RE: [Bug-gnubg] Why does gnubg allow itself to be gammoned?

[Ian Shaw]

> From:  Billie Patterson
> Sent: 12 February 2003 23:14
> To: address@hidden
> Subject: [Bug-gnubg] Why does gnubg allow itself to be gammoned?
> 
> 
> Here's the setup:
> 
> I'm going to win in the next roll no matter what gnubg
> does.  But on a 34 roll, gnubg plays 6/3, 5/1 rather than
> 8/5, 4/off or 8/4, 3/off.  I don't understand his logic
> here.  How does he benefit by not taking a man off?
> 
>  GNU Backgammon  Position ID: BQAAeLcrAQAAAA
>                  Match ID   : EoHxACAAKAAA
>  +12-11-10--9--8--7-------6--5--4--3--2--1-+     O: gnubg
>  |             O    |   | O  O  O  O  O  O |     2 points
>  |                  |   |       O  O  O  O |     Rolled 34
>  |                  |   |       O     O  O |     
>  |                  |   |                O |     
>  |                  |   |                  |    
> ^|                  |BAR|                  |     7 point
> match
>  |                  |   |                  | XX 
>  |                  |   |                  | XX  
>  |                  |   |                  | XXX 
>  |                  |   |                  | XXX 
>  |                  |   |             X  X | XXX 5 points
>  +13-14-15-16-17-18------19-20-21-22-23-24-+     X: Billie
> (Cube: 4)
> Pip counts: O 47, X 3
> 
> * gnubg moves 6/3 5/1
> 
Any move GnuBg makes gives it a match winning chance of 0%, so it just
picks one from the list (probably based on the first one found by the
legal-move-finding routines.)

Gammons don't count in this position, i.e. with the cube on 4. You win
the match irrespective of whether the score is 7-2 or 9-2. The extra two
points don't count for anything (technically known as overage). You
probably noticed that GnuBg redoubled to 4 as soon as you doubled to 2.
This is correct behaviour because it had nothing to lose but a lot to
gain by playing for 4 points instead of just 2. 


--Ian