Ok, so this is something I also want to do someday. It's not what Squib is at all, but I feel like we need better tools board game simulation and machine learning. In particular, I'd like to simulate the game with very basic AI, then analyze the games to look for overpowered moves (e.g. if 75% of all winners used Black Lotus).
I'm trying to do this in an ad hoc way on my own games, but progress is slow :(
I'm trying to do this in an ad hoc way on my own games, but progress is slow :(