A machine learning algorithm that learns and plays mancala better than my wife
- using the algorithm, beat my wife in mancala
- become the first mancala repo with at least 1 star on github search "mancala-ml"
- follows the rules of mancala
- descends at all
- beats random moves 99% of the time
- beats shelby
- Deep Q Learning currently beats random moves ~90% of the time
- Minimax currently beats random moves ~100% of the time with a depth of ~10
- Minimax can beat shelby ~100% of the time when going first. When going second, it needs a depth of ~15 to be competitive.