A simple othello/reversi AI written in Haskell.
Switch branches/tags
Nothing to show
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.



This is a simple implementation of the Minimax algorithm for Othello/Reversi. Is my first "real" code after a few days attempting learn Haskell, so it might not be very idiomatic, elegant or readable.


Why not? Othello is a nice game to play, that requires no graphics and the Minimax algorithm can be implemented fairly easily. I had written it once before in C (and failed horribly in Visual Basic), so I knew what I was in for. It is complex enough to be more than a small exercise. Plus, I get to brag about how 1337 I am implementing AI in Haskell. I could find only one other implementation when I wrote this, and it was written in a kind-of-imperative style.


To play, just enter the X/Y coordinate as a tuple, ie. (4,2). You play as "O".

The basic datastructure is a map of an Int,Int tuple to a Piece data. This let me use an immutable datastructure, and still refer to the pieces in a simple x/y coordinate system.

The heuristics are pretty simple. The Minimax algorithm is set up to optimize for advantage in the number of pieces but most importantly, the number of possible moves. This ensure the AI will dominate the match, leaving you no room to make any smart moves for yourself. I'm not a terrible othello player, but I had a hard time beating the AI.


  • Update to be more idiomatic Haskell. I had some code review from a Haskell hacker friend, but I'll need some more Haskell time to "get" some of the point free style and evaluate what I like.

  • Implement alpha/beta pruning. This should be simple, but I've never actually done it.

  • Parallelize? I have two cores and multithreading. This should be the ideal algorithm to split over a number of threads. We'll see as soon as I get the urge to Haskell.