This is an argument for a more nuanced fatigue model, not making the existing system pointless by capping it before serious symptoms occur.
If you understand the mechanisms here we need outlines of how it works in reality more than we need implementation details.