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

BGonline.org Forums

Rollout w/3 ply cube

Posted By: Chuck Bower
Date: Monday, 23 February 2009, at 6:16 p.m.

In Response To: Rollout w/3 ply cube (Christian Munk-Christensen)

The following is my understanding of how GNU-bg works, and that might not be accurate. But FWIW:

The wins shown in the GNU-bg report aren't directly related to takes or passes. These are the cubeless expectation (evaluation) at the time the game ended (often with a pass).

You are right to be suspicious when the cubeless values disagree in the two branches (double and no double). Even though these shouldn't be the same for matchplay (I contend they should be the same for $-cubes, within statistical uncertainty), when they diverge considerably it should raise a flag that the bot doesn't understand the position very well.

In the Fogerlund-Rockwell position I posted a week ago ("hidden lesson" thread) the large disagreement in the 2-ply rollout is what led me to dig deeper and find that GNU-bg was taking subsequent cubes in the no-double branch which it should have passed. (If your opp is going to take the cube next time, rightly or wrongly, you can't lose your market!)

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.