matafan a écrit :It shouldn't really matter that the efforts are not exactly the same. The fact remains that despite those efforts feeling harder and harder, as they should, MPA stays higher and higher with each interval. You can turn it one way or another, that means MPA is not, in this case, a true depiction of what's really going on. It doesn't take fatigue into account.
I can see how MPA could be useful for short intervals well over threshold though.
Also, how does it differ from anaerobic capacity (W')?
You are correct that getting to MPA is hard and that it can feel like it gets harder and harder as a workout or race progresses. Some of this can be attributable to other fatigue mechanisms that are not captured *yet* such as lowering glycogen levels or increasing body core temperature. MPA analysis can assist you in identifying the occurrence of these other limiters.
What you'll find is that training improves both your fitness and your ability to express it. What this means is that you train to improve your limits of performance but you also train to be able to reach these limits more readily.
MPA is similar to W'bal modeling only instead of giving you a somewhat arbitrary intermittent capacity available that can only be validated when it reaches 0kJ, Xert determines MPA from available capacity and thus can be validated at any moment simply with a maximal effort. This can occur during a sprint, all-out effort when fatigued, or being unable to sustain the target power during an interval/time-trial, for example. So if your data has frequent maximal efforts, Xert can more readily determine your fitness signature variables. This also means that you may likely never need to do a structured training test (such as a 20 minute FTP test) or that the pacing and execution of such as test may not be so important.
There are many benefits to using this approach and we're only getting started with the many applications.
We are seeing very consistent and repeatable fitness signatures from athletes doing a broader variety of training - from crits to 20 minute TTs - all yielding similar fitness signatures for them, sometimes identical even. This is indicative that our model is working well and that people can trust the MPA values that are being derived.