Assign A Number?
I wrote the majority of this essay in 2013. Five years later I added some more current details and observations. This remains an interesting topic and the deeper one dives the greater the potential benefit — think about how continuous glucose monitors can produce actionable data for diabetic patients as well as folks wishing to learn about their individual carbohydrate sensitivity. That said in 2017 a fellow named Chris Matthews posted an entertaining look at sleep data collected by different devices that illustrates some of the issues I describe below. See his article on Medium
The concept of "The Quantified Self" won’t go away: individuals and their capacities represented with numbers, displayed as 0s and 1s. The promise that data may be used to turn us into meta-humans is an integral part of the concept. Rather, that promise is the premise of the marketing of this concept.
I am not against using devices to collect data regarding fitness or performance. I own several of them. I do believe that—in the current stage of development—the majority of the tools are useless for the majority of users. I will therefore object and critique until my resistance—combined with inherent curiosity—produces a genuinely useful idea or tool.
My first question has always been, "How can an App on my phone be more accurate than my own sensitivity?" The second question is, "Why would I cede the development of this sensitivity and awareness to such a device?"
With a few exceptions, I don't believe the current tools are collecting the right data (and admit that I don’t know what the “right data” is in every context), and those that do are missing the most critical element when it comes to applying the data.
Sensing and collecting is a beginning. It is the first of the 1000 steps that must be taken before a useful thesis and plan may result. Because most tools can only collect and display the data, I think users get hooked on the act of collection, then the display, and finally the comparison to their own previously collected numbers. But that's a dead end: there is no means of interpreting the numbers built in to the tool, the user isn't educated enough to do so him or herself so all that remains is the development of a unique metric and then a language to relate socially. Unfortunately, 999 steps remain.
I spoke with some pro bike racers who collect data. The difference between the regular user and these guys is that they turn the data over to someone who is expert in the interpretation and individual application of the information. This is the what is missing from the mass-market and massively marketed tools folks are so excited about: accurate interpretation.
A knowledgeable and experienced friend has been using the Zephyr Bio-Harness, which is the most sophisticated data collection device out there right now (in 2013 and in the context of general gym exercise). His conclusion was the same, "It takes a good amount of knowledge to infer performance-based metrics."
So the choke point here is understanding and then applying the data. The existing tools cannot help with this so something else fills the vacuum and it is not productive. That's often the case with the nature of man. Uninterpreted data produces a shallow, un-useful representation of experience and capability (unless we count the social aspects as useful, of course, and many do).
A shallow data and fitness profile taken as "truth" is subversive. It allows or even causes delusion vis-a-vis "real" sport performance (you do not learn how to race a bike by chasing KOMs on Strava). It produces a lexicon and subculture within the context of the device or service that displaces or becomes the experience itself. This is great for selling stuff but …
The pro cyclists use the data in an entirely different way. Externally collected data may be matched with internally measured characteristics (blood analysis, muscle biopsy, etc) to produce a remarkably accurate fitness profile. This is then quite simple to manipulate on an individual level. This data may also be applied on a larger scale. Accurate, detailed knowledge of individual racer characteristics informs team strategy: who can cross the gaps, who can sprint but doesn't have a five-hour tank so needs protection, who can work at 85% all day, who recovers best overnight, etc. Same goes for any team sport or military unit.
IMO, this is the direction research into these devices should take: external collection, plus minimally invasive yet functionally useful internal collection, and most importantly, an overarching interpretative solution.
The rush to market the concept and the devices has focused attention on commerce instead of human potential, but I suppose that comes as no surprise. Funnily enough, if I stick to the recreational (not Pro) cycling example, the use of data collection, display and communication has actually driven physical characteristics away from competitive sport performance by pushing performance within the context of the App or service itself.
The Strava App—which is actually a social media platform rather than a quantified self tool—wherein collected data is matched against actual terrain features and the performances of individual participants, is specific to cycling and running. The competitive or comparative nature of man is exploited and users measure themselves against others. This drives folks to work harder so I am all for it but I do not believe the App and service positively influence characteristics relevant to bike racing.
Strava trophies are earned with hard, steady-state effort, related only to the TT discipline on a bike, and once the segment "finish line" is crossed, the rider may slack off to rest. Real racing is reactive among the competitors: you must respond to the behavior of others. The guy with the biggest engine and steady-state power is very rarely the winner. Yes, you can hold 400 watts for five minutes but it's not enough because at 5:01 you must accelerate to 600w for 30 seconds, drop back to 400w for another minute and then try to recover at 380w hiding in the wheels in case you need some gas to follow another attack.
On Strava the 5-minute 400-watt effort (followed by 10 minutes of coasting) might net a trophy and earn kudos. It will not produce understanding or the characteristics needed in a race. It can however produce the delusion that such exists — and a whole subculture of those for whom that is enough. Maybe it is.
I suspect that the marketers will outshine the scientists in this lucrative world of data collection-display-communication devices and the result will be money spent, delusions fostered, and ultimately, unfulfilled potential and progress on a human level, which is what interests me.
I hate the idea that a device can replace sensitivity and personal awareness. I want people to take responsibility for their development instead of handing it over to a device. But if using the device is easier and reduces individual effort it will be far more attractive than education, application, analysis, and responsibility.
I've used and use several data capturing tools and after years—decades—of trial and error I have developed a modest ability to assess and apply the data. In a perfect world the data I collect may be used to cause behavioral change.
I used a sleep tracking App for a while that (ostensibly) analyzed sleep quality by using the accelerometer in my phone, which was placed on the bed to measure movement. At the time I was excited that an external device could tell me whether I slept well or not—as if I couldn't tell on my own. I tried to use the information to adjust habits and improve sleep but ultimately couldn't interpret the data so could not turn it into useful action. A graph was not enough. I stopped using the device.
For years I measured and recorded waking pulse, using it as an indication of recovery status. Later I improved the precision by measuring how HR increased from waking prone to standing and how quickly pulse settled down afterward. I added a pulse oximeter to sample blood O2 levels. For awhile I was measuring blood lactate in the morning because some science suggested it could indicate status of muscle glycogen replenishment and that value could be used to determine training intensity for the day.
Later I used a heart rate variability monitor because it was lauded as the new, more precise metric to assess recovery status. For three months every morning save one the monitor advised me to seek medical attention. Waking HR was generally in the low-40s and blood pressure was low-normal during the period but I had consistently low variability between heartbeats. This jives with the stress I was under at the time, my generally depressed attitude, and poor sleep habits. That said I was riding my mountain or road bike most days and finished well in several MTB races (three category wins, a 5th overall and a 3rd overall). The only day the HRV monitor showed me fully recovered and ready to go was the morning of a 50-mile MTB race and, while I suspected it was an anomaly, I did not retest because the result put my head where I wanted it to be that day. In any case, I never managed to improve my HRV score and stopped using it because the numbers were unhelpful.
The value of the "numbers" delivered by these various devices paled when compared to the value I got from the questions and ruthlessly honest answers of a daily self-interrogation regarding physical and psychological condition. So far, as long as I practice honesty, the subjective trumps the so-called objective. But the subjective cannot be sold because no one believes in themselves.
So in 2015 I started working on a project that might deliver more accurate and appropriate data in the context of lifting weight and/or moving body weight. Some friends in the medical profession were developing force sensors to help analyze and prescribe the volume and intensity of weight-bearing activity as it related to healing fractures in the lower limbs. One of them was also quite an accomplished climber who I trained at Gym Jones. Together we hypothesized sensors at the contact points (feet and hands) that measured peak force, rate of force production, and total time under load.
This data could be interpreted by an experienced coach to accurately prescribe training loads, cut workouts short when quality declined, increase training frequency according to an athlete’s recovery, etc. Adapting the medical sensors wasn’t a stretch and initial testing showed some promise but before we could make significant progress I left on a job and the main protagonist on the medical side moved away too so momentum waned. It was an interesting idea yet fraught with all of the aforementioned issues and above all would require even more expert interpretation and application for it to be useful, which takes me back to the very organic concept of personal sensitivity.
Around the time I originally wrote this article I visited to Chamonix after many years away. Some guys were making a video about a new route I climbed with Andy Parkin in 1992. During the weekend we never discussed numbers. We all spoke about feeling. We shared true emotion—the heart and obsession that drove us to climb hard in the first place. When Andy and I attempted that climb I was looking for an experience that would transform me. I wanted to do something hard and scary enough to erase what I believed of myself because that belief was a limitation. I wrote recently that, "heart—and passion—is the true source of power, while muscle is just the tool we use to express it." I learned this in the mountains, where I learned to believe in the open-ended potential of the Self.
I don't need numbers because I already know what I feel and what it means but the market for tools that attempt to quantify experience and to make those experiences "efficient" or meaningful vis-a-vis personal progress remains strong. Outside of in-depth professional application, I don't think the numbers will ever be more than a stepping stone—no matter how convenient or modern they appear—because the numbers or conclusions we assign to individual capacity prevent progress as often as they are used to cause it.