The title says it all. Here, I'll start:

Submitting your solution for a problem after testing it on example input a thousand times and it working properly

"Wrong answer on test 2"

Is it just me, or are the differences between the two ratings too big? It goes from basic arithmetic (9A) to quite hard optimizations (1875A, i got tl'd on it). Maybe lower ratings for some of the 800s will help imo, the variance in difficulty is just unreal.