[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

GNUBG RO Blunder(BGB/XBG right)

Posted By: Frank Berger
Date: Monday, 29 June 2009, at 9:29 p.m.

In Response To: GNUBG RO Blunder(BGB/XBG right) (neilkaz)


The rollout posted at hardy's forum was just 1-ply without variance reduction (it had to be fast because I couldn't keep my eyes open).

I just did a 1296/2-ply with VR which shows pretty much the same results. A 384 3ply-with VR shows this: 

 

Rollout 

========= 

Number of Games:384 

 

Move : 10-4, 8-4 

       Wins    G+BG    BG  

Green   6.1     0.0     0.0  

Red    93.9    87.3    81.1  

Match Winning Probability Red: 0.969 

95% Confidence Interval:       2.511 /  2.613 

 

Move : 8-2, 10-6 

       Wins    G+BG    BG    

Green   6.7     0.0     0.0  

Red    93.3    86.6    81.2   

Match Winning Probability Red: 0.966 

95% Confidence Interval:       2.492 /  2.596 

 

Move : 8-2, 7-3 

       Wins    G+BG    BG   

Green   6.8     0.0    -0.0  

Red    93.2    86.1    77.9   

Match Winning Probability Red: 0.966 

95% Confidence Interval:       2.455 /  2.555


Because BGB shows unusual changes of the evaluation with different ply, the real value might even be a little higher and your evaluation looks excellent.

Altough the evaluation of BGB is untypically off due to the unusual position, the relative ordering is still o.k. and BGB suggests reasonable moves

ciao

Frank






Messages In This Thread

 

Post Response
Your Name:
Your E-Mail Address:
Subject:
Message:

If necessary, enter your password below:

Password:
 

 

[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums is maintained by Stick with WebBBS 5.12.