A structured review process is much more productive than randomly opening hands from a database.
Start by selecting a specific area of your game. For example, you could review:
- big blind defense;
- single-raised pots;
- three-bet pots;
- c-betting on the flop;
- turn barreling;
- river decisions;
- blind-versus-blind situations;
- pots played out of position.
Once you choose one category, collect a reasonable number of hands. Ten hands may reveal an interesting mistake, but they are rarely enough to establish a reliable pattern. A larger sample allows you to distinguish between an actual leak and normal variance.
For every hand, reconstruct the situation before evaluating the action. Consider stack sizes, positions, effective stack, preflop action, board texture, opponent tendencies, previous history, and the ranges that each player can realistically have.
Only after establishing this context should you evaluate your decision.
A useful review question is: "What was I trying to accomplish with this action?"
If you cannot clearly answer that question, the decision deserves additional attention. Every bet, check, call, and raise should have a strategic purpose. That purpose might be value extraction, protection, bluffing, denying equity, controlling the pot, inducing aggression, or realizing equity.