How do you refine optimal trade entry for higher win rates?

OTE uses Fibonacci retracements after a sweep, but entries often fail without confluence. Traders refine it with order blocks, displacement, and session timing. What refinements improved your OTE win rate, and where does it fail?