Jump to content
  • Advertisement
Sign in to follow this  
DeepButi

Learning methods in trick-taking card games

This topic is 5039 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

I'm looking for learning methods aplied to card trick-taking team games (bridge-like). As so, it's: a) NOT a two-player game b) NOT a complete information games -> minimax, alpha/beta prunning does not fit well. Adding Montecarlo would help a little, but that's not my problem just now. c) NOT a war-like, strategic-like game and most links don't apply. I cannot see how Genetic/neural theory apllies in such kind of games because I'm stopped on how to evaluate "moves", strategy, ... Actually my program (for a local card game) wins 46% games against humans (well, one human as ally and two human opponents), it plays ... mmm ... moderate, and I would like it to learn from history. Any suggestions? Thanks a lot

Share this post


Link to post
Share on other sites
Advertisement
Guest Anonymous Poster
One thing you may wish to investigate is Temporal Difference Learning.

Share this post


Link to post
Share on other sites
Guest Anonymous Poster
You may want to check out POKI, a poker playing agent that must deal with many of the challenges you are facing. http://www.cs.ualberta.ca/~games/poker/

The papers are your best bet for pertinant information but are very verbose. This is an active and challanging area of research.

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Advertisement
×

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

Participate in the game development conversation and more when you create an account on GameDev.net!

Sign me up!