2019-07-22

Hole In The Data

Over the years, I've really come to love cycling. As any cyclist will tell you, there is something magical about pedaling rapidly down a long, scenic road or path with the wind in your face and the sounds of the world around you. It is a thrilling way to travel, and a wonderful way to get some exercise.

In that spirit, I recently decided to incorporate a long bike ride into my weekly workout regimen, on Sundays, in lieu of a "rest day." Cycling is far less aerobically intensive than running, at least when you cycle like I do, and so it has the overall feel of an active rest day.

As longtime readers of this blog know, my bicycle is a single-speed Windsor Clockwork. Riding a single-speed bike means that my average and maximum speeds are limited on the upper end. While people riding multi-gear bicycles can easily achieve speeds above 30 miles per hour on flat ground, such speeds seem virtually impossible to me. That was an intentional choice, by the way; when I first started shopping for bicycles, I was so uncomfortable on thin-tired road racers that I wanted to cap my top speed as an added safety measure. It's hard to smash your face if you're never going fast enough to cause real damage in the first place. Another consequence of all this is that I tend to expend more physical effort at the same level of speed than riders with fancier bikes. Again, I knew this going into my initial purchase of a single-speed bicycle; having a more difficult time pedaling the bicycle means that I'll get a better workout than others. No "cheating" by down-shifting on the uphills, and 30 miles of riding is a good, hard ride for me. And, of course, the frame itself is made of steel; comfortable, durable, and inexpensive, yes, but certainly not light or fast.

In light of my bicycle's inherent challenges, I've set sane goals for my weekly ride. At some point in the future, I'd like to build up to a 50-mile ride,  A.K.A a "half-century." That would say something about my endurance on the bike. At another point (almost certainly not on a 50-mile ride!), I'd like to log a ride for which my average speed is 17 miles per hour. For me, these are do-able goals, and since I have no time constraints getting in my way, they're fun things to shoot for, good ways to add some focus to my cycling when otherwise I'd just be out riding for kicks.

Yesterday, I got up before dawn, at four o'clock in the morning, to be exact. I wanted to start riding before the heat kicked in, and I also wanted to be back in time for breakfast. I hit the road while it was still dark outside, attracting moths and gnats and bats. I rode through the industrial part of town, where roads are wide and sparsely trafficked on Sundays, then along the river here in the city to a few cultural landmarks before circling back through some of the nicer neighborhoods until I reached home.

I rolled up to my garage and my GPS had me at over 37 miles of cycling - my longest ride yet. I still had plenty of aerobic energy; the data from my ride shows that I spent almost all of the ride at or below heart rate zone 1. But, after 37 miles of single-speed work, my legs were exhausted. My quadriceps burned with lactic acid and they had long since lost their "bounce." I burned over 1,300 calories while riding and gave my leg muscles a good, hard workout. It felt great, and I'm already looking forward to next week's ride.

Reviewing my statistics on Strava and Garmin Connect, however, I was a little disappointed to find that my 37-mile ride, my longest ride ever, my great Sunday morning excursion, was basically scored as "no activity" in my health and fitness statistics. The reason is entirely a data classification issue.

That is, at both Strava and Garmin Connect, my primary activity type is set to "running." This is as it should be, of course, but the problem with that is that my fitness and training load statistics are thus calculated only from running activities, i.e. activities tagged explicitly as "running." I could do thirteen hours of intense bicycle intervals across 300 miles of tough terrain, and it wouldn't impact my "training load" or "fitness curve" at all. In fact, both my load and my fitness curve would decrease that day, since I would have failed to log a "running" activity.

Since I've grown accustomed to monitoring my training load using my Strava and Garmin accounts, I was chagrined to put in an awesome ride yesterday morning and not get a little numerical "boost" for my good, hard work. In some ways, this is a problem with the way these companies calculate their statistics.

In other ways, though, it is a purely epistemic problem. I don't need to log on to the internet to know that I had a great workout yesterday. I don't need to see my "fatigue curve" climb in order to know that my muscles are tired. I don't need to see a little trophy graphic to know that it was my longest ride ever and that I climbed more vertical feet than on any previous road ride. I don't need to see a "Relative Effort Score" of over 100 to know that my two-and-a-half-hour ride was more exhausting than last week's tempo run. But the brain sees numbers and processes them as cardinal values, anyway.

The truth is that I know how hard my ride was, what a great workout it was, how much good it did me as an athlete, and how much fun I had doing it. Internet algorithms might not know it, and internet kudos values might not reflect the intrinsic value of the ride to me. But this is one problem technology cannot solve. In at least this regard, we may have been better off in the days before GPS watches, when all we really had to go on was time, distance on the (paper, hard-copy) map, and the subjective fatigue we felt after the fact. We never felt bad about not showing a numerical improvement, because there were no numbers staring back at us. We had memories, instead.

Well, it's an epistemic problem, but not an insurmountable one. Just as the brain can be trained to think carefully about its cognitive biases and deliberately work against them, so can we train the brain to accept that internet fitness data is just there as a fun thing to track. It doesn't make or break your workouts, nor even a year's worth of workouts. We have more tangible means of assessing our progress. Sometimes the absence of an uptick in the data merely serves as a reminder to us that we should pay more attention to the physical sensation of exercise, and perhaps less attention to quantitative, algorithmically calculated benchmarks.

No comments:

Post a Comment