You go first. Adjust the slider to change how long it will take one each move. The longer you give it, the better it plays. In the actual competition, it was at 119 seconds. Yes, this is what I will be entering, not the other one.
Thanks to @PullJosh for making the competition and the base. Also, thanks for testing my old AI on your computer. This uses a monte carlo tree search. That means that it cycles through each of the moves, and then plays out a random game (except it always wins and blocks when it can). It keeps track of how many times each move wins, and chooses the one that wins the most. Thanks to https://en.wikipedia.org/wiki/Monte_Carlo_tree_search and https://jeffbradberry.com/posts/2015/09/intro-to-monte-carlo-tree-search/ for the theory behind this.
