On Feb 2, 2009, at 9:40 AM, Jason House <jason.james.ho...@gmail.com>
wrote:
On Feb 2, 2009, at 6:57 AM, "Isaac Deutsch" <i...@gmx.ch> wrote:
Hi Issac,
You should be more in the range of +200-300 ELO, at least with
pattern
based
playouts.
Sylvain
Isaac. They are not pattern based playouts, but as I said uniformly
random.
I reckon the effect of RAVE is less with these?
"How many playouts per second do you get with each version?"
Actually, to make comparable tests with both versions, the version
"without
RAVE" just sets the coefficient of RAVE in the UCT-RAVE value
calculation to
zero. I get about 41k games/s on 9x9 (using 2 cores, about 20k/s/
core).
The playouts are fairly optimized but the tree search isn't at all
yet, so
there is still some potential.
My first (braindead) multi-threaded UCT played weaker with two cores
than one core. How do you combine search trees/results? How do you
pick a move to play?
Single-threaded RAVE with no parameter tuning, an ancient RAVE
formula, and 10k light playouts per second got me into the 1600's on
CGOS. I used Don Dailey's AMAF methodology for RAVE (first color to
play on a point, keep 7/8 of move list).
Actually, that rating was probably using my own home-grown RAVE
formula. Sorry for the misinformation.
Also, I noticed your rank measurements were based on CGOS results
after relatively few games. It can retain significant bias for quite
a while.
--
Jetzt 1 Monat kostenlos! GMX FreeDSL - Telefonanschluss + DSL
für nur 17,95 Euro/mtl.!* http://dsl.gmx.de/?ac=OM.AD.PD003K11308T4569
a
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
