stuck in a fitness vortex: another OLH

The past weeks have been dominated by priorities other than cycling or running -- the occasional ride, including a few in the gorgeous Amalfi Coast of Italy, but a lot of moving to a new neighborhood in San Francisco, travel, recovering from jet lag, and solving problems at work.

So after resolving to "make time" for yesterday's Noon Ride Old La Honda, my rational expectations, my self-assessment of fitness, were low. Indeed, after having missed the Red Kite Patterson Pass Time Trial the weekend prior, an event I normally would have been excited about especially since it had a 3-4 master's category in which I'd have the reasonable hope of getting a top 3, I wasn't too disappointed. I'd not even looked at the racing schedule. I don't believe in racing without proper preparation. Long rides? Nope. Interval sessions? Not unless you count one ride in Amalfi. Extended climbs? Virtually none. At least my weight was fairly good, within 1 kg of my 57 kg target.

That morning I had little margin for making the train, and thus had to grab the Ritchey Breakaway and roll. Tire pressure was low (approximately 65 psi versus my nominal 70-85 psi in 26 mm tires, although I'd take that up to 90 psi if I was planning on doing OLH) and the chain was filthy: the Pro Lube I'd been using seems to pick up more dirt than the Rock & Roll red that I'd used before. Rolling resistance from low pressure would at least show up on my Powertap, but drivetrain loss is truly lost to the black hole that is unrecorded power.

But OLH beckoned, so off I went.

At the start, I failed to recognize most of others. Stefano from Low-Key was there, as was Chris. Both were clearly stronger than me: I've been closer to Chris than to Stefano, but certainly not now. Stefano had a teammate in a UC Santa Cruz jersey.

As we rolled, Stefano told me he'd been training for duathlon world championships in Switzerland, although now he'd not be able to go. Still it was clear he'd be fast. I was shocked when he told me he'd never before climbed Old La Honda. Experience on the climb, like virtually any climb, counts for a lot. I was interested what sort of time he'd do.

Unlike in previous Wednesday Nooners this year, I didn't feel compelled to take any pulls prior to the base of the climb. The pace was decent but not challenging: probably ideal for doing a decent climb. I turned onto Old La Honda toward the front, in good position. And then the explosion occurred.

Stefano's Santa Cruz teammate took off in a full sprint, Stefano and Chris in tow. Behind there was a wake of destruction, riders trying to follow. I simply shook my head. No way... simply no way.

I was basically solo from here. I looked down at my Garmin twice in the opening 500 meters: first I saw 290 watts, then 310. The third time I looked at the average power: 262. 262??? How was that possible. This was an extreme example of sampling bias: I tend to look at my power when I'm producing relatively more of it. I made an effort not to increase my harder efforts, but to spend less time slacking.... not that I recalled slacking, but the average didn't lie.

Tracking average is a dangerous thing. It's very easy to over-correct. Average's too low, so go hard until it's back where you want it, but in so doing you've dug yourself into the red zone, and then it fades again. I didn't want to make that mistake.

But I sort of did. The average quickly rose to 268... 270... 272. And then it stopped... then reversed. 271 ... 270 ... (ouch!) 269. By this point I'd hit 5 minutes to go, when I knew I'd get the end-game boost, unless I'd totally fried myself. Sure enough... 270... 271.

Toward the finish, I had Stefano's teammate in my sights, closing the gap. He'd done his job toward the bottom, and was on survival mode now. We finished fairly close.

My final numbers: 18:52.14, 271.48 watts. The more things change, the more they stay the same. All 4 OLH's in June-July have been within a 17 second, 6 watt window, with the last 3 within a 5 second, 3.3 watt window. I've been remarkably consistent.

The accumulated average power plot is one I used to really like, and I still think it's useful. But the limitation is that just because your average power is below target, or even below the eventual average, doesn't mean you're not riding too hard. As an example, if I coast for the first 10 seconds, then go 10% over target for the next 90 seconds, I'm still at only 99% target power for those 100 seconds, yet I've obviously been riding at an unsustainable pace for the preceding minute and a half. The smoothed power will show this. On the other hand, smoothed power loses some of the big picture, making it hard to separate the trend from the short-duration fluctuations which naturally arise from variations in the slope.

I had a bit of goal creep on this sone.. I went in hoping I'd limited the fitness loss to a minute, but ended up, completely without justification, frustrated at my lack of ability to crack the 18:50 barrier: