An engine for playing the board game Go.
This is a Monte-Carlo tree search based engine, with very basic heuristics for the playout.
Days since birth (Jan 22, 2016) | Elo (1000 playouts, 0 elo is random mover) | Net Speed Difference |
---|---|---|
41 | 28.6 | 100% |
47 | 39.7 | 150.2% |
50 | 310.2 | 140.7% |
63 | 506.0 | 134.8% |
90 | 505.7 | 202.5% |
95 | 1366.4 | 246.4% |
105 | 1607.2 | 87.6% |
111 | 1752.7 | 88.3% |
124 | 2204.7 | 98.3% |
A special thanks to Donnie Pinkston (Caltech), who has given me valuable advice throughout this project.