document updated 8 years ago, on Dec 14, 2015
I like to write game AIs, just as an exercise in programming. Usually these are single-player puzzles, but sometimes I do others.
Games I've done:
- "10", by iojoe
Smaller boards are solved, then I discovered that the problem is NP-hard, so larger boards will always take forever.
Games I might do:
- "Mamono Sweeper"
I would just look for 3x2, 4x2, etc patterns. (I think I know most of them already though?)
It could also, given a particular small section of a board, give you the probabilities of each block, to help when you're forced to guess.
- "No Thanks" by Thorsten Gimmler
The strategy is a little complex, but not tooo complex. I'd like the computer to figure out what the equivalent value of a chip is, and how it changes as the game gets closer to the end.
Implement in Go? Rust? C#? D? PyPy?
- Rummy
It seems like the strategy changes a lot based on the number of players? If there's 2 players, then it's usually a very good thing to pull lots of cards from the discard pile. If there's 4 players, then it's much riskier to pull lots of cards.