting too fast.

I'm not a campfire guy or an artificial light guy when camping. When it's dark, I go to sleep. When it gets light I get up. This seems the most natural approach, and the most restful. I was awakened by the howling of coyotes during the night. Not much land for coyotes, sadly, but fortunately what land was spared was only just: the Marincello housing development, for example, had been a done deal before protests stopped it.

The next morning, legs still tired from the previous day's run, I put on my running clothes and headed out of camp again. This time I went the opposite way from the end of the camp access road: out to the Tennessee Valley trailhead, past the stables, up Bobcat, out and back to the Hawk campground, Alta, SCA, down Coastal, up Miwok, down Old Springs, past the stables again, and back to camp. It was just over 20 km. That made a solid two days of running, although my legs were tired.

Despite the tired legs, it was time to return home. So we packed our stuff and back we went. The cats were hungry, Cara said.

Saturday, March 15, 2014

runs per week

In the March 2014 Ultrarunning, page 16, Gary Cantrell writes: "The biggest obstacle to accumulating the mileage necessary to achieve the results we desire is inconsistency". He advocates running every day, maybe for just a mile, maybe for just 15 minutes, but to get in the habit of running every day.

So how have I fared? Not well.


It's easy to see the biggest killer on my consistency has been races. The solution may simply be to get out and do more super-easy runs after races. But easier said then done: if I feel hobbled, that even running across the parking lot at work is infeasible, it seems better to just let my body recover and not go out and force myself to run out of "habit". I'm doing better in the plot once I get a few weeks out from racing.

Yesterday my non-running status wasn't directly from a race, but from my right leg bothering me a little, which it's done since that downhill run on Empire Grade. But could I have gotten a short, easy run in? Probably. I was "too busy", though. Or was I really?

Keeping this curve up in the 5-6-7 range is clearly something I need to work on. I've got precious little time before Woodside. But running every day before then seems like a worthy goal.

Friday, March 14, 2014

projecting training to Woodside Ramble 50 km

I initially posted this at 1:59:26 am PDT on π day, which made me feel slightly clever, but then I felt considerably less clever when I realized I'd mistakingly concluded I had 5 weeks until the Woodside Ramble 50 km, rather than the 4 weeks I actually have. Analysis revised...

With only 4 weeks to go until my 50 km race, I'm becoming more and more obsessed with it. After Lake Chabois, I took a week off from running, but helped fill the void with cycling, made even more irresistable by my new bike, the Winter Allaban.

Woodside Ramble 50 km, Michigan Bluff photography

But with "rest week" over, my body fatigued from the rapid ramp-up in riding volume, it was time to start running again. And that meant Caltrain to work instead of SF2G. For Caltrain, it would be folly to take the Allaban, so I've been riding my "rain bike", a 1989 Trek 1500. That bike needs a new drivetrain, the derailleur and 7-speed cassette both well past their replacement date, and the Speedplay Frog cleats need to be swapped, but despite these issues it's still fun to ride. The geometry actually works fairly well for me, the same 73-73 angles and 75 cm top tube on the Allaban, although it runs 12 cm handlebar drop instead of 8 cm. The extra drop is okay, though, because with the brake-only levers the typical riding position is on the ramps rather than the hoods, which reduces the effective drop to my hands. I've really enjoyed being on the bike and it would be very difficult for me to part with it.

So I didn't come out of "rest week" as fresh and ready to run as I'd have liked. But I've been getting the runs in, culminating with 32.8 km on Saturday, 15.6 km hilly km on Sunday, then after two days of forced recovery due to the downhill pounding from that run, 12.2 km on Tue and 21.7 km on Wednesday. That's a nice solid block, but I've got to keep at it.

The goal is to continue my trend line of +61 meters/day of CTS increase per day, or +430 meters/day per week. I want to keep pushing that curve up. Right now, I'm sitting at 7 km per day:


Click on the plot to see higher resolution.

I've indicated on the plot my race dates, tagged with what my CTS was on those days relative to the race distance. At the Inside Trail Runs half-marathon distance at Montara Mountain, a very hilly course, I was at 18.5% of race distance and did a very nice run by my standards, finishing 11th. Then at the Lake Chabois 30 km (actually 29 km), also Inside Trail Runs, I was higher: 21.0% of race distance, on a flatter but still hilly course. That race also went well, and I finished strongly, in 11th.

Projecting CTS to the 50 km race, with 2 substantial climbs and additional "rollers", if I can keep to the +430 m/day/week rate of CTS increase I will be at 17.6% of race distance, significantly below the level I had at the half marathon, and well below the 21% I had at Lake Chabois. So if I believe this metric, then I need to be extra-conservative at the Woodside Ramble: throttle back on the opening climb, save more for the finish. But then this is sort of no brainer when starting the longest race you've ever run.

My recovery from Montara was haphazard, coming back too quickly then taking two extra weeks before I could run on a near-daily basis. Recovery from Lake Chabois, despite my overindulgance in cycling, has gone much better. This was critical for giving me the time for a solid training block between the two events.

The last point on the chart is within 100 meters of the CTS trend. To keep my CTS moving along the trend line I need to keep daily distance around 2.58 km/day above CTS, or since CTS is at 7.0 km/day now, I need to run around 9.6 km tomorrow to keep it moving upward at this rate. ATS is smoother than daily km, but it's lagging, so to keep CTS on the trend, ATS needs to be 2.15 km/day higher than CTS (dashed red line on plot). At present that's 9.2 km/day and it's 10.3 km/day so it's where it needs to be.

However, I need to do a bit better than these minimal values since I'll need to taper distance the last week, at least. But I've done this in the past: post-race CTS progress halts, pre-race CTS progress slows or halts, but in between I'm able to make up for lost time. This becomes more difficult as CTS rises, however. Obviously there's a limit. Competitive runners don't train distance in direct proportion to race distance. 10 km runners may run close their race distance every day, but marathons certainly do not, and 100-milers don't run even close to four times the training distance of marathoners. Longer distances are about trading speed for distance. And even if I stick to trend, I'll be a lower % of race distance than either of my previous two races. Is it enough? I have to think so, and treat the course with the respect it deserves.

This weekend: a bike camping trip with Cara, the Allaban's first tour. It's a short one, a beta test, and will leave plenty of time to get one or two solid runs in on the trails. With only a month to go I seriously need it.

Thursday, March 13, 2014

joining the torque wrench age

I feel like such a wimp. Old-school is to go by feel. Torque wrench? I don't need no stinkin' torque wrench!


torque wrench

Wednesday, March 12, 2014

POC bait and switch: Octal comes in 55 grams heavy

Since it debuted last year on selected Garmin riders and on Lance Armstrong, among others, the POC Octave road helmet looked to be a game changer.

POC Octal

Combining features which appear to provide a competitive advantage in safety with weights as light as anything else in the market, the POC looked to be the best of all worlds, especially considering the relatively round shape looked to be a good match to my relatively round head. I'm a good match for the Specialized helmets, for example the Prevail which I now wear. But the Prevail is going to need replacing, and the POC was a leading candidate. Just wait to March....

"190 grams in a size medium... certified worldwide". Amazing. Consider me in. The video was posted September 2013, 6 months ago.

Then there were the reviews. From BikeRumor:.

Tested against the two top selling high performance helmets, the Octal has 20, and 27% better ventilation, though POC isn’t leaning on that for a selling point. The Octal is rather a complete package with better protection, fit, safety and performance. All of this, and the helmet is still one of the lightest on the market. At a claimed 193 grams, the Octal is well within the ultra light crowd with the lightest samples at 188g and the heaviest of 100+ samples so far at 195g. POC isn’t necessarily going after the lightweight crown, but if the helmet is already one of the safest around and it’s one of the lightest – why not?

From VeloNews:

The weight is impressive, just under 200 grams (194g to be precise, for our large).

From BikeRadar/CyclingNews:

If there's one thing we wouldn't have guessed from looking at the lid it would be the weight – our pre-production sample tipped the scales at just 193g for a size medium.


And, to seal the deal, on the POC Octal web page:

SIZES: Small (50-56 cm), Medium (54-60 cm), Large (56-62 cm)
WEIGHT: 190 g (S), 195 g (M), 205 g (L)
CERTIFICATIONS: EN 1078, CPSC 12.03, AS/NZS2063-2008

In store March 2014

In January I posted this to the blog. I was obviously very interested.

Well, it's March now and I popped into Palo Alto Bicycles and there they were: freshly in stock. I'd ridden to Palo Alto from work in Mountain View, taking advantage of the 15 km distance to spin my legs out, still slightly tweaked from running downhill on Empire Grade on Sunday. A Caltrain delay due to a stalled engine blocking the tracks gave me some time to try the helmet on.

And, as anticipated, it fit great. The retention system was nicely adjustable. And it didn't look as bulky as it does in photos. It seems like an excellent helmet. The price was marked $270. I get a Silicon Valley Bicycle Coalition member discount, 10%, which covers the regional sales tax of 8.25%. That's a lot. But a helmet's important. You need to carry it on your head for thousands of miles. And the weight's important: it's weight saved over all of your bikes, on all of your important rides. And weight savings, as everyone knows, comes at a premium price. In Palo Alto Bikes customers are regularly spending $5/gram or more to buy lighter components or frames versus cheaper, heavier ones which would otherwise serve them just as well.

So when I looked at the box and saw "small: 230 g; medium: 250 g; large: 285 g" I was naturally disappointed. What?

I interrupted the sales woman who was telling me how the helmet had all these nice features and was "incredibly light". I pointed to the numbers on the box. "That's not particularly light." She seamlessly switched to damage control. To her credit, she didn't argue with the facts, but instead ran into the back of the store and fetched the owner's pre-production sample, the same one seen in the reviews.

Here's a photo from BikeRumor:


See that huge vent running down the center? It's gone. Instead there's now two vents with a new cross-bar. She explained this to me. It was necessary to pass US standards, to stop the helmet from breaking after hit by the falling mass used in the test. Instead it looks more like this:


So much for "worldwide certification". That's a 55 gram mass increase, assuming the claimed mass is representative.

Honestly, I think the original POC is a fine, safe helmet, safer than most of the US-certified helmets on the market. This is because it was designed not to the distorted certification test, but to real-world considerations. The added temple coverage, for example, does nothing to improve a score in the CPSC certification test, nor does the advantage of a smoother surface versus one with potentially neck-twisting points. And lighter is safer, as every additional gram provides a more fatiguing load to the neck and shoulders, and adds to the load which must be absorbed during a violent crash. "50 grams isn't much", you might say, but it's a straw-that-broke-the-camel's back scenario. At some point 50 grams does matter for safety. And, as the market has spoken, it matters on value.

So I'm disappointed with the bait-and-switch.

Some photos I took the next day:




Tuesday, March 11, 2014

Cortland Hurl: power comparison Powertap, Vector, Strava

In Dec 2011 I compared Strava power estimates to Powertap measurements for the Cortland Hurl, a previously popular climb for SF2G. Unfortunately the Hurl has gone out of favor with the Google riders, but recently I did a ride with a leisurely spin up the climb, which has been an occasional fitness test for me.

New to the ride was my use of Garmin Vector, mounted on my heavy (10.5 kg) Winter Allaban. You'd expect a relatively high watts at a given VAM with the big bike with fat tires at relatively low pressure. So I was curious to see how the watts I reported on the climb compared to my previous trend.

Here's the result:

watts vs VAM

The data are sparse, and it would have been a lot better had my Vector ride been at a higher VAM. But it is strongly compelling that the Vector is reading low. The Strava estimates are certainly low in particular due to an underestimation of mass. The Strava method of asking for "bike weight" and "rider weight" omits the weight of equipment, which in my case typically includes a loaded backpack.

Here's the same data, but converted to an effective mass, which is calculated assuming all power goes into moving mass against gravity. Since I'm typically 57 to a historic maximum of 60 kg, obviously a substantial fraction of power is going into wind resistance and rolling resistance and, perhaps, some acceleration.

kg vs VAM

Fortunately, I finally got replacement batteries for my Powertap, so I'll need to put the Powertap wheel on the Allaban and see how the two compare, directly. The small complication is the need to use multiple Garmin head units, and pair each Powertap with a separate Garmin.

Monday, March 10, 2014

Spring, Classic!

I've done my friend Murphy's "San Mateo Wheelman club ride" each of the three times he's led it. A group of friends get together, ride at a liesurely pace to somewhere, then ride back to the city via public transportation. Or something like that.

Anyway, with my focus on running right now, and not very good at deviating from a focus, I volunteered this year, because his ride is such fantastic, relaxed fun that I felt important to support it even if I wasn't going to participate as a rider.

This year's route started in San Francisco and worked its way in a dirty, indirect manner to the Ibis Cycles headquarters in Santa Cruz. Consistent with the leasurely, informal nature of the ride, the route was chosen only the day before. It's always exciting to see where Murphy's going to send his friends.

After a flurry of reorganization related to who had a car and who was was able to drive a car, I was assigned to checkpoint 2, a new parking lot (empty in satellite view but not yet on StreetView) off Skyline Boulevard. I hate to see trees replaced with pavement, and this is no exception, but the growing demand for mountain biking, hiking, and running on the trails around Skeggs Point were putting a load on the road-side parking, apparently. Certainly the lot proved popular during my time there.

I went down with Peter, who'd normallly be riding the event on his single-speed, but that was in distress, and so he opted to help instead.

Our task was to count the riders passing through, give them water, milk, soda, coffee, or other beverages, cook them sausages, and help them with whatever misclaneous needs they had, including calling the SAG wagon for them if they chose to bail out at mile 45. For just two people this was going to be a fun challenge.

Hide that!

Only when we arrived did we assess the collection of stuff in the car. There was a grill, in two parts, with a propane tank. There were various beverages. There was 14 gallons of water. There was an ice chest for which Peter had purchased a small bag of ice. There were sausages, and one very small container of Osmo hydration powder.

I thought we were marginal on the water, as the day was warm, especially for March, but clearly on the Osmo we were low, and on bananas we were nonexistent.

We managed to figure out the grill and the water dispensers. The grill was hot, even on the lowest setting, and I was afraid the sausages would overcook (they were pre-cooked to start with) before riders arrived. We targeted an earliest possible arrival at 11:15 am, but given the technical challenge of the route, that was optimistic. After all, this was a leasurely ride.


A friendly range came by and told us in his friendly ranger way that the grill wasn't allowed. No problem, I responded, and we shut it off. So it was cold sausages for the riders today.

An end to the hot sausages.

Surprisingly or not, hungry riders will eat almost anything.


The bigger problem was the general shortage of food. Peter made several trips 3 miles to the 84-Skyline junction to refill water jugs. He also got more ice, more Coca-Cola, and all of their meager supply of bananas. Still, some riders arrived at a poor time and didn't get water. I directed these riders to the store, a 3 mile descent, where there was a spigot.

After, Peter and I and a sagging rider got into the Windstar van, along with the bike I'd ridden to the start, and drove to Santa Cruz, the finish. The sag van took more sagging riders.

No cash for you!

I'd brought my running shoes, expecting an opportunity to get a run in at some point, and I wasn't disappointed. In Santa Cruz I left Ibis, ran up Western Ave, then High Street connecting quickly with Empire Grade. The route was essentially 7.6 km up, 7.6 km down. It was a great run made somewhat less great by the combination of high speed traffic and disappearing shoulders on Empire. I'd run into blind corners, running on the left side of the road, holding my arm outstretched so oncoming cars would see me sooner. If there was overtaking traffic from behind, eliminating the possibility of oncoming cars going into the opposite lane, I'd just bail to the shoulder if I heard cars coming. But I love the opportunity to be able to run up an extended climb, and the descent was good work for what has been a weakness.

I felt okay on the downhill, arriving at the finish of the ride just in time to see Paul rail the final corner and finish smooth-as-silk. I then chatted a bit with the finish volunteers before heading inside. There I enjoyed the excellent food, and the excellent friends, which were assembled there.

After a fun post-ride party, I got in with the first group home. The ride back across the city from Roaring Mouse near the Golden Gate Bridge at night was a good cool-down from my run and the long trip back from Santa Cruz. My Garmin Edge 500 was incapacitated during the ride, however: it was too busy updating all of its FIT files with daylight time. I have no idea why it does this instead of using universal time. But I'm not a developer.

Five weeks to Woodside Ramble 50 km

Five weeks to go until the Woodside Ramble 50 km trail run, which means prime time is now for building my fitness. I have at most 3 weeks of build ahead before I need to taper the distance.

Following the Lake Chabot 30 km race, I took a planned week off from running. This was originally going to be 6 days off, but I extended it to seven to do a Saturday bike ride to Mount Diablo with Patrick and others from SF2G. That capped off a week in which I did two solo rides to work, totaling 153 km, making for 276 km total on very little cycling base. That's not exactly "recovery", but at least my running-specific muscles got a break. But it left me tired, rather than fresh and ready to go, on the Sunday of the weekend after Chabot.

So a modest 12.8 km on Sunday (actually a bit more due to Strava app problems), then a rest on Monday, but 14.8 on Tue, a solid 22.8 on Wed, Thu another rest, only 1.7 on Friday, but a year's best 32.6 km on Saturday, then a straight-up, straight down 15.6 on Sunday put me back into the thick of it on my CTS and ATS score.

Here's my plot, where for "ATS" I use an exponentially-weighted average with a 7-day time constant, and "CTS" a 42-day time constant, using a "TSS" (training stress score for the day) of just km run or hiked (no hiking in the recent past).


The plot shows 100 days. The CTS trend line of 0.43 km/week was fit from 100 days back to 2 weeks back. After my post-Chabot off-week I've got CTS back to that trend, almost 7 km per day. In 5 weeks I should be able to push that up 2.1 km further, to 9 km per day, but I'll want to make the substantial majority of that progress in the next 3 weeks before I start to taper on distance. Even a "taper" on distance can continue to improve the CTS: I just need to keep ATS (red) at or above the CTS line, and there's plenty of margin there.

Friday, March 7, 2014

Winter Allaban part 10: shake-down

Continued from last time...

After leaving home I took off for the Embarcadero, where there is an extended section of cobblestones along the MUNI rail tracks. I was really looking forward to how the 30 mm nominal, 32 mm measured Strada Bianca tires and high-rake fork took the rough surface compared to my road racing bikes.

And, of course, it handled it quite well. But it was still somewhat uncomfortable. No super-plush ride like that Jan Heine likes to brag with his 42 mm by 650B tires on his custom Rene Herse (Boulder Bikes). But I was fine with that.

After the cobbles, I came to the Maritime Museum. There's some vestigal tracks here, no longer used, which can cause issues for narrow tires. I rolled over these tracks, even at a relatively oblique angle, without problem. The wider rubber was much more forgiving.

Fort Mason was next. This is a short, relatively steep climb which gave a chance to see how it went up. And it passed: the geometry, despite being smaller trail than I've ever ridden before, tracked nicely. It was really fun being on this bike.

Jan Heine likes to talk about bikes which "plane", his word for when the frame flexes with the pedal-stroke, sort of a frame-flex version of eccentric chainrings. Indeed, he likes bike with exceptionally narrow tubes to maximize this flex. The tubes on this bike are light, but not as thin as the ones Jan prefers. I don't know if I felt any "planing", but it certainly felt good enough on the Fort Mason climb.

For the descent, I took the dirt trail which goes around the green at Fort Mason rather than the conventional paved route. It did well on the dirt. No problem with debris getting caught between the tires and fenders.

I did notice one thing at Fort Mason, however, which was that there was overlap between my shoe and the fender. Jan Heine would not approve but it doesn't bother me much. The first road bike I owned had ferocious overlap and I quickly learned not to turn my front wheel into my foot. It's second nature, to my detriment mountain biking, because you need to do this when riding around switchbacks. Mountain bikes obviously are not designed with overlap. But while this bike is for many things, I don't plan to ride it through too many switchback turns.

The next test was the Golden Gate Bridge. Going around the bridge supports invaraibly provides a cross-wind challenge. and while the winds weren't strong on this day, I still got a feel for how it handled in the cross-winds. Jan Heine argues that low-trail geometries work well in cross-winds for two reasons. One is that the lever arm of for the wind force against the front wheel relative to the steering axis is less if the rake is less. Additionally, the steering moment induced by bike lean is less if the trail is less. So generally lower trail will result in less of a steering challenge with gusting cross-winds than larger trail. The bike passed no problem.

Across the bridge I paused in the parking lot to take this photo:

at bridge

Now was the time for some legitimate climbing. I took the bike up the Conzelman climb. This winds its way up the hillside just north of the Golden Gate Bridge. It's a fantastic climb, unfortunately diminished by an excess of tourist car traffic. I tried to maintain a steady effort here, but my time was slow. No magical power boost offsetting the greater gravitational burden! But then my legs were tired from running, and my cycling was off from neglect.

I got to the top without issue. Then came the equally interesting test: descending the west side. The west side of Counzelman is a one-way narrow, winding, steep descent with world-class views of the rugged coastline. But the focus wasn't ont the views, but rather how the bike handled. And the answer was very well. I wonder if the 2 cm of extra handlebar width helped a bit with the control, but a greater factor was the low trail, which in theory contributes to the bike steering with less effort than longer trail bikes. Meanwhile the relatively fat tires provide more confidence, that if I hit a pothole or rock on my descent it's far less likely to take me down.

Counzelman descends to a sharp right turn onto Field Road took me past some buildings from the old army base, down past the visitor center, and then onto Bunker Road. Bunker Road, heading back east, is a gradual uphill grind. I've never made a real effort here. The real climbing begins later.

The usual way for a road bike is to turn right on McCullough. McCullough is a steady, moderately steep climb, also with excellent views, back to the logical half-way point of the Counzelman climb I'd already done. But before the turn onto McCullough there's an alternate route: the Coastal Trail's steady, moderately graded climb to the McCullough-Counzelman junction. Recent rains had left the trails slightly damp, and my bike was shiny clean, but the bike was bought to be ridden so I decided to see how I did here.

First, I paused just past the gate to where the climbing starts to take a quick shot. I think this photo captures the spirit of the bike nicely:


Then I set off. First, immediate disappointment, not in the bike but in myself. The flow from the rains had left ruts in the trail, ruts which could catch a skinny tire. My tires were a bit less skinny (although still so by Jan Heine standards) and should have been able to handle the surface. But I dismounted, in shame. Oh, the shame....

Back on the bike, I started climbing again. The bike was handling the trail very nicely. Then, just as I hit a relatively rough part of the trail... a rattling sound. I tried to ignore it. Surely it was just an anomaly and would go away any second. But of course it didn't. My momentum was lost. I had to stop.

And this is what I saw:

One of the fender screws had come loose. Obviously having a fender stay flap around in the breeze wasn't the best idea: I wanted to find that lost screw. It had been attached with a lock washer, but perhaps I'd failed to tighten it enough. Or better yet, perhaps I should have used Loc-Tite.

I backtracked, slowly, to the start of the rough section. That must have been where the screw came free. Then I went a bit further. No luck finding the small screw on the dirt. So I turned around and retraced my pedal strokes back up the trail. Still no luck. This wasn't going to work.

I decided to ride carefully the rest of the trail then try to find a piece of string or wire to clamp the stay shut. Then I'd stop at Box Dog Bikes on the way home and get a proper replacement. The ride to the top went without incident: the loose stay didn't appear to be a hazard of flying into the spokes. I regretted that I'd failed to replace the zip tie I normally carry in my tool bag the last time I'd used it, which was to ironically fix a fender on my rain bike (an old Trek 1500). A zip tie would have solved the problem.

I made it to the top of the trail then began the paved descent of Counzelman. This also went fine. At the bottom is a parking lot where I thought I had a good chance to find something useful. So I searched the dirt verge to the parking lot. It was strewn with strange pieces of flimsy string, obviously left over from some construction project or something. But they were too old and flimsy to work. Maybe a paper clip or wire would turn up. I didn't find one, but instead I found.... a bolt.

Surely the bolt wouldn't work: the threads would be wrong. But I had to give it a try. I put the bolt up against the C-clamp holding the fender stay in place and gave a turn. It mated without difficulty. Amazing!

It was long, but there was clearance, and the fender stay was held in place better than ever:

on-road repair

the rest of the ride home went without incident. Along the way, as planned, I stopped at Box Dog where admiration was expressed for the bike and a screw and lock washer were donated. I made sure to tighten it this time, as well as its mate on the opposite side.

Box Dog

From this and a following loose bolt experience, I've concluded that especially for the primary fender bolt at the fork crown, Loc-Tite is the best approach. For the stay bolts, at least I can easily check those for looseness, but I'll probably Loc-Tite those as well.

So the shake-down ride accomplished it's nominal goal: the bike was shaken and a weakness was identified. But the ride was fantastic. I couldn' wait to take it out again. My only regret was that my training is focused on running right now, so truly worthy adventures will need to wait.

Thursday, March 6, 2014

Winter Allaben part 9: getting ready for shake-down cruise

After the photo shoot, Eric took his saddle back, packed the Winter Allaban up, and shipped it to me. And it looked even better in person than it looks in the photos. Really, it's stunning. A really quality paint job such as the one Kim Anderson did, even in monchrome, is so much nicer than the usual production bike work.

Putting it together presented a few issues. One was I was unfamiliar with the Paul Racer brakes. I accidentally detensioned them, and not realizing how tension was controlled, I was unable to get the brakes to retract properly from the rims. After carefully reading the on-line instructions for the brakes I finally figured out how to tension the springs which are responsible for pulling the brake pads away from the rim. The cable pulls them toward the rim, the spring pulls them away. The latter would have allowed dispensing with the brake bridge. But he said he uses the bridge and single bolt mount on his own bike. So I have to be happy with that.


The next issue was my usual bad habit of not tightening bolts enough. I put the Bebop pedals on, mounted my bike in my Keen sandles, and set off for a quick cruise around my very steep San Francisco block. But then I hit a ridge in the asphalt which has a history of testing handlebar stem bolts, and sure enough, the bars rotated. I hadn't tightened the single retension bolt enough. I then tested other bolts, and I noticed I had left several others loose in my preliminary set-up. This problem was easily fixed.

But then a more serious problem. I was checking dimensions on the bike and noticed the Q-factor (pedal stance) on the crank arms was high: 167 mm. I was checking the White Industries web site and noticed the crank was the mountain version, not the road version. I contacted Eric and he said he'd asked for the "VBC crank", and for some reason the mountain version had been shipped. So I had to remove the crank, box it up, and ship it back to Eric. He, meanwhile, ordered a road one and shipped that to me. This caused a delay of about 1.5 weeks. But the road crank came in right on spec at a Q of 150 mm. My hips are narrow and so my knees like smaller Q's, so getting this right was more than worthwhile.

So finally, it was time for my first real ride, and I wanted a legitimate test. Since it will take some shoe surgery to make the Bebop cleat fit on the Giro mountain bike shoes I'd just purchased, I mounted a pair of Garmin Vectors I'd not yet used. I was in a rush and didn't want to go through the proper installation procedure for proper power measurement, so just used them as dumb pedals. Any ANT+ data they optistically transmitted were lost in the vacuum of space.

As I set off, I found I liked the pedals more than expected. They have a better walking surface than the Speedplay road cleats, by far, and clip-in is perhaps easier than the Shimano SPD-SL's I recently sold. I didn't miss the Speedplay rotational float much, either. But even if they were easier to clip into than the Shimanos, they're still single-sided, and occasionally I'd try to clip into the wrong side. So I look forward to trying the Bebops with their double-sided entry.

I'll continue the story next time...