At the moment, Pebbles is creating a huge drift. Brian - CGOS requires
us to use new names on the server each time we change our bots. It
computes the strength using all games (heavilly biased with the
results of the first 100 games)
Pebbles is probably closer to 2000 ELO than 1000 ELO
From: Jason House jason.james.ho...@gmail.com
CGOS requires us to use new names on the server each time we change our bots.
It computes the strength using all games (heavilly biased with the results of
the first 100 games)
Hypothetically speaking, if a
it would slowly grow in (measured) strength over time.
s.
2009/4/20 terry mcintyre terrymcint...@yahoo.com:
From: Jason House jason.james.ho...@gmail.com
CGOS requires us to use new names on the server each time we change our
bots. It computes the strength
The best way to get an accurate picture is to ignore the main page and go to
the Latest Bayeselo which does a full performance rating on all games
every played on CGOS. There will be no bias and you have error bars.
The main page just tracks ratings incrementally and is not a very good
The Bayeselo page would not deal well with it, since it assumes each player
is a unique individual with a fixed playing strength. In this case, the
incrementally rated main page would do a better job.
What I'm thinking about doing is to measure the ratings of active players
compared to the
On Mon, 20 Apr 2009, ?ukasz Lew wrote:
Is there a rating drift? I remember that pure UCT no RAVE with 100k
playouts got over 1700 elo.
That seems a little high. My 50k-pure-UCT searcher is around
1580 for a long time.
Christoph
___
computer-go
Rated: 1713 as of 2007-12-29 09:28:46
http://cgos.boardspace.net/9x9/cross/libEGO-v0.115-100k.html
I am running exactly the same binary file to check it recent rating.
ego-v0.115-100k
Lukasz
On Mon, Apr 20, 2009 at 20:16, Christoph Birk b...@ociw.edu wrote:
On Mon, 20 Apr 2009, ?ukasz Lew
That should be interesting.
So we have
1. ego-v0.115-100k
2. libEGO-v0.115-100k
Is that correct? We can watch it's bayelo rating as well as it's
incremental rating.
- Don
On Mon, Apr 20, 2009 at 3:13 PM, Łukasz Lew lukasz@gmail.com wrote:
Rated: 1713 as of 2007-12-29 09:28:46
Earlier today, I looked up my identical 50k RAVE bots and found
ratings of 1827 (old) and 1468 (new).
Sent from my iPhone
On Apr 20, 2009, at 4:08 PM, Don Dailey dailey@gmail.com wrote:
That should be interesting.
So we have
1. ego-v0.115-100k
2. libEGO-v0.115-100k
Is that
Jason,
This means nothing - can you give us more details? What did the error bars
look like? Which hardware were each run on? etc.
- Don
2009/4/20 Jason House jason.james.ho...@gmail.com
Earlier today, I looked up my identical 50k RAVE bots and found ratings of
1827 (old) and 1468
that is correct.
2009/4/20 Don Dailey dailey@gmail.com:
That should be interesting.
So we have
1. ego-v0.115-100k
2. libEGO-v0.115-100k
Is that correct? We can watch it's bayelo rating as well as it's
incremental rating.
- Don
On Mon, Apr 20, 2009 at 3:13 PM, Łukasz
hb-797-RAVE-50k
1827 -181 +142
hb797-50k
1477 -88 +84
Identical hardware. AMD64x2 with one search thread and no pondering.
The error bars don't overlap, but they come close. I'm rarely patient
enough to wait for 500 games to get +/-50 ELO.
Sent from my iPhone
On Apr 20, 2009, at 4:43 PM,
Hardware might be not important for fixed number of playouts.
Can you give us logins?
2009/4/20 Don Dailey dailey@gmail.com:
Jason,
This means nothing - can you give us more details? What did the error bars
look like? Which hardware were each run on? etc.
- Don
2009/4/20 Jason
On Mon, 20 Apr 2009, ?ukasz Lew wrote:
Is there a rating drift? I remember that pure UCT no RAVE with 100k
playouts got over 1700 elo.
There is no 'anchor' (FatMan-1 ?) runnig on CGOS-9x9 for at
least 36 hours. That could create a drift.
Christoph
Just started.
Hideki
Christoph Birk: pine.lnx.4.64.0904201451410.16...@andromeda.ociw.edu:
On Mon, 20 Apr 2009, ?ukasz Lew wrote:
Is there a rating drift? I remember that pure UCT no RAVE with 100k
playouts got over 1700 elo.
There is no 'anchor' (FatMan-1 ?) runnig on CGOS-9x9 for at
least
I've started two bots: hb797-10k and hb797-50k
They are pure UCT+RAVE with light playouts, one search thread, and no
pondering.
The number at the end represents the playouts per move.
I forget the approximate ratings for each version. I'll guess
1200-1300 for 10k and 1500-1600 for 50k.
There are actually very few programs playing on CGOS (9x9). My engine is
rated around 1000, which means that it plays 75% of its games against
AverageLib (rating 670), and occasional games against Aya (rating 2300). All
games have the predictable result, so I haven't been learning much lately.
I
On Fri, 17 Apr 2009, Brian Sheppard wrote:
I saw on Sensei's Library page http://senseis.xmp.net/?CGOSBasicUCTBots that
there are a range of basic UCT implementations that would be excellent
opponents (rating 1171 through 1603), but I haven't seen these players in
weeks. Is it possible to get
I put something out there. Not sure how well it will play, but it
probably has a rating as it's played before.
- Don
2009/4/17 Brian Sheppard sheppar...@aol.com
There are actually very few programs playing on CGOS (9x9). My engine is
rated around 1000, which means that it plays 75% of
19 matches
Mail list logo
