[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-gnubg] probably bug

From: Jim Segrave
Subject: Re: [Bug-gnubg] probably bug
Date: Wed, 5 Dec 2007 05:45:00 +0100
User-agent: Mutt/1.5.16 (2007-06-09)

On Wed 05 Dec 2007 (01:58 +0100), Wojciech Zaremba wrote:
> So gnubg should to move from 4 to off and from 2 to off (or looking on the
> board from 21,23 to the board)

Sorry, I misread your initial post as being a report that gnubg simply
got stuck playing that position, not that gnubg's move was absurd. 

It's a legal move, using the 2 first - 4/2/off, it just seems to make
little sense. However, the other players chances of winning are so
tiny that it almost makes no difference what move gnubg chooses - the
equity difference between taking two pieces off on this roll and one
piece is absolutely tiny. Basicly, the opponent needs to roll four
doubles in the next 5 moves while gnubg rolls no doubles, The chances
of that happening I make to be:

0.482 - that gnubg rolls no doubles in its next four rolls
0.000776 - that the opponent rolls at least four doubles in the next
five rolls. 

the net result is that the opponent wins 0.031 percent of the time and
the difference in equity between taking one or two off now is
negligible. I haven't checked, it may well be that it makes no
difference at all in reducing the worst case number of rolls for gnubg.

Jim Segrave           address@hidden

reply via email to

[Prev in Thread] Current Thread [Next in Thread]