"Ingo Althöfer" | 25 Nov 21:45 2014
Picon
Picon

K-best mode in MCTS?

Johannes Laire wrote:
> Sometimes the top move gets 99% or even 99.9% of the playouts, even
> though it's not clearly the best move...
> I've found myself wishing for something similar to Ingo's proposal,
> but I would do it very simply on the UI level. If the Move List
> allowed you to select one row to mean "ignore moves above this", 
> that would be enough. By default the top row would be selected 
> and CrazyStone would do whatever it normally does. If the first 
> move is getting too much attention, just click the second row 
> and then CrazyStone will consider all moves except the first one.

Yes, such a simple procedure might help if the analysis mode
itself works with very few mouse clicks...

> Changing between analysis mode and "resume game" is also 
> very cumbersome if you're analyzing an ongoing game.

You speak out of my soul.

> Only 2) should be enough.

+1

> There's a long way to go to match chess engines, but CrazyStone 
> is quite nifty already. :)

Yes, for analysis purposes CrazyStone is a mile-stone
in the world of go bots.

Cheers, ingo.
(Continue reading)

"Ingo Althöfer" | 25 Nov 21:40 2014
Picon
Picon

Congratulations to AyaMC

Hi Hiroshi,
congratulations to Aya's torunament win also from my side!

I wish that you get further improvement of your engine
by parameter tuning and other measures.
Try to catch up with Zen and CrazyStone!

Ingo.
Alexander Kozlovsky | 23 Nov 21:44 2014
Picon

Intel Xeon Phi card big discount

For anyone interested, currently you can buy Xeon Phi card with huge discount (for less then $200):

http://www.phoronix.com/scan.php?page=news_item&px=MTgzNjY

Comments:
https://news.ycombinator.com/item?id=8648220
http://www.phoronix.com/forums/showthread.php?109498-Intel-s-Xeon-Phi-Is-Being-Sold-For-An-Insanely-Low-Price-Right-Now

Initially I was very excited and thought I can buy one, install it to my desktop and use it for my go engine. But as it turns out this is a server card with passive cooling which is not intended for desktop computer. But if anyone uses server rack case for its go computer, it may be a nice purchase.

Regards,
Alexander Kozlovsky
_______________________________________________
Computer-go mailing list
Computer-go <at> dvandva.org
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
Nick Wedd | 22 Nov 12:30 2014
Picon

Congratulations to AyaMC!

Delayed congratulations to AyaMC, winner of last Sunday's KGS bot
tournament, with 12 wins from 12 games!

My report, which says nothing about the games themselves, is at
http://www.weddslist.com/kgs/past/107/index.html

My thanks to Hideki Kato, for pointing out an error in the table
of annual points at http://www.weddslist.com/kgs/annual/index.html ;
and to Petr Baudiš, for reasons explained in the report.

Nick
--

-- 
Nick Wedd
nick <at> maproom.co.uk
Stefan Kaitschick | 18 Nov 17:23 2014
Picon

opponent modelling with simulation balancing?

Is Rémi Couloms' simulation balancing widely used?
It could also be used asymmetrically in handicap games.
Instead of making strong attacking moves less likely for both sides, black would be deprived of strong moves, until the playouts come back at 50%.
If the position improves for white, black would then be modeled as a stronger player.

just a thought,
Stefan
_______________________________________________
Computer-go mailing list
Computer-go <at> dvandva.org
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
Dave Dyer | 16 Nov 21:05 2014
Picon

Re: Delivery Problem of the list

I have to scramble several times a year to keep my server's outgoing mail
working.

The spam wars aren't going to end soon. New anti-spam methods
and misconfigured anti-spam methods pop up all the time. 
"Ingo Althöfer" | 16 Nov 20:38 2014
Picon
Picon

Delivery Problem of the list

Hello,

in my eyes the best solution for the "non-delivery problem"
would be to have the list on a server with an IP address
that is trustworthy for "all" mailing services.

Ingo.
Detlef Schmicker | 16 Nov 10:04 2014
Picon

new kgsGtp client does not work for me, tournament soon....

Hi, I use this chanel:

kgs has a new client kgsGtp-3.5.20, but I get AM FEINSTEN: Got
successful response to command "time_settings 900 30 10": = 
java.lang.NullPointerException
	at com.gokgs.client.gtp.y.a(kgsgtp:22)
	at com.gokgs.client.gtp.z.a(kgsgtp:58)
	at org.igoweb.igoweb.client.gtp.aT.a(kgsgtp:107)
	at org.igoweb.igoweb.client.gtp.aM.b(kgsgtp:85)
	at org.igoweb.igoweb.client.gtp.aa.a(kgsgtp:246)
	at org.igoweb.igoweb.client.gtp.c.a(kgsgtp:395)
	at org.igoweb.igoweb.client.gtp.c.a(kgsgtp:378)
	at org.igoweb.igoweb.client.gtp.c.d(kgsgtp:368)
	at org.igoweb.igoweb.client.gtp.ad.a(kgsgtp:164)
	at org.igoweb.igoweb.client.gtp.c.a(kgsgtp:158)
	at org.igoweb.igoweb.client.gtp.al.a(kgsgtp:48)
	at org.igoweb.igoweb.client.gtp.am.a(kgsgtp:431)
	at org.igoweb.igoweb.client.gtp.aq.run(kgsgtp:298)
	at org.igoweb.igoweb.client.gtp.at.a(kgsgtp:18)
	at org.igoweb.igoweb.client.gtp.au.run(kgsgtp:97)
	at java.lang.Thread.run(Thread.java:745)
Nov 16, 2014 10:02:23 AM com.gokgs.client.gtp.GtpClient e
SCHWERWIEGEND: Fatal internal error
java.lang.NullPointerException
	at com.gokgs.client.gtp.y.a(kgsgtp:22)
	at com.gokgs.client.gtp.z.a(kgsgtp:58)
	at org.igoweb.igoweb.client.gtp.aT.a(kgsgtp:107)
	at org.igoweb.igoweb.client.gtp.aM.b(kgsgtp:85)
	at org.igoweb.igoweb.client.gtp.aa.a(kgsgtp:246)
	at org.igoweb.igoweb.client.gtp.c.a(kgsgtp:395)
	at org.igoweb.igoweb.client.gtp.c.a(kgsgtp:378)
	at org.igoweb.igoweb.client.gtp.c.d(kgsgtp:368)
	at org.igoweb.igoweb.client.gtp.ad.a(kgsgtp:164)
	at org.igoweb.igoweb.client.gtp.c.a(kgsgtp:158)
	at org.igoweb.igoweb.client.gtp.al.a(kgsgtp:48)
	at org.igoweb.igoweb.client.gtp.am.a(kgsgtp:431)
	at org.igoweb.igoweb.client.gtp.aq.run(kgsgtp:298)
	at org.igoweb.igoweb.client.gtp.at.a(kgsgtp:18)
	at org.igoweb.igoweb.client.gtp.au.run(kgsgtp:97)
	at java.lang.Thread.run(Thread.java:745)

Nov 16, 2014 10:02:23 AM com.gokgs.client.gtp.GtpClient e
FEIN: Normal disconnection from server.

With the kgsGtp-3.5.11 client it seems to work, but I get Unknown
message type -105 for channel org.igoweb.igoweb.client.gtp.b <at> 443f6b14

How do we get the tournament working in about 2 hours:)

Detlef
"Ingo Althöfer" | 15 Nov 09:06 2014
Picon
Picon

K-best mode in MCTS ?

Hello all,

CrazyStone is a fine (commercial) go program.
It has a nice analysis mode, where for a position
%-values for the top 20 or so moves are shown,
including the numbers of playouts for these moves.

When analysing the games from the codecentric Challenge
very often the top move got more than 90 % of the
playouts. Such a distribution may make sense for
a playing program. However, for ANALYSIS mode it would
be better when SEVERAL candidate moves got many playouts.

QUESTION: Who has ideas how such a k-best mode for
MCTS might best be organized?  (Here k is the number
of candidates who get large numbers of playouts - and thus
"precise" evaluations.)

Ingo.

PS. In computer chess such k-best modes are THE tool
for analysis (in GM practice, in correspondence chess ...)
Nick Wedd | 28 Oct 10:30 2014
Picon

November KGS bot tournament: 19x19

The November KGS bot tournament will be held on Sunday November 16th,
starting at 11:00 UTC and ending at 19:00 UTC.  It will use 19x19
boards,  with time limits of 14 minutes each plus very fast Canadian
overtime, and komi of 7.5.  There are details at
http://www.gokgs.com/tournInfo.jsp?id=930 .

I will be on holiday at the time. This is the reason for the
non-standard start time. It also means that from a few days before the
event, I will not have access to my usual email address from which I am
posting this, so you will need to use maproom <at> gmail.com to register.

Please register by emailing me, with the words
"KGS Tournament Registration" in the email title, at
maproom <at> gmail.com .

Nick
--

-- 
Nick Wedd
nick <at> maproom.co.uk
"Ingo Althöfer" | 27 Oct 12:31 2014
Picon
Picon

codecentric Challenge

Hi Rémi,

> Thanks Ingo. I enjoyed this match very much.

The same did I.
It was so nice that both you and FJ were so easy-going
during the arrangements.

By the way, it was a lucky coincidence that the event 
became real in 2014. In German computer go forum there
was a discussion on the Yoda exhibition games from March 
2014, where CrazyStone hadhad a nice H4-win.

FJ also participated, and out of the blue he wrote
"I would be willing to play some match against CS
when someone is willing to sponsor that". For a moment
we discussed crowd funding via the forum, but than I
had an idea which turned out to be better. Raymond-Georg
Snatzke was my assistant at Jena University. He got a
doctor degree for a thesis on Combinatorial Game Theory
applied to the game of Amazons. For several years he is 
now with codecentric. 
I told Georg about FJ's idea to have a match with sponsor.
Georg asked his boss, and to his surprise the answer
was directly positive. 

For me the match is one (little) step on the long march:
to have a bot beating a top human player without handicap
by 2027 (the year of my retirement from University job).

Cheers, Ingo.

PS: It is not an accident that I always write "codecentric"
in small letters, but Challenge with a capital "C".
The official name of the company is with small "c".
_______________________________________________
Computer-go mailing list
Computer-go <at> dvandva.org
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go

Gmane