Ongoing GO programming?

[jshriver]

MoGo and CrazyStone represent the new wave of Monte Carlo-based Go programs.

Many more brilliant developers than I have decided on the MC route. But me personally I don't feel it's the path to a really strong engine. As for my own efforts I'm going a different route, and working directly with structures.

In fact 100% of the code I've written so far has been in generation and analysis of structures. Results seem promising. We'll see though. I just hope to get an engine out of vaporware status and onto a IGS unlike my chess engine attempts.

I really wish there was a forum for Go dev kinda like CCC. Though it seems all attempts at creating one go blank in favor for the popular computer-go mailing list.

-Josh

[Denis P. Mendoza]

I have made my best estimate of the top ten 19x19 computer Go programs on Sensei's Library, a wiki dedicated to the game of Go. This estimate is hindered by the lack of participation of top programs in recent computer Go competitions. (There is currently no equivalent to the WCCC in the Go world. The well-funded World Computer Go Championships died off with the Ing Prize in 2000. Participation in the Computer Olympiad by top programs has been spotty.) My results are summarized here; source material is at the first link above.

• 1. MoGo
  2. CrazyStone
  3. KCC Igo (Silver Star)
  4. Haruka
  5. Go4++
  6. HandTalk (Goemate)
  7. GNU Go
  8. Many Faces of Go
  9. Go Intellect
  =10. Wulu, Aya, Katsunari, Go Ahead

Notes:

MoGo and CrazyStone represent the new wave of Monte Carlo-based Go programs. There was recently a short article about the technology in Scientific American.

CrazyStone is by Remi Coulom, author of The Crazy Bishop. Other than him, there has been little successful cross-pollination between computer Go and computer Chess circles.

KCC Igo from North Korea had a lock on the Gifu Challenge title from 2003-2006. Silver Star, Haruka, and HandTalk are only sold in asian countries, as far as I know.

Ian

That's a cool news, but you just forgot one promising engine, Mango.

And I quote:

Mango is a new Go program based on Monte-Carlo Tree Search, developed by Guillaume Chaslot and Jahn-Takeshi Saito at the University of Maastricht.

Tournament results:

9x9:
Mango was 5th out of 10 in the 12th Computer Olympiad in Amsterdam.

13x13:
May KGS Tournament: 2nd out of 7 (First: CrazyStone Second: Mango Third: Mogo)
April KGS Tournament: 3rd out of 10 (First: CrazyStone Second: Mogo Third: Mango)
January KGS Tournament: 2nd out of 10 (First: Mogo Second: Mango Third: GNU)

19x19:
12th Computer Olympiad: 4th(shared) out of 8 (First: Mogo Second: CrazyStone Third: GNUGo )
March KGS Tournament: 4th out of 12 (First: Mogo Second: GNUGo Third: ManyFacesOfGo)

[mschribr]

I kinda quit work/research into Chess to persue Go development. Oddly even though it's supposedly harder to code for than Chess, I'm finding it a lot easier.

I think when people say it is harder to code a go program than a chess program they mean it is harder to code a go program than a chess program to be successful in beating strong human player. Are you saying your go program is more successful at beating stronger humans than your chess program?
Mark

[mschribr]

I really wish there was a forum for Go dev kinda like CCC. Though it seems all attempts at creating one go blank in favor for the popular computer-go mailing list.

Have you tried the Computer go forum at Go Discussions? It is at www.godiscussions.com/
Mark

[jshriver]

Are you saying your go program is more successful at beating stronger humans than your chess program?
Mark

Na, just meant for Chess I had a hard time for the longest time understanding recursion and tree traversal and the way 0x88 and bitboards work.

Go is a lot different, you don't use alpha-beta or trees as much. At least the approach I took signature structure evaluation. The way my wetware is wired it makes more sense to me.

Right now I have 3 subprojects going on. 1 is a 5x5 solver, another is a Go database with statistics on various group structures (breakpoints, etc). Then finally the engine.

[mschribr]

I kinda quit work/research into Chess to persue Go development.

Why did you switch from chess to go programming? Is it because computer chess has reached its goal and computer go will not for many decades? Or do you enjoy playing go more than chess?
Mark

[jshriver]

Why did you switch from chess to go programming? Is it because computer chess has reached its goal and computer go will not for many decades? Or do you enjoy playing go more than chess?
Mark

Little bit of both. I enjoy both games a lot but chess has been pretty much refined. Go is still in it's baby phase and lots of raw work going on.

I like to think of Monte-Carlo as the alpha-beta for chess. Not algorithm wise but progression wise. AB lets chess prune but wasn't the greatest so better algorithms came out like nega-max and many more enhancements. Go still has a lot more growth potential, software wise.

-Josh

[adams161]

i didnt think winboard supported shatranj. is this a mod winboard? where do you get it?
Mike Adams

[adams161]

When i found out winboard_F supported shatranj, i pulled out my old shatranj code and hooked it into the current pulsar. Seems to be working, but there was a bug in winboard_f in how it accepted one move , move 88 qe3+ from icc. I"ll probably release next month or when there is full interface support. winboard_f is in beta so i'm not surprised i found a bug and i'm sure with future releases it will get fixed.

features of pulsar that i have for most all wilds it supports and will add for shatranj are:

user settable evaluate options. i let users configure some evaluate options like piece values.

a book for white and black. i will only write a tiny shatranj book but users can hand write more.

handicapping of difficulty level with limited depth searches or reduced time searches.

yours
Mike Adams
www.adam16mr.org/pulsar.html

[Dirt]

where do you get it?

Look through the site the WWW button at the bottom of H.G.'s post links to. The download link may not be easy to find, but the search is worthwhile.