[Computer-go] Pachi Authors

David Fotland fotland at smart-games.com
Thu Apr 21 17:03:44 PDT 2011


I do almost the same thing in Many Faces.  When taking black in a handicap
game, the dynamic komi starts at 7 points per handicap stone, and decreases
linearly to zero at move 90.

 

David

 

From: computer-go-bounces at dvandva.org
[mailto:computer-go-bounces at dvandva.org] On Behalf Of Jean-loup Gailly
Sent: Monday, April 18, 2011 8:49 AM
To: computer-go at dvandva.org
Subject: Re: [Computer-go] Pachi Authors

 

> One more question on the recent Pachi wins against pro players
> in handicap games: Did you use dynamic komi; and if so, in which way?

 

pachi supports several dynamic komi options, including a rather
sophisticated

"adaptive" method. But in distributed mode I found that the linear method

works best. Black gives itself initially a disadvantage of 7 points per
handicap

stone (so 49 points in the h7 game) and this disadvantage drops linearly to
0

after 200 moves. This way all machines in distributed mode compute

the same dynamic komi.

 

The actual value of one handicap stone is twice the normal komi so about

15 points(*), but I tried this and it makes pachi play too agressively. 7
points

per stone is a compromise which seems to work well.

 

Jean-loup

 

(*) http://www.mail-archive.com/computer-go@dvandva.org/msg00096.html

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://computer-go.org/pipermail/computer-go/attachments/20110421/ec094734/attachment.html>


More information about the Computer-go mailing list