| |
BGonline.org Forums
Mini-manifesto on evils of PR gammon.
Posted By: Bob Koca In Response To: A BMAB Move (Chuck Bower)
Date: Tuesday, 26 May 2026, at 7:26 p.m.
So PRgammon was used to choose the team. Is the London tournament itself a PRgammon tournament?
There are issues with PRgammon going far beyond the artificial tricks to up the decision count.
For winning a BACKGAMMON match the goal is not to win the PR battle but to win the match. Some styles get unfairly punished by the "better PRgammon player means better BACKGAMMON player" mindset. For example, suppose in a certain position that a wild play which leads to a complex position and a conservative play leading to a simple position have the same equity. A player whose style veers toward the wild play will average a worse PR, but his opponent would likely also. So just a stylistic difference which is neutral on average is punished. And it gets worse. If that player is the better player, the wild play is almost certainly the better practical play to win the BACKGAMMON match, but is heavily disincentivized if the match is actually PRgammon.
How many just look at PR results out of context? One stupid game which XG misunderstands can do crazy things. In the Monte Carlo Tournament Championship match a few years ago between Zizska and Sander they were both playing about 3 for first several games. Then came one super backgame and they were both around 10. Were they suddenly intermediate players that day?
The PR approach is not even needed at all. Variance can be reduced by just using variance reduced match win %ages which avoid almost all the issues of PR gammon. Nearly always would make the practical play to win the BACKGAMMON match.
Is it clear that I have strong opinions on this topic?
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.