Most coaches recommend keeping a training log, a list of rides completed along with their length. Most coaches also recommend measuring the length of those rides in minutes rather than miles. One reason for preferring minutes is illustrated by a consideration of two rides of equal mileage, one hilly, one flat. Imagine riding these two rides, riding each with the same intensity. To do that, you will need to slow down going up the hills of the hilly ride (and no, the downhills will not make up for the uphills.) That means the hilly ride will take longer and, as a result of being longer in minutes, leave you more tired (and more fit). However, if you track miles in your training log, they will look the same. Tracking minutes makes up for that and therefore is a better metric. However, let’s change the scenario a bit. After the above rides you decide to see how fast you can complete the hilly ride. Because you moderated your effort the first time, imagine it took an hour to complete the ride. When you repeat it going all out, imagine that it only took you fifty minutes. Obviously the faster ride will leave you more tired (and more fit) than the slower ride, but by tracking minutes, you get less credit for it in your training log. Aren’t coaches aware of this problem? What do they suggest we do about it? Coaches are fully aware of this fact (they would not call it a problem) and what they would say about it is that every ride should be ridden at a planned level of effort based on a well thought out training plan so there would never be a reason to compare the faster and slower rides. Exercise scientists have different goals and would take a different approach, they would say that level of effort does need to be taken into account and have developed quantitative measures for doing so. For better or for worse, I decided to start tracking a quantitative measure of the effort of my rides in my training log. To explain what I mean by that, let me start with some definitions.
Fitness is the goal of my training. It is the process of getting faster and stronger and developing more endurance. Fitness is a very complicated topic which I will describe only briefly towards the end of this post. Training also has a dark side and that is Fatigue. When I ride I develop Fatigue such that if I attempt to ride again too soon, my performance would be worse than before I trained. However, if I rest, Fatigue declines and Fitness increases and if I ride again after the proper amount of rest, my performance will be improved. Excessive training can even lead to Overtraining, Fatigue that won’t go away. Thus, it is important to do the right amount of training and not too much. How is training measured? The term I use as a measure of how much training has been done is Load.
Load = Volume x Intensity
Volume is how long a ride lasts in minutes. Intensity is much more complicated, but roughly corresponds to how hard a ride is, hard because it is ridden fast or because the route over which it is ridden is hilly, for example. Perhaps the most common way Coaches talk about Intensity is in terms of Training Zones. Different coaches use different Training Zone systems. The coach I follow, Coach John Hughes, uses a 7 zone system, where Zone 1 is the easiest, the lowest level of Intensity and Zone 7 is the hardest. The boundaries of Training Zones can be defined by a variety of indicators including Heart Rate, Perceived Exertion, Power, Blood Lactate, and Rate of Breathing. In this post I will only be discussing Heart Rate. Dividing Intensity into Zones is merely a convenience, Intensity is continuous. For example, the Intensity of the very top of Zone 2 is virtually identical to the Intensity of the very bottom of Zone 3. Training zones are a great way for coaches to communicate training plans to their athletes, but are not quantitative, they do not allow one to calculate a quantitative measure of Load. In the scientific literature, the term used as a measure of training is TRIMP (which stands for TRaining IMPulse). Confusingly, there are several different versions of TRIMP, each corresponding to a different way of calculating that number (which is part of the reason I prefer to use the term Load.) Which version should I use? In the end, I decided I didn’t like any of them and developed my own. My discussion of the various versions of TRIMP and the development of my own version is described in my blog post of August 2021 entitled “TRIMP, Intensity, and Fatigue”. How I decided to use the Average Heart Rate for a ride provided by my TranyaGo sports watch to calculate the Load (my version of a TRIMP score) for that ride is described in my blog post of April 2023 entitled “Improved Training Load Estimate”. Between October 3, 2022 and March 3, 2024 I calculated and recorded that Load in my training log. Three to four months ago, I decided to review that data and blog about how that worked out. After two months of writing and deleting drafts of that post, I gave up and wrote a blog post on how the attempt to do so had given me “Blogger’s Block”. Where did I go wrong?
I’m not sure I did go wrong. I often say on this blog that in order to think, I write, and I believe there was a lot of thinking to be done on this topic and thus the writing took a long time. What made it feel frustrating is that the process of this kind of writing is not linear. I write down an idea, think about it, decide it is wrong, and then delete it and start over. It feels like I am making no progress, that I am stuck, that each step forward is followed by a step backwards, but that feeling is missing the fact that the thinking is the goal, not the writing, and thinking often involves exploring lots of blind alleys before finding the true path. (Well, maybe not the true path, but a better path.) But what made the thinking so hard?
Consider this whole project an experiment the purpose of which is to test the following hypothesis: Does the Load score I generated by using the Average Heart Rate for a ride provided by my TranyaGo Sports Watch provide a better measure of the Fatigue generated by my training than minutes of riding?
Something else I often say on this blog is that I shouldn’t let Best be the enemy of Good. The fact of the matter is that I do not have enough data of high enough quality to determine with a high degree of confidence if that hypothesis is true or false. However, as as result of looking at the Load data in my training log and of all the thinking I did while preparing this post, I am of the opinion that the original way I calculated Load was probably not more useful than minutes of Riding, but with minor modifications I could come up with a score that does have a reasonable likelihood of being better.
How could Load not be better than minutes? As I noted at the beginning of this post, it seems like this should be a no-brainer. Tracking Load by tracking minutes completely ignores the difference between a hard ride and an easy one. If one could reliably determine how hard a ride is, its Intensity, that would be true. Unfortunately, there is no consensus as to how to determine Intensity, or even what Intensity means, or if Intensity is even one thing. For the sake of the Good and the purposes of this post, let’s ignore Best and assume Intensity is one thing. If one ignores Intensity, the Load of a fast ride will be underestimated. If, however, the Intensity of that fast ride is overestimated, so will the Load of that fast ride. Because there is no limit as to how badly one can overestimate that Intensity, Load can end up being a worse measure of Fatigue than minutes. Thus, the first question I asked when I started working on this post is, does the estimate of Intensity I am using seem reasonable? To answer that question, I decided to simplify things by only comparing rides on the same route, on the same bicycle, etc. These are the rides I have named the Cañada rides, a 17 mile ride with 1,100 feet of climbing. The lowest heart rate I have recorded on those rides is 113 beats per minute (BPM). The highest is 145 BPM. Using the equation from my Improved Training Load post, I calculated that the 145 BPM ride had an Intensity 6.8 times that of the 113 BPM ride. The effect of that on Fatigue (Load) is less than that because the faster ride is completed in fewer minutes so the Load of the faster ride is 5.7 times that of the slower ride. This is completely unreasonable. Yes, I am more tired after the fast ride than the slow one, but it is not five to six times more tired! I might believe that the Fatigue of the fast ride is twice as much as the slow, though one and a half times seems even more likely. This is where I got stuck. I have spent the last three months trying to figure out where I went wrong. The good news is that in the process of trying to explain these discrepancies, I clarified my thinking on a whole range of issues. Here are some of them:
- Heart Rate depends on many things in addition to Intensity. A few examples include worry, caffeine, and Fatigue. Thus, Heart Rate can give misleading Intensity measures. In my experience, Average Heart Rate is fairly reliable, my Average Heart Rate recorded for most rides seems to be a pretty good indicator of Intensity, but when I selected the rides with the highest and lowest Average Heart Rates, I was selecting the outliers, those rides where Heart Rate was most likely not to be a good indicator. To get around this problem, I selected my ten slowest rides and my ten fastest rides and compared the Intensity calculated from the Average Heart Rates recorded for those two groups of rides. There was only a 2.2-fold difference, much more reasonable. (The average heart rates were 137 and 120 BPM.) Is this the solution to my problem? Maybe, in part at least. This idea occurred to me very recently. Had I thought of it early on, I might have avoided the whole “blogger’s block” thing, and that would have been unfortunate. As I struggled to explain why my Intensity estimates were so far off, I reconsidered every aspect of how I calculated my Intensity and many good insights resulted.
- I stand by my post on Trimp, Intensity, and Fatigue. If I were to write it today, I would change some of the numbers a bit but the overall conclusions of that post, that the commonly used TRIMP calculations are not consistent with coaches’ recommendations or athletes experiences, would not change. What I am reconsidering is how the conclusions from that post ought to affect my day to day cycling. That is a work in progress and may be a topic for a future post, but I have nothing more to say about that in this post.
- In my post on Improved Training Load Estimate, I proposed to calculate Intensity from Average Heart Rate by using this equation:
Intensity = 0.00065 x e(0.06 x heart-rate)
As noted in Point 1 above, I don’t think this was a ridiculous choice, but I am not sure if it is better or worse than just using minutes as a measure of Load and I also think I can do better. Again, that is a work in progress and is likely to be the topic of a future post, but I have nothing more to say about that in this post. - In that same post, I noted that Average Heart Rate for a ride is not a good way to estimate the Load of an Interval Session. More recently, I have noticed that some of my rides look a bit like Interval Sessions in that they contain short stretches of very high Intensity embedded in a long ride of more moderate Intensity. This might be important because such rides might be generating more Fatigue than would be predicted from the Average Heart Rate of that ride. I did several such rides after my move from a flatter neighborhood to a hillier one in 2020 and this might be what prevented me from riding the 2021 Art of Survival. This is not even a work in progress, but rather the seed of an idea that might or might not develop into something.
- Different rides have different effects on my fitness even if they have the same Load. One might primarily improve endurance while another might primarily improve speed. Thus, I should not substitute one ride for another just based on Load. For that reason, I have started adding a sentence to my Training Log explaining, for each ride, the purpose of that ride, why I chose to ride it instead of a different ride.
Where does this leave me? As of March 4, I have changed my training log to emphasize minutes over Load. I am still tracking Load but I am experimenting with different ways of calculating that Load. Once this settles down, I’ll post about it.
