Modern video games are designed, or at least vetted, by psychologists. The idea is to make the game challenging enough but not to challenging so players give up.
Use the same principle to design an HFT algorithm:
Trade where the capitulation points are,
Trade where participants are likely to ‘continue playing’
Trade the liklihood of capitulation vs reversion
Build a pain and capitulation scale/indicator
It probably doesn’t have to be an HFT algorithm.
This might be another approach to forecasting regime changes and momentum days (Ovation and Peer Effects models).