Implementation of cribbage for two players. Text-based GUI for use on your command line (so you can appear to be working). Several kinds of opponents.
-
Random. This opponent plays a random legal move
-
Expert. This opponent uses an enumerative strategy to determine the move with the greatest possible reward (points), and takes that move. The author wrote the optimal policy
-
Student. This opponent is a deep neural network that has been trained via self-play using Google Keras and OpenAI Gym (see
/trainfor details) to execute an optimal policy
Install using pip
pip install cribbage
Or install from the source (for development, to run the tests)
git clone git@github.com:dacarlin/cribbage.git
cd cribbage
pip install .
pytest
From the command line, specify the kinds of players you want in the game
cribbage human random
You can also name the players
cribbage human random --name1 Osha --name2 Otto
If you would like to try playing one of the AI opponents, you can run with the following
cribbage human expert --name1 Alne --name2 Arin
Both the Student and the Expert players achieve performance that is better than most humans.
Run cribbage --help for an explanation of the command line arguments