bug-gnubg
[Top][All Lists]
Advanced

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

Re: [Bug-gnubg] Missed doubles problem ?


From: Holger
Subject: Re: [Bug-gnubg] Missed doubles problem ?
Date: Tue, 26 Aug 2003 23:01:57 +0200

On Mon, 25 Aug 2003 08:47:23 +0200,
address@hidden wrote:
> I've analyzed a game (part of a 7pts match) and in the game stats I have :
>                                       gnubg                MaX
> Missed doubles around DP              0                    1 (-0,078, -0,910%)
> Missed doubles around TG              0                    4 (-0,041, -0,476%)
> 
> Now the only marked cube decisions I can find (using the game record and
> annotation windows, desperatly needing the integrated GUI !!) are the
> following two (both in game 7, moves 8 and 9) :

The statistics counts every missed or wrong cube action. Whether any
particular action gave away enough equity to mark it depends on your
threshold.

> Cube analysis
> 2-ply cubeless equity  +0,908 (Money:  +0,840)
>    85,2%  14,9%   0,1% -  14,8%   1,3%   0,0%
> Cubeful equities:
> 1. Double, pass         +1,000
> 2. Double, take         +1,180  ( +0,180)
> 3. No double            +0,959  ( -0,041)
> Proper cube action: Double, pass
> 
> Cube analysis
> 2-ply cubeless equity  +0,855 (Money:  +0,796)
>    83,9%  12,9%   0,2% -  16,1%   1,3%   0,0%
> Cubeful equities:
> 1. Double, pass         +1,000
> 2. Double, take         +1,070  ( +0,070)
> 3. No double            +0,922  ( -0,078)
> Proper cube action: Double, pass
> 
> What's going on here ?!
> Aren't these 2 missed doubles around DP ?!

An equity of short to +1 looks to me almost too good. I don't know
your match score but think your doubling window opens a bit lower.

> Also note how the equity differences in the stats (-0.078 and -0.041) are
> exactly the equity differences of the two moves !

At least with this I can neither explain the counts you got above. Or
maybe a rounding problem.

> Problem in the problem, saving and reloading the game (.sgf), the stats read :
> 
>                                       gnubg                MaX
> Missed doubles around DP              0                    1 (-0,078, -0,911%)
> Missed doubles around TG              0                    1 (-0,041, -0,476%)
> 
> Now the count looks ok (!!++), but I still cannot understand why one of the
> two is considered as a 'Missed double around TG'.

See above.

> The count problem can be reproduced : load the game (saved without any
> analisys), then analyze it and get the strange count (1 missed around DP,
> 4 missed around TG). Now save the game, reload it and the count now says
> 1 missed around DP, 1 missed around TG.
> Somebody should be able to look into that, email me if you need the game file.

You can attach _small_ files right away.

> In fact, I also have a saved game (part of another 7pts match) where the game
> stats are :
>                                       gnubg                MaX
> Missed doubles around DP              0                    3 (-0,178, -1,137%)
> 
> while in Game record/Annotation window I can only find this marked cube
> decision :

> This time, the equity difference is not the same and, also, the count doesn't
> change saving and reloading the game.

See again above the threshold.

Regards,

        Holger




reply via email to

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