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

BGonline.org Forums

Bot Comparison 140 / 500 games.

Posted By: neilkaz
Date: Friday, 27 November 2009, at 8:33 p.m.

In Response To: Bot Comparison 140 / 500 games. (Marv Porten)

Marv "The GNUBG 3-ply results for missed double, wrong double, wrong take, and wrong pass appear to be incorrect (or anomalous) compared to GNUBG 2-ply and 4-ply, and the last three categories (i.e., excluding missed double) appear to be incorrect or anomalous compared to the other bots. Any explanation?"

Neil replies, that GNU evaluations have a clear oscillation effect that makes odd plies, ie the 3 ply we are interested in, value the chances of the side on roll as greater than what 2 and 4 ply value them. This means that since over the wide majority of positions, GNU 3 ply overvaluing the side on roll, it will 1) Rarely fail to cube in a possible doubling opportunity, 2) double too soon rather often, 3) rarely make a bad take, 4) pass too quickly rather often.

That is exactly what is shown in the table for GNU 3 ply .. neilkaz ..

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.