Tuesday, March 25, 2014

Stories We Tell Ourselves

I have been compulsively documenting my training experiments on this blog. At present, I do six different kinds of rides: MAF Tests, Pace Rides, Long Rides, Fast Rides, Social Rides, and Challenge Rides. The first four are "training" rides. I put the word "training" in quotes because it has too large a connotation of painful hard work and sacrifice; by and large I really enjoy my "training" rides. The fifth, the social rides, are easy rides my wife and I do together and I would not give these up for the world. Although these are my easiest rides, I believe they also contribute to my fitness. The challenge rides are rides like the Tour de Pink, the Back Roads Century, and brevets. They are done for fun, but also, as their name suggests, as a challenge: can I successfully prepare for and complete these rides? Not only are these rides fun in their own right, they make my "training" rides more fun by giving them a purpose: to get ready for the challenge rides.

Of all my training rides, the MAF Test has attracted the most criticism from my readers. It is a 45 minute ride (with another 45 minutes total of warm up and cool down before and after) ridden within Heart Rate Zone 2 on Joe Friel's Heart Rate Zone scale of 1 (slowest) to 5 (fastest). By comparison, my Pace Rides are about as long as my MAF test rides but are ridden significantly faster (perhaps Zone 3), my Long Rides are ridden at about the same speed as a MAF test or perhaps a little slower but are up to 6 or 7 hours long, and my fast rides are ridden as fast as I can go (Zone 4 to Zone 5) with anywhere from 6 to 30 minutes spent at this higher level of exertion. Thus, I view MAF Test Rides as relatively easy; strenuous enough to provide some training but easy enough to allow for recovery.

The criticism from my readers, if I understand it (a big if), is that I am underestimating the difficulty of my MAF Test rides. As a result, I experience no recovery on a day I ride a MAF test and if I do too many MAF tests, I am riding fatigued, which is harmful to fitness. The explanations of my critics as to why I am misled about the difficulty of my MAF Test rides vary, but include the following:
  1. My Heart Rate Zones are set at heart rates which are too high, so when I think I am riding in Zone 2 I am really riding in Zone 3.
  2. Due to chronic fatigue, my heart rate is no longer a good measure of my exertion.
I want to start by saying that I take the suggestions of my readers very seriously, even when I argue with them, and am constantly re-evaluating my training based on (among other things) these suggestions. That said, I have previously discussed these theoretical issues ad nauseam and so won't repeat my rationalizations for rejecting the above two assertions but rather want to come at this from a different angle. For reasons both deliberate and fortuitous, my training schedule has varied over the years. If I look at MAF Test results and compare them to those changes, will I see patterns that suggest MAF Test rides are relatively easy and allow for recovery, or will I see patterns that suggest MAF Test rides are causing me to over-train? Unfortunately, this approach is susceptible to "over-fitting", a technical term which describes the problem when you have too many ways of explaining things and too few facts. I might look at the results from my MAF Tests and tell one story while someone else might look at exactly the same ride data and craft a completely different one. In short, this whole post is a plunge down the rabbit hole of subjective nihilism. Unfortunately, I believe that this subjective nihilism is inescapable, so onward into the abyss.

MAF Test results (blue, in miles per hour, scale on the left side of the graph), wind speed (green, in miles per hour, scale on the right side of the graph), and cumulative training effort (red, in arbitrary units.) The gray line through the MAF Test results is the linear best fit trend line. The date of the MAF Test is indicated on the bottom of the graph, ranging between January 2 and March 20 of 2014.

The rather complex graph above is my usual MAF Test plot in blue overlaid with two other kinds of data. The green bars indicate the wind speed as reported by Yahoo Weather for the City of Houston on the days and times of the MAF tests. The red graph is my effort to indicate how hard I was training leading up to the MAF test. In Joe Friel's Training Bible, he suggests a metric of training volume which is calculated by multiplying the heart rate zone at which a ride is ridden times the hours the ride is long. Thus, a two hour ride in zone 3 would get a training volume of 6. To make this easier to look at and perhaps more meaningful, what I plotted was the average training volume for the week preceding the MAF test; the red line is that running average. So that I could talk about this graph, I marked five points on the graph with a 1, 2, or 3.

Story 1: Wind decreases MAF Test results

I begin with this story because it is simple, obvious, and hopefully uncontroversial. On a flat road, the biggest factor (by far) influencing bicycle's speed is wind resistance. The relationship between wind speed and bicycle speed is complex but I think it is not totally unrealistic to assume that the a headwind will reduce a rider's speed by at least the same amount that the equivalent tailwind will increase it. Consider a flat, out and back course, 17 miles long. On a windless day, I might be able to ride that in two hours, one hour to ride the 17 miles out at 17 miles per hour, one hour back. On a day with a stiff headwind going out which reduced my speed to 12 miles per hour, the most the tailwind coming back could help me would be to increase my speed to 22 miles per hour. At these speeds, it would take me 1.4 hours to complete the ride out and 0.8 hours the ride back for a total of 2.2 hours. As most experienced cyclists know, the damage a headwind does exceeds the benefit a tailwind provides. I ride my MAF tests on an oval course which is similar to this out and back course in that head and tailwind are present for an equal number of miles, and so I know that my speed on a windy day will be lower than on a still day. This is confirmed by the data on the above graph. In particular, the two rides marked by a number 1 show high winds and slow rides. To my eye, every day it is windy, there is evidence that my ride is slower than it would have been otherwise. Thus, when I have a slow ride on a windy day, my story is "Ignore that poor result, it is due to the wind."

Story 2: There is unexplained variation ("noise") in the MAF Test results

Although I believe that every windy day is a slow day, I do not believe every slow day is a windy day. Note the two rides on the above graph labelled with the number 2. The ride on the date 2/6 is clearly noise; it is a slow day but not a windy day. In addition, nothing about my training volume leading up to the ride explains this slow day. On the other hand, the ride done between 3/13 and 3/20 lends itself to two different explanations, and therein lies a story. Although this slow day cannot be explained by wind, note that a drop in training volume precedes this slow ride. In fact, I initially explained this slow ride by the fact it was preceded by three days in which life prevented me from riding; three days off the bike and my fitness was gone! However, upon reflection, it is obvious that it would be rash to jump to that conclusion, this slow ride could equally be due to noise. In short, the story I should not tell myself is "Taking off as few as three days completely destroys my fitness; I must never take off more than one day a week." Rather, the story I should tell myself is "Hm, a bad day. I wonder if it is just noise, or if it is due to missing three days of riding in a row? I will pay attention in the future and see if a pattern develops."

Story 3: The effect of daily MAF Tests on performance

Note the four rides around the number "3" on the graph above. Each one is faster than the one before. The story I am telling myself about these rides is that "If I ride three or more MAF Test rides in a row, I tend to get faster for each one. Apparently, riding back to back MAF Test rides is not harmful, and may be helpful." How can I justify that story given what I have just said about noise? Because, in this case, I did pay attention and a pattern did develop over time. I have been riding MAF Tests since December of 2012, and I went over all that data and found 14 times I did 3 or more MAF tests in a row. In 10 of those cases, the last MAF test was faster than the first. In only 2 cases the last MAF test was slower than the first, and in 2 others the first and last speeds were the same. The trend is clear; my performance improves when I ride back to back MAF tests.


The point of this post is not in the specific stories, I don't think they are all that novel. Rather, the point is that the stories we tell ourselves about our training are inherently ambiguous. In that context, there are many more players I could add to my stories. I have the impression I ride slower on cold days, maybe because the heavier clothing I wear on those days or maybe because of the effort required to stay warm. Maybe I ought to discount cold days the same way I discount windy days?  I keep a daily training log, and in that log I record a subjective measure of how fatigued I am feeling. Maybe if I quantified that subjective fatigue and included it on the graph, some of the "noise" might be explained? Maybe cause and effect take place over longer periods of time than I have considered. What if a slow ride is due to cumulative training effort not the days before a MAF test but weeks or months before? The problem is, each time I add a new player, the problem of "over-fitting" becomes greater; by bringing in the effect of more players, it becomes possible to explain anything and reach any conclusion I like. As a scientist, I deplore my efforts to be my own experiment of one. As a person living a real life in the real world, I realize I have no choice but to do just that. I certainly do not claim to be a philosopher, but the same necessity that drives me to be an experiment of one forces me to develop a philosophy for keeping my experiment of one from leading me astray. My feeling is that the application of skepticism, a heavy dose of common sense, and a willingness to reconsider and re-evaluate as necessary may be the key to moving the stories I tell myself from the category of fiction to that of non-fiction.

No comments:

Post a Comment