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

BGonline.org Forums

21$ 63 oddity

Posted By: Dmitriy Obukhov
Date: Tuesday, 5 August 2014, at 7:05 a.m.

In Response To: 21$ 63 oddity (Bob Koca)

I see that 3 ply thinks the plays are very close at money and for many scores in 4 point match thinks split is correct when slot is correct by a decent amount. For example in 3 point match it thinks split by .016. Does this indicate a weakness in how XG incorporates the raw numbers into match play?

Are we talking about 53P-21 here? Without opening book XG thinks split is best by about 0.02 (depending on ply level) for money. It is typical for XG to hate slotting plays. You can expect a similar effect at 3a-3a, and it does exist. However, this big swing at 3a-3a cannot be just because XG underestimates slotting plays. Bots neural net is trained by playing a bunch of cubeless money games, so it makes sense that it understands money play better than match play.

There is effect of killed gammons on a 4 cube. Comparing with money play, slot gains some equity here, just as opening 64S gains at 3a-3a and 4a-4a.

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.