Is Training a Markov Process?

This is one of the most esoteric of questions, but one I've been mulling over for a while: Is training and racing a Markov process? A Markov process is one in which the conditions at any moment determine what happens in the next moment, but the overall history of how you got to that moment is unimportant. It's actually a bit of a free will vs. determinism question; if every moment decides the next moment, then you have no choice in what happens.

The immediate response of most people would be that, if you run a certain time at a race, it doesn't matter how you got there. The time you run is the time you run and that's that. Of course, if you used performance-enhancing drugs, then all of a sudden everyone thinks it's very important how you got to where you are. So - it's not important unless it is?

I think what's important is where you are in your training. Training responses tend to form a logistic curve:

If you're early in your training, at point "A," then if you race, you probably are going to run about as expected (not all that well, but better than in training). Similarly, if you're late in the training, at point "C," then you can be fairly confident that you're going to race well and within a narrow range of possible times. It's when you're in the middle, at point "B," that things get tricky; you're rapidly improving, so it's possible that you might have a surprisingly good race. I think most runners don't have any idea where they are, because they don't train consistently, but if they did, they'd know whether they're improving rapidly or starting to reach a peak.

This is why, if you see me running, you shouldn't ask me what I'm thinking about.