I recently took first place in an AI competition aimed at perfectly balancing popular board games

I compared the speed between different implementations of the solution for the 0-1 knapsack problem. I found quite a nifty vectorized solution to the problem.