[Top][All Lists]

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

Re: [Bug-gnubg] Bug(?!): Rollouts involving noise

From: Massimiliano Maini
Subject: Re: [Bug-gnubg] Bug(?!): Rollouts involving noise
Date: Tue, 12 Aug 2008 11:43:16 +0200

Hi Achim,

I guess you lostpasr of Bob's message ... what are  1D, 2C, 2D ?
Can you repost Bob's original message ?


address@hidden wrote on 11/08/2008 13:13:38:

> Hi folks,
> I'm forwarding an e-mail by Bob Koca (he also tried to show me this  
> behavior of gnubg whwn we met in Monte Calro):
> ***************************************************
>   Here is a summary of some of the weird behavior I?ve noticed when  
> doing rollouts involving noise. I suspect it may just be an issue of  
> something being done in a wrong order. I used a 1 point match, zero  
> skill position so the weakening of either player and the clicking of  
> cubeful on or off should not matter but it does. It is the latest  
> build which I just downloaded today.
> In particular:
>   Comparing 1A to 1C shows that if cubeful is on then weakening one  
> player can influence the result.
>   Comparing 2C to 2D shows that if cubeful is off then determinsitc or  
> nondeterministic noise makes a difference. However this is not the  
> case if cubeful is on as can be seen from comparing 1C to 1D.
>   On a related issue it would be nice for the rollout results to give  
> more data. For example
>   Whether the databases were being used, what type of noise was used,  
> what MEQ was used. Basically every setting that is alterable seems  
> like it should be included in the summary.
> Thanks for any help, Bob Koca
> Player 1 = bobk Orange checkers on bottom
> Player 0 = gnubg White checkers on top
> Rollout of 15 on ace vs. 15 on ace 1 point match with various settings:
> All use no variance reduction,  1296 trials, quasi random dice,  no  
> bearoff database, player 0 on roll,  and seed = 1,000,000
> 1)  Using CUBEFUL ON
> A)  expert  with cubeful eval checked vs  expert with cubeful eval checked
>      win% of  .658951   and s.d of .013173
> B)  expert with cubeful eval off vs expert with cubeful eval off
> same as A
> C)  advanced (.015 det noise and cubeful eval)  vs expert
>      win % of .592607 and s.d of .011854
> ***************************************************
> Ciao
> Achim
> _______________________________________________
> Bug-gnubg mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/bug-gnubg

reply via email to

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