by Douglas Zare
1 November 2012

When a backgammon bot disagrees with a human, it's usually right to bet that the bot is correct.
However, there are many backgammon positions the bots don't understand well, including quite simple ones.
All bots have weaknesses, but I'll show a few positions Gnu Backgammon misevaluates.
As usual, in all positions, Red bears off to the bottom left.
Here is a classic example: Rolling an outside prime home.
|
| ||
|
| ||
| The rest of this article (19.19 K) is premium content. Please subscribe below. |
Article text Copyright © 1999-2013 Douglas Zare and GammonVillage Inc.
Gnubg has neural net inputs that only look at primes out to the 9- or 10-point. if I recall correctly, so it's unsurprising that it misevaluates primes further back. I'm pleasantly surprised to hear that it can roll them home well.
I find it interesting that you show how GNU does not understand these positions, yet I believe most agree that eXtremeGammon is superior to GNU. I put a couple of your positions into eXtremeGammon and XGR+, which is less than a complete rollout, gets them correct according to your definition of correct.
I think it is important to point out that your "bot breaking" positions don't necessarily apply to all bots, and if the reader truly cares about having a more accurate bot, he might look somewhere else besides GNU.
In spite of this, I found your article, like all of your articles, very interesting.
You must be signed in to post comments.
