Re: [computer-go] MoGoRel3_3550pps

[Michael Williams]

Sylvain Gelly wrote:


As other said, but just to confirm:
--playsAgainstHuman 0
Also you have to specify
--totalTime 300

if 300 is the number of seconds of the games. If not, MoGo does not care 
about the time left, and will just play a constant time per move, 
loosing by time with no other worry :).




It looks like MoGo does respect the time_left commands from GTP, so I 
don't think the totalTime parameter is required in this case.

___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] CGOS web page

[Michael Williams]

Don Dailey wrote:

Take a look at the 9x9 web page now.

- Don



I think it looks great.  Thanks, Don.
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] CGOS web page

[Don Dailey]

Take a look at the 9x9 web page now.

- Don



Michael Williams wrote:
> I don't know about everyone else, but I would like to see the programs
> displayed in rank order even if they are not totally stable yet.
> ___
> 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/

Re: [computer-go] CGOS web page

[Don Dailey]

Michael Williams wrote:
> I don't know about everyone else, but I would like to see the programs
> displayed in rank order even if they are not totally stable yet.
We can probably do that if that's what people want.   The initial rating
(which is almost meaningless) should be set pretty low so that brand new
players don't get high on the list.

- Don


> ___
> 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/

[computer-go] CGOS web page

[Michael Williams]

I don't know about everyone else, but I would like to see the programs 
displayed in rank order even if they are not totally stable yet.

___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] How to get more participation in 19x19 CGOS?

[Don Dailey]

You should be able to run for days - I have sometimes had bots run for
weeks.

A broken pipe usually means your program crashed.If the connection
to the server was lost you would get a different error probably.

- Don


Mark Boon wrote:
> I came back to my computer to see an error message about a broken
> pipe. I also see my program lost a game on time to Odie, which is
> probably caused by this. Is this common? Or does it mean I have a
> problem on my end? I do have a rather slow internet connection.
>
> Mark
> ___
> 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/

Re: [computer-go] How to get more participation in 19x19 CGOS?

[Mark Boon]

I came back to my computer to see an error message about a broken  
pipe. I also see my program lost a game on time to Odie, which is  
probably caused by this. Is this common? Or does it mean I have a  
problem on my end? I do have a rather slow internet connection.


Mark
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

RE: [computer-go] How to get more participation in 19x19 CGOS?

[David Fotland]

The new level of participation is terrific.  It's great to see Goliath, the 
strongest program for many years.  I love the new 15 minute time limits.  Thank 
you everyone.

I'd like to remind people to follow the link to Sensei's computer go page and 
add a brief description of your program when you have time.

David

> -Original Message-
> From: [EMAIL PROTECTED] [mailto:computer-go-
> [EMAIL PROTECTED] On Behalf Of Don Dailey
> Sent: Thursday, January 10, 2008 10:03 AM
> To: computer-go
> Subject: Re: [computer-go] How to get more participation in 19x19 CGOS?
> 
> I also put up "Odie" on the 19x19 server so that we would have some
> weaker players.Odie should be quite weak and I can also put up even
> weaker versions.
> 
> - Don
> 
> 
> 
> 
> David Fotland wrote:
> >
> > I tested the three versions only against each other, and tuned them
> by
> > removing go knowledge.  I didn�t spend much time on it so I didn�t
> > have time to tune against gnugo.  I thought that tuning on the server
> > would be better.  Since programs do better against themselves, the
> > actual strength difference is probably less than 5 stones.
> >
> >
> >
> > Level 0 should be really stupid.  It can�t read ladders, and the move
> > generator only uses patterns that are 3x3.
> >
> >
> >
> > I tested with 100 games, and level 2 gives 5 stones to level 1 and
> > wins 45%.  Level 1 gives five stones to level 0 and wins 53%
> >
> >
> >
> > David
> >
> >
> >
> > *From:* [EMAIL PROTECTED]
> > [mailto:[EMAIL PROTECTED] *On Behalf Of *terry
> mcintyre
> > *Sent:* Wednesday, January 09, 2008 7:39 AM
> > *To:* computer-go
> > *Subject:* Re: [computer-go] How to get more participation in 19x19
> CGOS?
> >
> >
> >
> > Thanks, David!
> >
> > I think you said that your three versions of Many Faces 12 are about
> 5
> > stones apart in strength?
> >
> > I'm guessing this means that version A can give B a 5 stone handicap,
> > and win 50% of the time,
> > while version B can give C a 5 stone handicap and win 50% of the
> games?
> >
> > Have you played the 3 versions against Gnugo? What handicaps seem
> > appropriate there?
> >
> > I think Don Dailey posted that there is a need for a GnuGo on the
> > 19x19 server? I can host if need be, have spare CPU here and full-
> time
> > connection.
> >
> >
> >
> > Terry McIntyre <[EMAIL PROTECTED]>
> >
> >
> >
> > �Wherever is found what is called a paternal government, there is
> > found state education. It has been discovered that the best way to
> > insure implicit obedience is to commence tyranny in the nursery.�
> >
> >
> >
> > Benjamin Disraeli, Speech in the House of Commons [June 15, 1874]
> >
> >
> >
> > - Original Message 
> > From: David Fotland <[EMAIL PROTECTED]>
> > To: computer-go 
> > Sent: Tuesday, January 8, 2008 3:00:44 PM
> > Subject: RE: [computer-go] How to get more participation in 19x19
> CGOS?
> >
> > You can test your bot on the 19x19 cgos now, since I added 3 weak
> > players today.  They are so fast they don�t load my machine, so I�ll
> > leave them up.
> >
> >
> >
> > David
> >
> >
> >
> >
> >
> > -
> ---
> >
> > Never miss a thing. Make Yahoo your homepage.
> > 
> >
> > -
> ---
> >
> > ___
> > 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/

___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] How to get more participation in 19x19 CGOS?

[Don Dailey]

I also put up "Odie" on the 19x19 server so that we would have some
weaker players.Odie should be quite weak and I can also put up even
weaker versions.

- Don




David Fotland wrote:
>
> I tested the three versions only against each other, and tuned them by
> removing go knowledge.  I didn’t spend much time on it so I didn’t
> have time to tune against gnugo.  I thought that tuning on the server
> would be better.  Since programs do better against themselves, the
> actual strength difference is probably less than 5 stones.
>
>  
>
> Level 0 should be really stupid.  It can’t read ladders, and the move
> generator only uses patterns that are 3x3. 
>
>  
>
> I tested with 100 games, and level 2 gives 5 stones to level 1 and
> wins 45%.  Level 1 gives five stones to level 0 and wins 53%
>
>  
>
> David
>
>  
>
> *From:* [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED] *On Behalf Of *terry mcintyre
> *Sent:* Wednesday, January 09, 2008 7:39 AM
> *To:* computer-go
> *Subject:* Re: [computer-go] How to get more participation in 19x19 CGOS?
>
>  
>
> Thanks, David!
>
> I think you said that your three versions of Many Faces 12 are about 5
> stones apart in strength?
>
> I'm guessing this means that version A can give B a 5 stone handicap,
> and win 50% of the time,
> while version B can give C a 5 stone handicap and win 50% of the games?
>
> Have you played the 3 versions against Gnugo? What handicaps seem
> appropriate there?
>
> I think Don Dailey posted that there is a need for a GnuGo on the
> 19x19 server? I can host if need be, have spare CPU here and full-time
> connection.
>
>  
>
> Terry McIntyre <[EMAIL PROTECTED]>
>
>  
>
> “Wherever is found what is called a paternal government, there is
> found state education. It has been discovered that the best way to
> insure implicit obedience is to commence tyranny in the nursery.”
>
>  
>
> Benjamin Disraeli, Speech in the House of Commons [June 15, 1874]
>
>  
>
> - Original Message 
> From: David Fotland <[EMAIL PROTECTED]>
> To: computer-go 
> Sent: Tuesday, January 8, 2008 3:00:44 PM
> Subject: RE: [computer-go] How to get more participation in 19x19 CGOS?
>
> You can test your bot on the 19x19 cgos now, since I added 3 weak
> players today.  They are so fast they don’t load my machine, so I’ll
> leave them up.
>
>  
>
> David
>
>  
>
>  
>
> 
>
> Never miss a thing. Make Yahoo your homepage.
> 
>
> 
>
> ___
> 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/

Re: [computer-go] To French speakers: computer Go on the French Radio

[Alain Baeckeroot]

Le jeudi 10 janvier 2008, Olivier Teytaud a écrit :
> > They announce that a match will be organized between MoGo and a 
> > professional 
> > player in March, during the Paris Go Tournament.
> 
> It will be the
> MPI version of mogo, and
> in various board-sizes.
> 
What will be the hardware for Mogo ?

Alain.

___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] How to design the stronger playout policy?

[compgo123]

I think the issue with MC or deterministic is following. With MC your value of 
the node can get more accurate with larger number of playouts. With a 
deterministic you don'thave such a mechanism here (there are other?mechanisms 
in UCT). 

DL


-Original Message-
From: Don Dailey <[EMAIL PROTECTED]>
To: computer-go 
Sent: Wed, 9 Jan 2008 3:46 pm
Subject: Re: [computer-go] How to design the stronger playout policy?





Mark Boon wrote:
>
> On 8-jan-08, at 17:04, Don Dailey wrote:
>
>> And yes, it slows down the play-outs.   Still, the play-outs seem to
>>
>> require a good bit of randomness - certainly they cannot be
>>
>> deterministic and it seems difficult to find the general principles that
>>
>> are important to the play-out policy.  
>>
>
> I was thinking about this and wanted to make sure I understand what
> you mean by 'cannot be deterministic'.
You could build a playing strategy that would always play the same exact
game from a particular position and this would be deterministic but then
the "Monte Carlo" component is completely missing from the
program.   You would still get some chaos from the tree itself since
you always expand a node once visited.   

I don't think this is fully understood so I could be wrong in my
statement - but it does seem likely that a certain amount of randomness
will improve the results. Attempts to make the play-outs stronger
and stronger have failed,  perhaps due to the fact that this makes them
less and less random,  or perhaps it's due to some other phenomenon.  

- Don


> For example a stone gets put into atari and it finds an escaping move.
> With plain playouts the stone gets captured anyway roughly 50% of the
> time. With heavy playouts it gives a weighting so that the chance of
> escaping becomes larger? Or does it always play the escaping move and
> just the rest of the playout is random?
>
> Mark
>
> 
>
> ___
> 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/



More new features than ever.  Check out the new AOL Mail ! - 
http://webmail.aol.com
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] To French speakers: computer Go on the French Radio

[terry mcintyre]

From: Olivier Teytaud <[EMAIL PROTECTED]>


> They announce that a match will be organized between MoGo and a
 professional 
> player in March, during the Paris Go Tournament.

> It will be the MPI version of mogo, and in various board-sizes.


What handicap, if any?
 
Terry McIntyre <[EMAIL PROTECTED]> 
“Wherever
is found what is called a paternal government, there is found state
education. It has been discovered that the best way to insure implicit
obedience is to commence tyranny in the nursery.”
 
Benjamin Disraeli, Speech in the House of Commons [June 15, 1874]





  

Never miss a thing.  Make Yahoo your home page. 
http://www.yahoo.com/r/hs___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] On question about Libego110

[compgo123]

Thanks,Lukasz.



DL

-Original Message-
From: Łukasz Lew <[EMAIL PROTECTED]>
To: computer-go 
Sent: Thu, 10 Jan 2008 7:47 am
Subject: Re: [computer-go] On question about Libego110




The code of any version is easy to get:
ttp://www.mimuw.edu.pl/~lew/hg/libego/?tags
The file you are talking about is here:
ttp://www.mimuw.edu.pl/~lew/hg/libego/?file/dfcd0a6db96e/uct.cpp

f you take a look at line 151 you see: (bias should be renamed to
umber_of_visits)
explore_coeff  = log (bias) * explore_rate;
where bias is equivalent to this->bias i.e. number of visits in current node.
 lines lower is a loop over all children where we compute:
   float child_urgency = child->ucb  (explore_coeff);
where the body of ucb is:
return
   (pl == player::black ? value : -value) +
   sqrt (explore_coeff / bias);
and "bias" variable refers to the child's bias.
If you have any more questions, just ask.
ukasz
On Jan 9, 2008 7:49 PM,  <[EMAIL PROTECTED]> wrote:
 I'm looking at the code of Libgo110. I have a question. In the file uct.cpp
 and the definition of class note_t, the explore_coeff is calculated from
 log(node->bias). But in the paper 'Modification of UCT with Patterns in
 Monte-Carlo Go' table 1 line 10 -17, the explore_coeff is calculated from
 log(nb), where nb is the summation of node->bias for all the child nodes.
 Whyis the difference? Or did I read the code wrong?

  Thanks for any explanation.

  DL
  
  More new features than ever. Check out the new AOL Mail!

 ___
 computer-go mailing list
 computer-go@computer-go.org
 http://www.computer-go.org/mailman/listinfo/computer-go/




___
omputer-go mailing list
[EMAIL PROTECTED]
ttp://www.computer-go.org/mailman/listinfo/computer-go/



More new features than ever.  Check out the new AOL Mail ! - 
http://webmail.aol.com
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

[computer-go] cgos server fix

[Don Dailey]

I deployed a potential fix to the 9x9 cgos server as recommended by
Michael Williams.

If it goes over 3 days without crashing, the confidence that it is fixed
will be high.

Then I will ask Olivier to deploy the fix on the 19x19 server.

- Don

___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] To French speakers: computer Go on the French Radio

[Olivier Teytaud]

They announce that a match will be organized between MoGo and a professional 
player in March, during the Paris Go Tournament.


It will be the
MPI version of mogo, and
in various board-sizes.

___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] How to get more participation in 19x19 CGOS?

[Mark Boon]

I see now what people mean with regards to the starting of rounds.  
Most bots are idle most of the time while a few slow ones slug it out.


The way it's currently configured was probably the simplest way to do  
it and get reasonably uniform results. Otherwise you may end up with  
fast bots playing lots of games against each other and skew their  
ratings.


Still, maybe it could be done so that when there are only one or two  
games going still that are far from finished (say less than 150-200  
moves played) then a new round gets started. The bots in the slow game 
(s) skip a round. This way you would still get reasonable pairings  
and have more bots active all the time.


Just a suggestion of course, I'm sure some thought has been given on  
the issue already.


Mark

___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

[computer-go] To French speakers: computer Go on the French Radio

[Rémi Coulom]

Hi,

Some readers of this list may be interested in this one-hour programme 
that will be broadcasted live on France Culture tomorrow afternoon:

http://www.radiofrance.fr/chaines/france-culture2/emissions/science_publique/fiche.php?diffusion_id=58397&pg=avenir
It will be available for download from that web site.

They announce that a match will be organized between MoGo and a 
professional player in March, during the Paris Go Tournament.


Rémi
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] How to get more participation in 19x19 CGOS?

[Mark Boon]

On 9-jan-08, at 20:43, Don Dailey wrote:


I picked a game at random and it was just dumb luck that I picked the
game of Goliath!That was a pure coincidence.



That was funny. Anyway, I managed to get the file and look at it and  
when I test it it does remove the dead stones. So I'm a bit puzzled.


I've made a change to make it more deterministic and I'll try again.

Mark

___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] On question about Libego110

[Łukasz Lew]

The code of any version is easy to get:
http://www.mimuw.edu.pl/~lew/hg/libego/?tags

The file you are talking about is here:
http://www.mimuw.edu.pl/~lew/hg/libego/?file/dfcd0a6db96e/uct.cpp


If you take a look at line 151 you see: (bias should be renamed to
number_of_visits)

explore_coeff  = log (bias) * explore_rate;

where bias is equivalent to this->bias i.e. number of visits in current node.
2 lines lower is a loop over all children where we compute:

   float child_urgency = child->ucb  (explore_coeff);

where the body of ucb is:

return
(pl == player::black ? value : -value) +
sqrt (explore_coeff / bias);

and "bias" variable refers to the child's bias.

If you have any more questions, just ask.
Łukasz

On Jan 9, 2008 7:49 PM,  <[EMAIL PROTECTED]> wrote:
> I'm looking at the code of Libgo110. I have a question. In the file uct.cpp
> and the definition of class note_t, the explore_coeff is calculated from
> log(node->bias). But in the paper 'Modification of UCT with Patterns in
> Monte-Carlo Go' table 1 line 10 -17, the explore_coeff is calculated from
> log(nb), where nb is the summation of node->bias for all the child nodes.
> Whyis the difference? Or did I read the code wrong?
>
>  Thanks for any explanation.
>
>  DL
>  
>  More new features than ever. Check out the new AOL Mail!
>
> ___
> 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/

Re: [computer-go] MoGoRel3_3550pps

[elife2008]

Hi Sylvain,
  Have you finished your thesis? We are eager to read it:-)

On 1/10/08, Sylvain Gelly <[EMAIL PROTECTED]> wrote:
> Hi,
>
> I guess the public version of MoGo was designed with a
> > focus on 9x9 and not 19x19.
>
> It was not more on 9x9 that 19x19, it was more or less the best settings of
> MoGo against gnugo at the moment I left the developpement (early september)
> for both 9x9 and 19x19.
>
>
> Or is there something else I should be
> > including on the command line?
>
> As other said, but just to confirm:
> --playsAgainstHuman 0
> Also you have to specify
> --totalTime 300
>
> if 300 is the number of seconds of the games. If not, MoGo does not care
> about the time left, and will just play a constant time per move, loosing by
> time with no other worry :).
>
> The release is designed to play against human on a server/client which
> supports a scoring taking into account the dead stones. On cgos you have to
> capture all dead stones.
> As for the rating, I don't know all the changes that has been done on CGOS
> since then, but on the old 19x19 one, the rating should be more than 2100 or
> 2200.
>
> Hoping this helps,
> Sylvain
>
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] MoGoRel3_3550pps

[Sylvain Gelly]

Hi,

I guess the public version of MoGo was designed with a
> focus on 9x9 and not 19x19.

It was not more on 9x9 that 19x19, it was more or less the best settings of
MoGo against gnugo at the moment I left the developpement (early september)
for both 9x9 and 19x19.


Or is there something else I should be
> including on the command line?

As other said, but just to confirm:
--playsAgainstHuman 0
Also you have to specify
--totalTime 300

if 300 is the number of seconds of the games. If not, MoGo does not care
about the time left, and will just play a constant time per move, loosing by
time with no other worry :).

The release is designed to play against human on a server/client which
supports a scoring taking into account the dead stones. On cgos you have to
capture all dead stones.
As for the rating, I don't know all the changes that has been done on CGOS
since then, but on the old 19x19 one, the rating should be more than 2100 or
2200.

Hoping this helps,
Sylvain
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/