bug-gnubg
[Top][All Lists]
Advanced

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

Re: [Bug-gnubg] GeneralCubeDecisionE, cubeinfo.fMove, and aarOutput[0][6


From: Joern Thyssen
Subject: Re: [Bug-gnubg] GeneralCubeDecisionE, cubeinfo.fMove, and aarOutput[0][6]
Date: Tue, 6 May 2003 20:50:47 +0000
User-agent: Mutt/1.4.1i

On Tue, May 06, 2003 at 12:05:09PM -0700, David Montgomery wrote
> Hi,
> 
> I'm getting results I don't understand from a call to GeneralCubeDecisionE.
> I'm hoping someone can help me understand them.
> 
> Here's the position:
> 
>  GNU Backgammon  Position ID: eN8PAAADAAAAAA
>                  Match ID   : AQHgAEAAGAAA
>  +12-11-10--9--8--7-------6--5--4--3--2--1-+     O: Oh (Cube: 2)
>  |                  |   |                O | OOO 4 points
>  |                  |   |                O | OOO On roll
>  |                  |   |                  | OOO
>  |                  |   |                  | OO
>  |                  |   |                  | OO
> ^|                  |BAR|                  |     7 point match
>  |                  |   | 6  X             |
>  |                  |   | X  X  X          |
>  |                  |   | X  X  X          |
>  |                  |   | X  X  X          |
>  |                  |   | X  X  X          |     3 points
>  +13-14-15-16-17-18------19-20-21-22-23-24-+     X: Ex
> Pip counts: O 2, X 77
> 
> 
> My understanding of GeneralCubeDecisionE is that:
> 
> output[0][6] gets the cubeful equity assuming that the roller doesn't
> double, and
> output[1][6] gets the cubeful equity assuming that the roller doubles and
> the non-roller takes.
> 
> But for this position I get
> output[0][6] = 0.1598 if cubeinfo.fMove = 0, or
> output[0][6] = 0.3792 if cubeinfo.fMove = 1.
> 
> The board I'm passing to GeneralCubeDecisionE has O (oh) is player 0 (zero).
> So,  if fMove == 0, I would expect output[0][6] = 0.

No, see my first reply and my answer below.

> And, if fMove == 1, I would expect output[0][6] = 0.945*0.169188 = 0.1598
> (Zadeh).
> 
> From this, I can see that fMove==0 is giving me the result I
> expect for fMove==1 (though I don't know why), but I don't have
> any idea where .3792 is coming from.

fMove=0: evaluate the position with X on roll, with score -3,-4 and X
owning 2-cube.

94.5% * (-3,-2) + 5.5% * (-3,0) = 38% MWC

fMove=1: evaluate the position with X ron roll, which score -4,-3 and O
owning 2-cube.

94.5% * (-4,-1) + 5.5% * (-4,0) = 16% MWC

Joern




reply via email to

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