Some Thoughts on Load Tracking

        Load tracking has been a component of athletics for as long as we really know, whether it be miles, weight x reps in the gym, hours, or more popularly as of late, effort based metrics like TSS and CTL. TSS and CTL are hugely popular, with both coaches and athletes. I use them myself on a daily basis, working with athletes and training prescription. I think they have a lot of value, however I also see people consistently place more stock in those values than I think is merited. What follows is an open table discussion, between me and myself, about the idea of “load”, is it worth tracking, how should it be tracked, what can we extrapolate from it. Be forewarned, all that follows goes no farther than opinion and trial and error. This is far from an objective science, but potentially that is where the greatest value lies.

Philosophy of Total Load

        Every athlete can run at a given “total load”, which individual to that athlete, let’s call it Load_ind (Load for an individual). Load_ind must consider not just an athlete’s workouts, but also EVERYTHING that adds STRESS to the system, such that components of Load_ind, may include:   

  • Workouts that give an objective load score (i.e. TSS)
  • Workouts that do not give an objective load score, but could be applied a subjective load score (i.e. fatigue ranking, RPE, etc)
  • Relationship stress
  • Sleep quality
  • Work stress
  • Nutrition and hydration state, in relation to optimal, a workout is undertaken in
  • List goes on, but the point being that "total load" must comprise more than what shows up in TSS and/or CTL (or similar such metrics)

        The graphic below shows an athlete, in three different scenarios. The athlete is running at their hypothetical 100% of Load_ind in each scenario, but you can see the composition is different.

Three scenarios, of an athlete at 100% of Load_ind, with varying composition.

Three scenarios, of an athlete at 100% of Load_ind, with varying composition.

        In each scenario, in theory the athlete really cannot take on any additional load, whether it is from training, health, or psychological stress. Certainly, everyone as a coach and athlete, has likely crossed over that Load_ind max point, and typically the result is unsustainable, until we return back below 100% Load_ind. The trap to fall into is thinking that if in June of 2016 an athlete handled “x” amount of training load, with no work obligations, and in June of 2017 an athlete has 20 hours per week of work, that they are going to handle “x” amount of training load in the same. That is a simple example that most people would agree with, but in reality it gets more complex and the lines between where the stress is coming from, and how to manage it, get blurred.

Load as a Tool for Forecasting Performance

        It is becoming more and more popular to use training load (i.e. CTL) as a forecasting tool for performance. When there is a possibility to replace grey area of guesswork, with black and white of trustworthy quantifiable data, I am all for it. However, I would venture to say that using CTL as the main driver in performance forecasting, and the target in driving training, such that the priority is getting CTL to a certain level, that is just trading grey area for a different grey area. Here is why:

  • Load does not account for variances in composition, such that a CTL of 100 can be achieved a myriad of ways
  • Equivalent loads of differing compositions are almost always going to result in different performances, in the same competition context. Load composition is just as important as load.
  • Load, measured as CTL, is influenced to a greater degree by aerobic powers than by higher intensity powers. Such that an athlete's perceived increase in load (and fatigue) for a weekly increase in volume of high intensity, may not be duly reflected in a metric such as CTL.

The Fix

        The reality is, if looking for a black and white, quantifiable metric, there is no fix. As I hit on with the idea of Load_ind, training load is just a slice of the pie. If there is already a lot of grey area in the total load, how much does it even matter if whatever metric we use to track training load is an accurate portrayal of how fatigue ebbs and flows?

        Well, it probably can be a more useful tool the more clear of a picture it paints. My opinion is that the more aerobically driven the training, the more closely a metric like CTL correlates to the performance and perceived fatigue of the athlete.  However, as training gets more “polarized” and incorporates a higher volume of intensity, the correlation starts to decrease. A specific example, we’ve already hit on,  a plateau in CTL, followed by a change in training composition. Athlete’s perceived fatigue changes, as do contextual competition performances, none of which is reflected in any real change in a metric like CTL, ATL, or even TSB. Are there metrics the specifically address the idea of changing how intensity is weighted, in a load tracker?

·      Yes and no. Chris Baddick has really chewed on some of these issues, and created the metric of CIL (chronic intensity load). CIL will display bifurcations from CTL, as the volume of intensity (measured by IF) changes, but Chris will be the first to argue that even IF is a really poor portrayal, as depending on the total volume of the ride, IF can easily be diluted. I continually come back to the idea of eschewing “advanced metrics” such as TSS, CTL, and IF – for more simple measures such as volume of intensity and total volume.

        Looking at training in hours, may be “old school” in this era – but it also seems to correlate, depending on the goal.  Sure tracking training by hours has an even bigger flaw in ignoring composition than CTL does, but if you combine that with weekly volume of time spent in different zones of intensity, maybe you start to get closer? It may not all feed into one catch all metric – but you can also see a black and white change in time spent riding at various intensities. Without diving into too much depth, I can say that after looking at a pool of riders very closely for three years now – continually the metric I see the highest correlation with summer competition performance is very simple, total hours ridden from November to February, of the prior winter.  Maybe cutting out the grey area means going back to simpler, but more black and white, metrics?

        The bottom line is that, no metric is perfect, and never will be. We can spend time tweaking the metrics to try and get them closer to perfection, but for me the whole idea of thinking about a “total load” – Load_ind – is that metrics do not, and will not ever paint the whole picture.  At the risk of now sounding like an infomercial, that is the value of a coach, in my opinion. The role of a coach should be to get much deeper than what is displayed by the popular metrics. The role of a coach is to cut through the grey area – which usually takes some trial and error, and it may take years for a coach and an athlete to build their successful model [aside - successful model is a total unicorn in my opinion, and if you believe you have one, you are settling]. If eliciting top performances was as easy as correlating it to a given training load, no one should have a coach. It’s not to say that having a coach all of a sudden makes it easy – but perhaps with the right relationship an athlete starts to get closer? The counter to that is that a coach can only be successful with clear athlete communication, due to the highly subjective, and oscillating, nature of everything that contributes to Load_ind. Good luck finding perfection, I certainly plan to keep searching!