I've been poking around with leetcode very occasionally for a couple years. I've got a lot of projects under my belt but I don't see trees/graphs/etc. very often.

I usually end up being very close to having the right solution to a leetcode problem but I overlook something, e.g.,

• Off by one errors
• It works for a node in the middle but fails on a node at the end or the beginning due to a silly mistake (forgot to check for null, etc.)
• It's not the optimal solution (I still don't have a good feeling of when to stop optimizing). A few times I gave up on a problem assuming my solution was "too slow" when it was in fact the optimal solution!
• Knowing when I just have to sort the inputs

I've read probably 3/4 of The Art of Computer Programming, I just bought Concrete Mathematics, I read "On Mathematical Induction" (the paper from 1960 by Henkin, it was a neat read). I know this is maybe not 100% applicable but I like math.

I feel like this is sort of where a Mathematical Induction mindset (base case, then n to n + 1) would help, and I can very easily do mathematical induction for e.g., summations (and I can even prove correctness), but I kind of have trouble visualizing it for programs. I know a for loop is basically a recursion.

I read a few chapters from about 5 algorithms books, and worked their problems, I kinda notice they tend to copy/paste each other at this point.

Also I don't know if it's just me or not but I don't notice any significant change in difficulty between Easy, Medium and Hard. Weird, huh?

The obvious answer is "just keep working programs and you'll find a pattern" (and I plan on doing that, as well as reading more books) but if anyone has any tips let me know.