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

BGonline.org Forums

42P-31P-42, updated rollout

Posted By: Nack Ballard
Date: Saturday, 29 May 2010, at 12:41 a.m.

In Response To: 42P-31P-42, updated rollout (mtuhtan)

Thanks for the extension, Miran. The variance from a rollout of less than a few thousand trials can throw off one's intuition when trying to analyze a range of positions.

At first, I was puzzled as to why XG rolled out 42P-42P-42D as poorly as -.027, in part because Snowie's 5k has it at only -.011, but also because 42P-31P-42D was, prior to extension, -.035 (i.e., not enough worse than -.027 to fit a decent curve).

But then I saw that you extended 42P-31P-42 (from a mere 2k) and D's error size increased to -.044, restoring some decency: basically, D's comparison with the best splitting play now progresses from -.031 to -.015 to +.027 to +.044, for the positions in which Opp owns the 5pt, 4pt, 3pt and 2pt, respectively. This curve is illustrated in the sequence of diagrams below.


2O ' ' '2X4X '2X ' ' '5O

2X ' '2O '4O '2O ' ' '5X

D is worse than Z by .031



2O ' '2X '4X '2X ' ' '5O

2X ' '2O '4O '2O ' ' '5X

D is worse than S by .015



2O '2X ' '4X '2X ' ' '5O

2X ' '2O '4O '2O ' ' '5X

D is better than S by .027



2O2X ' ' '4X '2X ' ' '5O

2X ' '2O '4O '2O ' ' '5X

D is better than S by .044


Nactation key: D = Down (13/11 13/9), S = Split (24/20 13/11), Z = reverse split (24/22 13/9).

Nack

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.