Re: [Bug-gnubg] new cube rating computation

[W.Stroop]

----- Original Message -----
From: "max-d" <address@hidden>
To: <address@hidden>
Sent: Wednesday, August 07, 2002 10:28 AM
Subject: [Bug-gnubg] new cube rating computation


> hi,
> The new cube rating computation does not look fair ...
> i can't agree i'm a beginner cube user  just because of this wrong take !
> 3 point match
> Move number 11: O doubles to 2
>  GNU Backgammon  Position ID: bJ2DAyDMngcHAA
>                  Match ID   : MBlgAAAAAAAA
> * user accepts
>
> Alert: wrong take ( -0,53%)! [doubtful]
> Cube analysis
> 2-ply cubeless MWC  55,67%
>   0,729 0,108 0,001 - 0,271 0,036 0,001
> 1. Double, pass         60,00%
> 2. Double, take         60,53%  ( +0,53%)
> 3. No double            58,88%  ( -1,12%)
> Proper cube action: Double, pass

Because its a 3 pointer I think it's correct to drop. In most cases 'X' has
to win the next
2 games if he takes, while if he drops this is not neccesarely the case.

In a 5 point match GNUbg takes.

        Win    W(g)   W(bg)  L(g)   L(bg)  Equity  (cubeful)
static: 0,724 0,115 0,001 0,036 0,000 (+0,527  (+0,877))
 1 ply: 0,740 0,120 0,001 0,035 0,000 (+0,564  (+0,969))
 2 ply: 0,728 0,109 0,001 0,037 0,001 (+0,529  (+0,881))


Double, take        : +0,881
Double, pass        : +1,000   (+0,119)
No double           : +0,821   (-0,060)

Correct cube action: Double, take

Greetings Rob