tag:blogger.com,1999:blog-4837627029742618106.post6993775208608998965..comments2016-12-23T17:17:02.688-08:00Comments on Computational Backgammon: Frustrating resultsMark Higginshttp://www.blogger.com/profile/16878070026852488955noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-4837627029742618106.post-33277793505372125802012-01-12T12:30:37.486-08:002012-01-12T12:30:37.486-08:00The real reason here: when I was evaluating possib...The real reason here: when I was evaluating possible moves to see which one is the optimal move, I was calculating probabilities assuming the player held the dice. This is wrong; after the player's move the opponent holds the dice. So the equity evaluation on the possible moves needs to assume the opponent holds the dice. I fixed that (eventually) and now training is much cleaner.Mark Higginshttps://www.blogger.com/profile/16878070026852488955noreply@blogger.com