I stopped re-solving the same problems and started a 'patterns' doc instead. Retention went way up.
For months my routine was: solve a problem, feel great, forget it completely two weeks later, re-solve it like I'd never seen it. Pure treadmill.
What finally worked: after every problem I write about 4 lines in a doc — the trigger ("sorted array + find a pair → two pointers"), the core idea, the gotcha that got me, and a link. I don't save the whole solution, just the pattern hook.
Now before a session I skim that doc for 5 minutes instead of re-grinding old problems. When I hit something new I recognize the shape faster, because I've been reviewing triggers, not solutions. Cut my "wait, I've done this before…" blank-outs a lot. Sharing in case anyone else is stuck on the re-solve treadmill.
5 comments
- 0
AnonymousAnon The pattern journal is underrated. I started tagging problems by the trick (two-pointer, monotonic stack, and so on) instead of by number, and suddenly new problems stopped feeling random. Recognition beats raw memorization.
- 0
AnonymousAnon I do something similar — a doc of 'when I see X, reach for Y.' Sliding window, two pointers, monotonic stack, etc. Way more useful than grinding 300 random problems. Do you organize yours by pattern or by problem type?
- 0
AnonymousAnon This is the single highest-leverage habit I've picked up too. One add: write the pattern in your OWN words, not the textbook's. The second I copy the official explanation it doesn't stick. How do you organize yours — by data structure, or by problem type?
- 0
AnonymousAnon Doing almost the exact same thing and it's been the single biggest change to my retention. One addition that helped: I tag each entry with the "tell" — the phrase in the problem statement that should trigger the pattern. Over time you're basically training pattern-matching on the wording, which is what actually happens under time pressure. Out of curiosity, how many entries in before you noticed the recognition speed-up?