One of my biggest time sinks during contests is sticking to a messy DP or greedy idea for 45 minutes just because I've already invested time in it, only to realize the editorial uses a 10-line math observation.
How do you train yourselves to recognize early that an idea is overcomplicated or doomed? Are there specific "smells" in a problem statement or constraints that make you immediately drop your current approach and look for something cleaner?









I personally struggled with this quiete a lot. The good way to do this now is to just first understand the problem and the system that they have given very well. Make some observations and try to move closer to solving it a bit. Then you would have sufficient intuition into what is going on and hopefully would be able to pick the right tool. Say you find an observation that you feel is correct then you might use it with a known idea say greedy and complete the coding. Hope it helps. Or did I just answer something completely different?
Nah, you nailed it. Care to drive a bit deeper into that?