How to study this material without drowning

A repeatable loop that beats reading solutions and hoping they stick.

The failure mode is predictable: read a solution, understand it, feel productive, forget it in a week. Understanding is not recall, and recall is what you need at a whiteboard.

The loop that works

  1. Attempt cold, for twenty minutes. No solution tab. Write the brute force if that is all you have.
  2. Name the bottleneck. Which operation repeats, and what does it cost?
  3. Only then read the solution — and read it for the idea, not the code. "Trade memory for time with a hash map." "Two pointers because the array is sorted."
  4. Close everything and rewrite it from scratch. If you cannot, you learned the words and not the idea.
  5. Write down the trigger, not the answer: "sorted array + pair sum → two pointers".

Space it out

Redo a problem the next day, then a week later. Twenty problems solved three times beat sixty solved once. The second attempt is where the pattern becomes yours.

Build the trigger list

Your notes should end up looking like this, short enough to reread in ten minutes:

Signal in the problem Reach for
"k-th largest / smallest" Heap
Sorted input Two pointers or binary search
"count occurrences", "seen before" Hash map or set
Substring or subarray window Sliding window
"all combinations", "all paths" Backtracking
Overlapping subproblems Dynamic programming
Grid or network reachability BFS or DFS

Track your errors, not your streak

Keep a list of mistakes you actually made: off-by-one on hi, forgot the empty input, mutated a collection while iterating it. That list is worth more than any problem count — it is the shortlist of bugs you write, and rereading it before an interview catches most of them.