These sites seem to suggest that unless current MPA is below profile MPA (by doing work above TP), the model assumes that your entire power profile is available and there has been no fatigue.
Ok, can I ask a few questions about that?
How does the model differentiate between the difficulty of a 30 second maximum effort and the second or third effort with sufficient rest to bring MPA up? The second is a lot harder, but it appears to be assuming you should have the same power available? What about a hard effort after 3 hours sub LTP?
How does the model accommodate for the adaptations given by a longer workout vs a shorter one? Would it make sense for XSS (aka fatigue) to increase with time without a break? Obviously other measures of strain (e.g. TSS) don’t account for this, but the fifth hour of a five-hour ride is vastly more fatiguing than the third hour even if nutrition needs are met. Has anything like this been considered before?
e.g. 60 Minutes at a power is identical XSS to 2 * 30-minute efforts at the same power. It appears to be reasonably well established that the adaptations and stress placed upon the body are quite different, even if those workouts were done in the same day. Is the model’s assumption more accurate? Xert - Active Recovery 60 (xertonline.com) Xert - Active Recovery 30 (xertonline.com)
Asking out of curiosity and to better understand the model!
I think the short answer is that in your Xert power-duration curve there is an allowance for long term fatigue (it doesn’t assume you can actually ride at TP forever), however in workout analysis and live MPA estimations the model does not allow for this. A related point is that from an XSS and therefore expected adaptations perspective, getting 100XSS riding at 95% TP is ‘the same’ as getting that same 100XSS riding at 70% TP for longer (both are purely ‘low’ XSS even though the likely recovery need and adaptation is different… the former would be in the ‘hard day’ category of most polarisation approaches but here it’s not).
It’s something they’ve been looking into for a while (search forum for Xert 2.0) but understand it’s taken a back seat due to focus on Forecast AI. It’s also not easy to model as e.g. fueling and perhaps form come into it too
@jjamesv Thanks - but not quite it - although it does show that the idea of durability is understood.
@wescaine That makes more sense and I think that’s what I’m getting at.
in workout analysis and live MPA estimations the model does not allow for this
Or in the accumulation of XSS.
I totally appreciate that no other measure of difficulty or strain encompasses this. The second hour of TSS is the same as the first hour of TSS. However - nobody prescribes work just based upon the TSS in the way that XERT does based upon the XSS (and high and peak loads).
I wonder if it would be possible to - like with ‘difficulty’ to exponentially increase XSS over time. Each person would have a “fatigue” or “durability” constant for Low, High etc (showing how well they could reproduce efforts for the same metabolic cost after fatigue).
Thus, your first hour of Z2 might incur 60XSS, but the second hour might be 65XSS. Or, if you had poor endurance durability, it might be 75XSS.
That way you could prescribe someone the optimum load within a given set of available hours (2hrs today or 1hr today 1hr tomorrow?)
Durability as we see it has a couple of dimensions. One is how the depletion of glycogen affects the fitness signature values. Unlike the way it’s currently evaluated - based on how mean-maximal-power changes after having performed a certain amount of kJ of work - we’ll take a more detailed/useful approach as we’ll don’t look at kJ. We evalulate how much glycogen has been utlized. If the athlete properly fuels and can match the glycogen that’s been depleted, we’d expect to see less reduction in fitness signature parameters and this would reflect in better breakthroughs later into longer rides. @wescaine You’re absolutely right that we have postoponed this slow recovery aspect of our analysis since Xert 2.0 will herald in an entirely new paradigm that we believe will allow us to model the energetics of cycling more closely as they relate to metabolism.
The second dimension and something that’s we see as an essential part of training is that durability improves with training load. Imagine two athletes with the same signature and weight but one has been training 25 hours a week whereas the other has only been training 5 hours. If you ask both to race a 5 hour road race, who do you think would fare better? That seems obvious. If you ask them to both race a 1 hour race, bets are off. Yes? This is encapsulated in our Event Readiness assessment currently shown when you set up Forecast AI for an event. It’ll become a key part of the new Race option as well. We’ll get more into how all this works once we release Forecast AI with the new Race option. In simplest terms, training affects your fitness signature but also your ability to do more work/handle more strain, i.e. durability. The more XSS strain (low, high or peak) you need to handle, the more trained each will need to be in order to have sufficient durability for the event. Forecast AI will enable you to train both using the new Race option.
