They had a notebook full of explanations, but every NBME felt like a new set of traps. Review was long; their decision process stayed unstable.
Review shifted from “what was the topic?” to “what exact clue should have forced the answer?” Misses were tagged as trap, misread, threshold, timing, or content.
They learned to name the pivot clue before looking at the choices, then use the explanation to repair the specific miss instead of rereading everything.