Python implementations of Alan Turing's TUROCHAMP (1950), John Maynard Smith's SOMA (1961), The Bernstein Chess Program (1957), Leonardo Torres y Quevedo's El Ajedrecista (1912), and some related engines
Install python-chess and then either run one of the chess engines with the UCI/XBoard interface (mainly for chess GUIs), e.g.:
$ pypy3 ptc_xboard.py soma
The parameter to ptc_xboard.py, in this case "soma", selects the engine to use; if you leave it out, PTC is selected.
Parameter | Engine | Notes |
---|---|---|
adapt | Simple Adaptive Engine | Needs an UCI engine with MultiPV (Stockfish by default) |
bare | Bare-bones engine | Has piece-square tables for positional play; is the strongest engine here (with all engines on default settings) except adapt |
bern | The Bernstein chess program | Selects plausible moves to prune search tree |
newt | Newt | Includes an opening book and other experimental features |
plan | Plankalkül | Is set to a search depth of four plies rather than a single ply by default to bring playing strength more in line with the other engines |
ptc | TUROCHAMP (Default) | The Turing/Champernowne engine after which this repo is named |
rmove | Random Mover | All other engines should win or draw against this |
shannon | Shannon chess engine | Pretty much an engine that maximizes mobility |
soma | The Smith One-Move Analyzer | Swap-off instead of tree search-based engine |
torres | El Ajedrecista by Torres y Quevedo | Plays only as White like the real machine! (Random moves as Black.) |
Or run an engine with the console interface (Unicode output; enter your moves as e.g. "e2e4"; use black on white text in the terminal for correct piece colors):
$ pypy3 pyturochamp.py
See the chess GUI page for details on how to set up the chess engines in e.g. Cute Chess or Arena.
Full documentation at mdoege.github.io/PyTuroChamp/