BGonline.org Forums

Janowski's formulae

Posted By: NJ
Date: Monday, 18 January 2010, at 2:53 a.m.

I recently made a post regarding computation of cubeful equities where I made some statements to the effect that Janowski's formulae were wrong. I wanted to clarify what I meant.

First, the Janowski paper is the basis of doubling strategy for all top backgammon programs, including my own. The paper laid the groundwork for all later work to come. The concepts described in the paper, including dead-cube and live-cube model with interpolation based on cube efficiency, are still the basis for evaluated (non-rollout) cubeful equities today.

In my program, I use all of these ideas. Therefore, the cubeful equities in my program are based on his theories. What I changed was for the live cube equity to be segmented instead of linear. I wasn't the first person to do this, the GNU authors made this modification long before I did. This modification fixes some problems with too good evaluations in situations that should not have too good evaluations. The fact that there have been modifications to the original formulae in no way diminishes the value of these original formulae. It really just shows that everyone still believes that these formulae are the best way to obtain cubeful evaluations.

Messages In This Thread

• Janowski's formulae
  NJ -- Monday, 18 January 2010, at 2:53 a.m.
  • Janowski's formulae
    Rick Janowski -- Tuesday, 19 January 2010, at 7:39 p.m.