Backgammon Cube Handling in Races

Cube Handling in Races

Forum Archive : Cube Handling in Races

Pip count percentage

From:   Jeff Mogath
Address:   mogath3@aol.com
Date:   25 February 2001
Subject:   Pip Count percentage
Forum:   rec.games.backgammon
Google:   20010225163922.04686.00002037@ng-ff1.aol.com

I understand that if you're more than twelve percent behind in the pip count you should probably drop. How do you figure this percentage? Any help is appreciated. Thanks Regards, Jeff

hi writes:

Hi, If you are playing online usually there is a pip count displayed, so if the pip count for you is 100 and you are more than 12 pips behind you should drop (.12 x 100 = 12). If your pip count is 60 and you are 8 pips behind you should drop (.12 x 60 = 8 rounded up). If the pip count is not displayed, then you have to learn how to count pips, and there are many ways to do that. Look around online for the various methods. In Paul Magriel's "Backgammon" he recommends taking if you are behind 10 pips or less in a 60 pip race, and 13 pips or less in a 100 pip race. How do you know how long a 60 pip or 100 pip race is approximately? If all your men are in your inner board and equally distributed, you are in a 60 pip race (15 men, average position is on the 4pt-->15 x 4 = 60). A 100 pip race is one in which there are about 4 or 5 rolls before the bear off begins (average roll of the dice is 8 pips, 5 x 8 = 40).

Gregg Cattanach writes:

A slightly more accurate system goes like this. (And the calculations are easier). Take the leader's pip count and multiply by 10%. Then take that number, subtract 2, and that's the pip difference for an initial double. Subtract 1 from the 10% number, and that's the difference for a redouble. Add 2 to the 10% number and that's the pip difference where the opponent should pass. This applies if both sides have similar high variance positions, (not a lot of checkers piled up on the low points.) This turns out to be remarkably accurate for any pip count from around 65 up to any large number. And the math is easier :) The 8%, 9%, 12% system overestimates the differences necessary when you get up to 120 pips or more, and underestimates it when you are down below 80 pips. This 10%-2, 10%-1, 10%+2 system is based on long computer simulations of the 'single-checker model'. This method does require getting an accurate pip count for both sides. I recommend Jack Kissane's Cluster Counting method: http://www.bkgm.com/articles/McCool/cluster.html

Did you find the information in this article useful?

Do you have any comments you'd like to add?