UEC cup article appeared Mainichi evening paper December 13 in Japan.
Some quotes are as follows.
---
Monte-Carlo Go dominates computer Go world.
MC based program are 10 programs out of 27.
Top five are all MC based programs except
Sgf files are available here.
First day.
http://jsb.cs.uec.ac.jp/~igo/sgf/uec1201.zip
Second day. (There is no exhibition sgf)
http://jsb.cs.uec.ac.jp/~igo/sgf/uec1202.zip
Hiroshi Yamashita
- Original Message -
From: Hiroshi Yamashita [EMAIL PROTECTED]
To: computer-go
Hi Don,
I never heard of this technique before. Are there any more you can share?
ManyFaces12 uses:
-Iterative deepening, with hash table
-Zero-window search (beta is alpha+1, and research when fail high)
-Null move (reduce depth by one, only try null when beta is not infinite,
only one null
David,
There are many variations of this technique, but what they all have in
common is the assumption that once you have searched some number of
child moves (perhaps 2 or 3) then (with good move ordering) there is
much reduced chance than one of the remaining moves will be
useful. Often
DF: thanks for the link to this new (to me technique). I'll implement it
soon.
In my alpha/beta searcher a very simple minded version is actually
working very well - it will take weeks though to test and refine
this. The only thing I am doing now is:
At any given node ...
If
David,
Just one comment on the ETC. This is not internal iterative
deepening. Internal iterative deepening is a technique to find a high
quality move that hopefully will give you a cutoff (when it gets
searched.) It's a good technique to use but it isn't ETC.
However, ETC will sometimes
for example, go books make a big deal about where to extend along the
side, or when to play in one corner or another, but the difference between
these various moves is usually only a few points.
The difference between similar-appearing various moves may well be one
of efficiency--and that
