My question concers the post-problem analysis techniques used to extract reusable patterns from problems, both solved and unsolved. I am considering journalling the complete thought process after each problem-recording how i interpreted constraints(eg n<=2e5 implying O(nlogn) as maximum time complexity), how i classified the problem(greedy,DP,binary search,ad-hoc), and what observations led to the final approach

I would like to ask experienced and high-rated problem solvers:what methods do you use to maximise pattern recognition and long-term retention during post-problem analysis? How do you ensure that insights transfer across problems rather than remain isolated?