Re: [computer-go] Re: Opportunity to promote ...
The "well" at the end of the title is implied. And computers still can't play 19x19 Go anywhere near the master level. Ingo Althöfer wrote: Dear Bob Hearn, it is not what you have been looking for, but nevertheless I want to ask you if the title of your talk "Games Computers Can't Play" is still up-to-date. I would accept something like "Games Computers Could not play well before 2003", but Monte Carlo has changed our world. Ingo Althofer. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Re: Opportunity to promote ...
On Nov 18, 2008, at 7:43 AM, Michael Williams <[EMAIL PROTECTED] > wrote: The "well" at the end of the title is implied. And computers still can't play 19x19 Go anywhere near the master level. I'm not very familiar with go terms, but I think kyu means student and dan means master. It may be better to say something like computers may have reached amateur master using super computers, but are still far away from professional play Ingo Althöfer wrote: Dear Bob Hearn, it is not what you have been looking for, but nevertheless I want to ask you if the title of your talk "Games Computers Can't Play" is still up-to-date. I would accept something like "Games Computers Could not play well before 2003", but Monte Carlo has changed our world. Ingo Althofer. ___ 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] Re: Opportunity to promote ...
On 18-nov-08, at 11:25, Jason House wrote: On Nov 18, 2008, at 7:43 AM, Michael Williams <[EMAIL PROTECTED]> wrote: The "well" at the end of the title is implied. And computers still can't play 19x19 Go anywhere near the master level. I'm not very familiar with go terms, but I think kyu means student and dan means master. It may be better to say something like computers may have reached amateur master using super computers, but are still far away from professional play I believe 'kyu' and 'dan' have a very similar meaning in that both mean something like 'step', 'stage' or 'rank' depending on the context in which it's used. If you want to compare to chess I believe a 'master' would be around 1-kyu/ 1-dan (pro) and a 'grand-master' would be a 9-dan. Due to commercialization, the Japanese 'dan' level has steadily eroded over the past century, whereas in China and Korea usually only pro's have a 'dan' rank. Mark ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Re: Opportunity to promote ...
When Myungwan Kim guesstimated that Mogo might be dan-level, he certainly was speaking of amateur levels, not pro. He was playing Mogo with 9 stones. If I recall correctly, he later beat Mogo in a seven-stone game. A pro can not offer another pro seven stones. The difference between pro shodan and pro 9 dan might be about 3 stones. Go programs have progressed a lot in the last few years, but there's still room for a lot of improvement. Terry McIntyre <[EMAIL PROTECTED]> -- Libertarians Do It With Consent! - Original Message > From: Mark Boon <[EMAIL PROTECTED]> > To: computer-go > Sent: Tuesday, November 18, 2008 5:51:43 AM > Subject: Re: [computer-go] Re: Opportunity to promote ... > > > On 18-nov-08, at 11:25, Jason House wrote: > > > On Nov 18, 2008, at 7:43 AM, Michael Williams > wrote: > > > >> The "well" at the end of the title is implied. And computers still can't > play 19x19 Go anywhere near the master level. > > > > > > I'm not very familiar with go terms, but I think kyu means student and dan > means master. > > > > It may be better to say something like computers may have reached amateur > master using super computers, but are still far away from professional play > > > > I believe 'kyu' and 'dan' have a very similar meaning in that both mean > something like 'step', 'stage' or 'rank' depending on the context in which > it's > used. If you want to compare to chess I believe a 'master' would be around > 1-kyu/ 1-dan (pro) and a 'grand-master' would be a 9-dan. Due to > commercialization, the Japanese 'dan' level has steadily eroded over the past > century, whereas in China and Korea usually only pro's have a 'dan' rank. > > 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] Re: Opportunity to promote ...
Many Faces gained 5 ranks when I added MCTS to it (with about 7 months of full time work), so I have to agree that Monte Carlo changed our world. David > -Original Message- > From: [EMAIL PROTECTED] [mailto:computer-go- > [EMAIL PROTECTED] On Behalf Of Michael Williams > Sent: Tuesday, November 18, 2008 4:43 AM > To: computer-go > Subject: Re: [computer-go] Re: Opportunity to promote ... > > The "well" at the end of the title is implied. And computers still can't play > 19x19 Go anywhere near the master level. > > Ingo Althöfer wrote: > > Dear Bob Hearn, > > > > it is not what you have been looking for, but nevertheless > > I want to ask you if the title of your talk > > "Games Computers Can't Play" is still up-to-date. > > > > I would accept something like > > "Games Computers Could not play well before 2003", > > but Monte Carlo has changed our world. > > > > Ingo Althofer. > > ___ > 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] Re: Opportunity to promote ...
On Tue, 2008-11-18 at 08:28 -0800, David Fotland wrote: > Many Faces gained 5 ranks when I added MCTS to it (with about 7 months > of > full time work), so I have to agree that Monte Carlo changed our > world. I remember that you were not a "true believer" at first :-) - Don signature.asc Description: This is a digitally signed message part ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Re: Opportunity to promote ...
> From: Don Dailey <[EMAIL PROTECTED]> > > On Tue, 2008-11-18 at 08:28 -0800, David Fotland wrote: > > Many Faces gained 5 ranks when I added MCTS to it (with about 7 months > > of full time work), so I have to agree that Monte Carlo changed our > > world. > > I remember that you were not a "true believer" at first :-) Skepticism can be a good thing, it drives one to root out the flaws and improve the idea. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Re: Opportunity to promote ...
On 18-nov-08, at 14:32, Don Dailey wrote: On Tue, 2008-11-18 at 08:28 -0800, David Fotland wrote: Many Faces gained 5 ranks when I added MCTS to it (with about 7 months of full time work), so I have to agree that Monte Carlo changed our world. I remember that you were not a "true believer" at first :-) I think it was the surprisingly useful combination of UCT with Monte- Carlo that got the attention of the 'old school' Go programmers. Mark ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] RefBot (thought-) experiments
On Mon, Nov 17, 2008 at 11:13 PM, Michael Williams <[EMAIL PROTECTED]> wrote: > No one ever alleged that pure AMAF or pure MC was infinitely scalable. My point is that in many cases, they doesn't even keep all of their benefits, after some number of trials have been run. So, running 10k playouts can be significantly worse than running 5k playouts. Weston ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] RefBot (thought-) experiments
Weston Markham wrote: On Mon, Nov 17, 2008 at 11:13 PM, Michael Williams <[EMAIL PROTECTED]> wrote: No one ever alleged that pure AMAF or pure MC was infinitely scalable. My point is that in many cases, they doesn't even keep all of their benefits, after some number of trials have been run. So, running 10k playouts can be significantly worse than running 5k playouts. It doesn't make any sense to me from a theoretical perspective. Do you have empirical evidence? ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Congratulations to Many Faces of Go!
My report on Sunday's KGS bot tournament is now available at http://www.weddslist.com/kgs/past/44/index.html Many Faces of Go was undefeated in both divisions. Nick -- Nick Wedd[EMAIL PROTECTED] ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
RE: [computer-go] Re: Opportunity to promote ...
I was a skeptic until the 2007 UEC cup. Then it was obvious that MCTS was stronger than the traditional programs. > -Original Message- > From: [EMAIL PROTECTED] [mailto:computer-go- > [EMAIL PROTECTED] On Behalf Of Mark Boon > Sent: Tuesday, November 18, 2008 8:56 AM > To: [EMAIL PROTECTED]; computer-go > Subject: Re: [computer-go] Re: Opportunity to promote ... > > > On 18-nov-08, at 14:32, Don Dailey wrote: > > > On Tue, 2008-11-18 at 08:28 -0800, David Fotland wrote: > >> Many Faces gained 5 ranks when I added MCTS to it (with about 7 > >> months > >> of > >> full time work), so I have to agree that Monte Carlo changed our > >> world. > > > > I remember that you were not a "true believer" at first :-) > > > > I think it was the surprisingly useful combination of UCT with Monte- > Carlo that got the attention of the 'old school' Go programmers. > > 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] Re: Opportunity to promote ...
> > I think it was the surprisingly useful combination of UCT with Monte-Carlo > that got the attention of the 'old school' Go programmers. I would say "Monte-Carlo + Tree Search" rather than "Monte-Carlo + UCT". You can have a very strong program without UCT. You can't without the incremental tree + the Monte-Carlo search. Best regards, Olivier ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] RefBot (thought-) experiments
On Tue, 2008-11-18 at 12:02 -0500, Michael Williams wrote: > Weston Markham wrote: > > On Mon, Nov 17, 2008 at 11:13 PM, Michael Williams > > <[EMAIL PROTECTED]> wrote: > >> No one ever alleged that pure AMAF or pure MC was infinitely scalable. > > > > My point is that in many cases, they doesn't even keep all of their > > benefits, after some number of trials have been run. So, running 10k > > playouts can be significantly worse than running 5k playouts. > > It doesn't make any sense to me from a theoretical perspective. Do you have > empirical evidence? I have never seen any evidence that it gets weaker with the number of playouts. In most normal situations I would think that it's inclined to prefer one specific move and that move is increasingly likely to get selected as you do more playouts. And since this style of bot is relatively weak, that move will not be best. But that is not the same as saying that it will play better in general if you limit the number of playouts. I suspect the strength level approaches some asymptote, in other words it will play better and better with the number of playouts, but will never quite reach some very modest level of play. At some point you may need to add a billion playouts to add 1 ELO point. - Don > > ___ > computer-go mailing list > computer-go@computer-go.org > http://www.computer-go.org/mailman/listinfo/computer-go/ signature.asc Description: This is a digitally signed message part ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Re: Opportunity to promote ...
On Nov 17, 2008, at 11:34 PM, Ingo Althöfer wrote: Dear Bob Hearn, it is not what you have been looking for, but nevertheless I want to ask you if the title of your talk "Games Computers Can't Play" is still up-to-date. I would accept something like "Games Computers Could not play well before 2003", but Monte Carlo has changed our world. Dear Ingo, First, the title is deliberately provocative. Also, though, the talk is not just about go: some of it is about formally undecidable games, that computers provably can't play well (and of course, that humans can't either!). Surprisingly, some of these games can be played with finite physical resources (unlike, say, an infinite Turing-machine tape). One real game that is a good candidate for being undecidable (though it hasn't been proven) is Rengo Kriegspiel: team blindfold go. I've played this a few times. Bob - Robert A. Hearn Neukom Institute for Computational Science, Dartmouth College [EMAIL PROTECTED] http://www.dartmouth.edu/~rah/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Re: Opportunity to promote ...
I had never heard of that. A Google search turned up this list of interesting Go variants: http://www.usgo.org/resources/downloads/deviantgo.pdf Bob Hearn wrote: On Nov 17, 2008, at 11:34 PM, Ingo Althöfer wrote: Dear Bob Hearn, it is not what you have been looking for, but nevertheless I want to ask you if the title of your talk "Games Computers Can't Play" is still up-to-date. I would accept something like "Games Computers Could not play well before 2003", but Monte Carlo has changed our world. Dear Ingo, First, the title is deliberately provocative. Also, though, the talk is not just about go: some of it is about formally undecidable games, that computers provably can't play well (and of course, that humans can't either!). Surprisingly, some of these games can be played with finite physical resources (unlike, say, an infinite Turing-machine tape). One real game that is a good candidate for being undecidable (though it hasn't been proven) is Rengo Kriegspiel: team blindfold go. I've played this a few times. Bob - Robert A. Hearn Neukom Institute for Computational Science, Dartmouth College [EMAIL PROTECTED] http://www.dartmouth.edu/~rah/ ___ 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] RefBot (thought-) experiments
On Tue, Nov 18, 2008 at 12:02 PM, Michael Williams <[EMAIL PROTECTED]> wrote: > It doesn't make any sense to me from a theoretical perspective. Do you have > empirical evidence? I used to have data on this, from a program that I think was very nearly identical to Don's reference spec. When I get a chance, I'll try to reproduce it. Weston ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] RefBot (thought-) experiments
On Tue, 2008-11-18 at 14:17 -0500, Weston Markham wrote: > On Tue, Nov 18, 2008 at 12:02 PM, Michael Williams > <[EMAIL PROTECTED]> wrote: > > It doesn't make any sense to me from a theoretical perspective. Do you have > > empirical evidence? > > I used to have data on this, from a program that I think was very > nearly identical to Don's reference spec. When I get a chance, I'll > try to reproduce it. Unless the difference is large, you will have to run thousands of games to back this up. - Don > > Weston > ___ > computer-go mailing list > computer-go@computer-go.org > http://www.computer-go.org/mailman/listinfo/computer-go/ signature.asc Description: This is a digitally signed message part ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
RE: [computer-go] Re: Opportunity to promote ...
It depends very much on what exactly you mean by "amateur master level". Is it a level that compares to amateur master level in chess? And what is amateur master level in chess? USCF master, FIDE master or international master? Some time ago I participated in a discussion about comparing chess titles to go ranks by evaluating effort, prestige and other factors [1]. In my opinion: 1: USCF master compares to about 4d 2: FIDE master compares to about 6d 3: International master compares to about 7d/1p If we assume that mogo on Huygens is about 2d, it think it hasn't quite reached "amateur master level" in any chess meaning of the word. It would only compare to a fairly strong USCF class A chess player. Dave [1] discussion and my proposal in a table near the bottom of http://senseis.xmp.net/?FIDETitlesAndEGFGoRatings Van: [EMAIL PROTECTED] namens Jason House Verzonden: di 18-11-2008 14:25 Aan: computer-go Onderwerp: Re: [computer-go] Re: Opportunity to promote ... On Nov 18, 2008, at 7:43 AM, Michael Williams <[EMAIL PROTECTED] > wrote: > The "well" at the end of the title is implied. And computers still > can't play 19x19 Go anywhere near the master level. I'm not very familiar with go terms, but I think kyu means student and dan means master. It may be better to say something like computers may have reached amateur master using super computers, but are still far away from professional play > > > Ingo Althöfer wrote: >> Dear Bob Hearn, >> it is not what you have been looking for, but nevertheless >> I want to ask you if the title of your talk >> "Games Computers Can't Play" is still up-to-date. >> I would accept something like "Games Computers Could not play well >> before 2003", >> but Monte Carlo has changed our world. Ingo Althofer. > > ___ > 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] RefBot (thought-) experiments
On 11/18/08, Michael Williams <[EMAIL PROTECTED]> wrote: > > > > It doesn't make any sense to me from a theoretical perspective. Do you > have empirical evidence? I agree that empirical evidence is required, but theoretically, if MC converges to something that is not perfect play, then as the number of playouts goes up, the probability of playing a different move (the perfect move) goes down, so the programme could get weaker. Does that make sense? The effect may depend on the opponent. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] RefBot (thought-) experiments
On Tue, 2008-11-18 at 19:42 +, Oliver Lewis wrote: > It doesn't make any sense to me from a theoretical > perspective. Do you have empirical evidence? > > > I agree that empirical evidence is required, but theoretically, if MC > converges to something that is not perfect play, then as the number of > playouts goes up, the probability of playing a different move (the > perfect move) goes down, so the programme could get weaker. Does that > make sense? No. Your reasoning is wrong. It's true that there are many cases where it won't play the best move with infinite playouts, but there are also many times when it will. Game playing ability is about errors, not playing good moves once in a while. A good move is really just a move that isn't an error. Playing at a high level is like walking through a minefield - if you can avoid the bad places, you win. This is real obvious in chess, not quite as obvious in Go. A 1 ply chess search still plays reasonable looking chess but with many tactical errors. Add an extra ply of search, and you eliminate a whole class of stupid errors, but you still have a lot of errors left. Each additional ply eliminates more classes of errors. As humans we tend to think in the other direction, we view moves in terms of goodness as if it's on a sliding scale. How many times have you heard the terminology, "that was a master level move", or "that was a 5 dan move."Or "that was ok, but this is better." It makes it sound as if each move has a rank associated with it. You should also consider that even weak players play the very best move some of the time. In chess I probably play the same moves a "grandmaster" would play a high percentage of the time, but I am far from a grandmaster in strength. It's those pesky bad moves that kill you. - Don > The effect may depend on the opponent. signature.asc Description: This is a digitally signed message part ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] RefBot (thought-) experiments
From: Oliver Lewis <[EMAIL PROTECTED]> On 11/18/08, Michael Williams <[EMAIL PROTECTED]> wrote: It doesn't make any sense to me from a theoretical perspective. Do you have empirical evidence? > I agree that empirical evidence is required, but theoretically, if MC > converges to something that is not perfect play, then as the number of > playouts goes up, the probability of playing a different move (the perfect > move) goes down, so the programme could get weaker. Not converging to the best move does not imply that stopping sooner would have found a better move. It might be looking in the wrong place entirely - like a drunk who searches for his keys under the lamp post because the light is better there, instead of near the car where he dropped them. An argument might be made that more playouts fail to improve the quality, and the time could be better spent elsewhere. I've noticed a trend to separate the discussion of "quality of move selection in the tree" versus "quality of move selection in playouts", for instance. One might argue for fast, light playouts for evaluation purposes, coupled with more intensive top-level search for evaluation candidates. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Re: Opportunity to promote ...
Dear Bob, thanks for your explanations. Now I see clearer. > First, the title is deliberately provocative. Accepted. > Also, though, the talk is not just about go: some of it is about > formally undecidable games, that computers provably can't play > well (and of course, that humans can't either!). I see. How much sense does it make to think about Monte-Carlo approaches to certain games that are undecidable? (Probably it does make a lot of sense in certain cases.) > Surprisingly, some of these games can be played with > finite physical resources (unlike, say, an infinite Turing-machine > tape). One real game that is a good candidate for being undecidable > (though it hasn't been proven) is Rengo Kriegspiel: team blindfold go. Here, indeed Monte-Carlo might "work" - for instance against opponents with limited capabilities. * When I read your original mail the following three main steps of game tree search came to my mind: (i) Zermelo's basic tree approach (with minimaxing) in 1911 or 1912. (ii) Shannon's paper from 1950 with the proposal to prune the (large) tree and to evaluate the artificial leaves by heuristic values. (iii) The Monte Carlo approach: getting an evaluation for a position not by applying a heuristic but by playing many "random" games from this position till the very end. These three mile-stones in mind, there are not many games left, where computers are really weak. Ingo. PS: When you meet Elwyn Berlekamp in the symposium you may tell him that looking at the game Clobber in the 2002-Dagstuhl workshop (where he was one of the organizers and I one of the participants) was my starting point for becoming a "game designer with computer help" and thus indrectly the starting point for me to investigate Monte Carlo, first in games with random elements. -- Ist Ihr Browser Vista-kompatibel? Jetzt die neuesten Browser-Versionen downloaden: http://www.gmx.net/de/go/browser ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] One-sided 2-inch Rules
Hello, one of the basic problems of go newbies is their tendency to place the next stone near to the latest stone of the opponent. Sometimes this is called the "2-inch heuristic of beginners". What do you think about a formalized variant of Go with one-sided distance-k rule? > Let k be some natural number. > The normal rules of Go hold, except for one thing: > When possible, White has to place his next stone > within distance k (in city-block metric) of the latest > stone of Black. If there is no feasible move of this type > the stone has to be placed within the smallest > possible city-block distance of the latest stone of > Black. White may pass at any time. Example: > On 19x19 board k=36 would mean no restriction at all.) * What should be fair values of komi(k) or fair numbers of handicap stones? * Main question: How strong would MCTS-based programs be in this variant(s)? * Would computers be stronger than humans for certain values of k? Ingo. -- Ist Ihr Browser Vista-kompatibel? Jetzt die neuesten Browser-Versionen downloaden: http://www.gmx.net/de/go/browser ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] One-sided 2-inch Rules
I think computers would be much better at this game (than they are at Go) because you have vastly reduced the branching factor of the game. Ingo Althöfer wrote: Hello, one of the basic problems of go newbies is their tendency to place the next stone near to the latest stone of the opponent. Sometimes this is called the "2-inch heuristic of beginners". What do you think about a formalized variant of Go with one-sided distance-k rule? Let k be some natural number. The normal rules of Go hold, except for one thing: When possible, White has to place his next stone within distance k (in city-block metric) of the latest stone of Black. If there is no feasible move of this type the stone has to be placed within the smallest possible city-block distance of the latest stone of Black. White may pass at any time. Example: On 19x19 board k=36 would mean no restriction at all.) * What should be fair values of komi(k) or fair numbers of handicap stones? * Main question: How strong would MCTS-based programs be in this variant(s)? * Would computers be stronger than humans for certain values of k? Ingo. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] One-sided 2-inch Rules
I think a computer would play this variant well if k was small. To make the move generation consistent, the first move should be played as if there was a previous move to the center perhaps. Ladders would probably still be an issue. - Don On Tue, 2008-11-18 at 23:20 +0100, "Ingo Althöfer" wrote: > Hello, > > one of the basic problems of go newbies > is their tendency to place the next stone > near to the latest stone of the opponent. > Sometimes this is called the "2-inch heuristic > of beginners". > > What do you think about a formalized variant > of Go with one-sided distance-k rule? > > > Let k be some natural number. > > The normal rules of Go hold, except for one thing: > > When possible, White has to place his next stone > > within distance k (in city-block metric) of the latest > > stone of Black. If there is no feasible move of this type > > the stone has to be placed within the smallest > > possible city-block distance of the latest stone of > > Black. White may pass at any time. Example: > > On 19x19 board k=36 would mean no restriction at all.) > > * What should be fair values of komi(k) or fair numbers > of handicap stones? > > * Main question: How strong would MCTS-based programs be in this variant(s)? > > * Would computers be stronger than humans for certain values of k? > > Ingo. > signature.asc Description: This is a digitally signed message part ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] One-sided 2-inch Rules
Well, "vastly" when k is small. The only way to find a good Komi would be testing and guesstimating. I think MCTS would be well suited to this variant because you still have the problem of difficulty in finding a good evaluation function and MCTS solves that. Computers would probably be stronger than humans for sufficiently small values of k. With the small branching factor, the computer would be able to build a very deep tree. Michael Williams wrote: I think computers would be much better at this game (than they are at Go) because you have vastly reduced the branching factor of the game. Ingo Althöfer wrote: Hello, one of the basic problems of go newbies is their tendency to place the next stone near to the latest stone of the opponent. Sometimes this is called the "2-inch heuristic of beginners". What do you think about a formalized variant of Go with one-sided distance-k rule? Let k be some natural number. The normal rules of Go hold, except for one thing: When possible, White has to place his next stone within distance k (in city-block metric) of the latest stone of Black. If there is no feasible move of this type the stone has to be placed within the smallest possible city-block distance of the latest stone of Black. White may pass at any time. Example: On 19x19 board k=36 would mean no restriction at all.) * What should be fair values of komi(k) or fair numbers of handicap stones? * Main question: How strong would MCTS-based programs be in this variant(s)? * Would computers be stronger than humans for certain values of k? Ingo. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] One-sided 2-inch Rules
On Tue, 2008-11-18 at 17:31 -0500, Michael Williams wrote: > Well, "vastly" when k is small. > > The only way to find a good Komi would be testing and guesstimating. > > I think MCTS would be well suited to this variant because you still have the > problem of difficulty in finding a good evaluation function and MCTS solves > that. > > Computers would probably be stronger than humans for sufficiently small > values of k. With the small branching factor, the computer would be able to > build a > very deep tree. I keep thinking about this. My first idea was that alpha/beta might be a better choice with a low branching factor but then you have the evaluation issue. But does MCTS handle that as well as we think? I'm not completely sure of this. It seems like that each move would require more precision to be right. It may be that a lot of move are very obvious. I think a heavier playout would be required.This is just guesswork of course. - Don > > > Michael Williams wrote: > > I think computers would be much better at this game (than they are at > > Go) because you have vastly reduced the branching factor of the game. > > > > Ingo Althöfer wrote: > >> Hello, > >> > >> one of the basic problems of go newbies > >> is their tendency to place the next stone near to the latest stone of > >> the opponent. > >> Sometimes this is called the "2-inch heuristic > >> of beginners". > >> > >> What do you think about a formalized variant > >> of Go with one-sided distance-k rule? > >> > >>> Let k be some natural number. > >>> The normal rules of Go hold, except for one thing: > >>> When possible, White has to place his next stone > >>> within distance k (in city-block metric) of the latest > >>> stone of Black. If there is no feasible move of this type > >>> the stone has to be placed within the smallest > >>> possible city-block distance of the latest stone of > >>> Black. White may pass at any time. Example: > >>> On 19x19 board k=36 would mean no restriction at all.) > >> > >> * What should be fair values of komi(k) or fair numbers > >> of handicap stones? > >> > >> * Main question: How strong would MCTS-based programs be in this > >> variant(s)? > >> > >> * Would computers be stronger than humans for certain values of k? > >> > >> Ingo. > >> > > > > > > ___ > computer-go mailing list > computer-go@computer-go.org > http://www.computer-go.org/mailman/listinfo/computer-go/ signature.asc Description: This is a digitally signed message part ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] One-sided 2-inch Rules
for small k, this should give a massive advantage to black. the additional requirement that white place a stone within the smallest cityblock distance of the last stone whenever he has no valid move within distance k of black's last move is an even more substantial advantage for black. i'm thinking that black can set up situations that cause white to make bad shape. if not, it's an unimportant rule, but otherwise it's a big advantage to black. just my $0.02. s. On Tue, Nov 18, 2008 at 5:20 PM, "Ingo Althöfer" <[EMAIL PROTECTED]> wrote: > Hello, > > one of the basic problems of go newbies > is their tendency to place the next stone > near to the latest stone of the opponent. > Sometimes this is called the "2-inch heuristic > of beginners". > > What do you think about a formalized variant > of Go with one-sided distance-k rule? > >> Let k be some natural number. >> The normal rules of Go hold, except for one thing: >> When possible, White has to place his next stone >> within distance k (in city-block metric) of the latest >> stone of Black. If there is no feasible move of this type >> the stone has to be placed within the smallest >> possible city-block distance of the latest stone of >> Black. White may pass at any time. Example: >> On 19x19 board k=36 would mean no restriction at all.) > > * What should be fair values of komi(k) or fair numbers > of handicap stones? > > * Main question: How strong would MCTS-based programs be in this variant(s)? > > * Would computers be stronger than humans for certain values of k? > > Ingo. > > -- > Ist Ihr Browser Vista-kompatibel? Jetzt die neuesten > Browser-Versionen downloaden: http://www.gmx.net/de/go/browser > ___ > 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] One-sided 2-inch Rules
I assumed the rules were symmetric, in that black also had to place his stones within distance k to white's last move. steve uurtamo wrote: for small k, this should give a massive advantage to black. the additional requirement that white place a stone within the smallest cityblock distance of the last stone whenever he has no valid move within distance k of black's last move is an even more substantial advantage for black. i'm thinking that black can set up situations that cause white to make bad shape. if not, it's an unimportant rule, but otherwise it's a big advantage to black. just my $0.02. s. On Tue, Nov 18, 2008 at 5:20 PM, "Ingo Althöfer" <[EMAIL PROTECTED]> wrote: Hello, one of the basic problems of go newbies is their tendency to place the next stone near to the latest stone of the opponent. Sometimes this is called the "2-inch heuristic of beginners". What do you think about a formalized variant of Go with one-sided distance-k rule? Let k be some natural number. The normal rules of Go hold, except for one thing: When possible, White has to place his next stone within distance k (in city-block metric) of the latest stone of Black. If there is no feasible move of this type the stone has to be placed within the smallest possible city-block distance of the latest stone of Black. White may pass at any time. Example: On 19x19 board k=36 would mean no restriction at all.) * What should be fair values of komi(k) or fair numbers of handicap stones? * Main question: How strong would MCTS-based programs be in this variant(s)? * Would computers be stronger than humans for certain values of k? Ingo. -- Ist Ihr Browser Vista-kompatibel? Jetzt die neuesten Browser-Versionen downloaden: http://www.gmx.net/de/go/browser ___ 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/
[computer-go] Re: Congratulations to Many Faces of Go!
Nick Wedd: <[EMAIL PROTECTED]>: >My report on Sunday's KGS bot tournament is now available at >http://www.weddslist.com/kgs/past/44/index.html >Many Faces of Go was undefeated in both divisions. Did Many Faces 1 and 2 share the same cluster or one for each? Hideki -- [EMAIL PROTECTED] (Kato) ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Re: One-sided 2-inch Rules
> one of the basic problems of go newbies > is their tendency to place the next stone > near to the latest stone of the opponent. > Sometimes this is called the "2-inch heuristic > of beginners". How about going the other way, forcing Monte-Carlo simulations onto a coarser grid in the hope of quickly finding information that might be lost in full-grid simulations (by analogy with multigrid methods - there even exists something called multigrid Monte-Carlo; perhaps some of the mathematically inclined can have a look?). As a trivial idea, if both playout-parties are forced to play on all odd intersections before being allowed onto the remaining ones (staking broader claims before entering detailed fights), how would that affect the statistics, and is there a benefit to be had from combining coarse and fine simulations? Claus ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Re: One-sided 2-inch Rules
I don't think that would be effective because you've completely changed the rules and hence the tactics of the game. Claus Reinke wrote: one of the basic problems of go newbies is their tendency to place the next stone near to the latest stone of the opponent. Sometimes this is called the "2-inch heuristic of beginners". How about going the other way, forcing Monte-Carlo simulations onto a coarser grid in the hope of quickly finding information that might be lost in full-grid simulations (by analogy with multigrid methods - there even exists something called multigrid Monte-Carlo; perhaps some of the mathematically inclined can have a look?). As a trivial idea, if both playout-parties are forced to play on all odd intersections before being allowed onto the remaining ones (staking broader claims before entering detailed fights), how would that affect the statistics, and is there a benefit to be had from combining coarse and fine simulations? Claus ___ 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] Newbie
I'm writing my engine from scratch and have a curious question. When my best friend an American 1dan lvl player (who has played in Japan) taught me the game. I love it. :) Though as I get more in depth, and programming wise, have no idea what "rule set" to follow. Not sure what I was "taught" guessing Japanese rules. But curious what the various "rule sets" and how do the differ/compare especially from a programmer and player point of view. Also, while I love computer-go.. is there a go-player mailing list? -Josh ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Newbie
Chinese rule variants are simpler to code than Japanese. This is because the game can be played to the bitter end without affecting the final score. So you should probably start there. There were recently some discussions on this list about how to handle Japanese scoring. Joshua Shriver wrote: I'm writing my engine from scratch and have a curious question. When my best friend an American 1dan lvl player (who has played in Japan) taught me the game. I love it. :) Though as I get more in depth, and programming wise, have no idea what "rule set" to follow. Not sure what I was "taught" guessing Japanese rules. But curious what the various "rule sets" and how do the differ/compare especially from a programmer and player point of view. Also, while I love computer-go.. is there a go-player mailing list? -Josh ___ 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] Newbie
On Tue, 2008-11-18 at 20:30 -0500, Joshua Shriver wrote: > I'm writing my engine from scratch and have a curious question. When > my best friend an American 1dan lvl player (who has played in Japan) > taught me the game. I love it. :) > > Though as I get more in depth, and programming wise, have no idea what > "rule set" to follow. Not sure what I was "taught" guessing Japanese > rules. But curious what the various "rule sets" and how do the > differ/compare especially from a programmer and player point of view. You've been on this list a long time, haven't you? Anyway, I don't think there is much of a question that Chinese rules are much better for getting started with computer GO. If you are newbie, Japanese rules will frustrate you. You can later migrate to Japanese rules if you want - but CGOS and KGS tournaments for go programs use Chinese rules. - Don > > Also, while I love computer-go.. is there a go-player mailing list? > > -Josh > ___ > computer-go mailing list > computer-go@computer-go.org > http://www.computer-go.org/mailman/listinfo/computer-go/ signature.asc Description: This is a digitally signed message part ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Newbie
> You've been on this list a long time, haven't you? > Yes :) I started by tinkering with Monte Carlo in VRML for a hardware solution, but aiming for a pure C, aimed at x86 engine. > > Anyway, I don't think there is much of a question that Chinese rules > are much better for getting started with computer GO. If you are > newbie, Japanese rules will frustrate you. You can later migrate to > Japanese rules if you want - but CGOS and KGS tournaments for go > programs use Chinese rules. > > - Don > I'll have to do some googling. To be honest I dont know what the diff is in Japanese, Chinese, Korean Go rules. -Josh ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
RE: [computer-go] Re: Congratulations to Many Faces of Go!
One for each. Actually they were running on a 128 core cluster. The current code only scales to 32 cores, so only half the cluster was used. David > -Original Message- > From: [EMAIL PROTECTED] [mailto:computer-go- > [EMAIL PROTECTED] On Behalf Of Hideki Kato > Sent: Tuesday, November 18, 2008 4:21 PM > To: computer-go > Subject: [computer-go] Re: Congratulations to Many Faces of Go! > > Nick Wedd: <[EMAIL PROTECTED]>: > >My report on Sunday's KGS bot tournament is now available at > >http://www.weddslist.com/kgs/past/44/index.html > >Many Faces of Go was undefeated in both divisions. > > Did Many Faces 1 and 2 share the same cluster or one for each? > > Hideki > -- > [EMAIL PROTECTED] (Kato) > ___ > 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] Re: Opportunity to promote ...
Bob, take a look at Arimaa. It's a relatively new game, but fun for people, and very very difficult for computers. David > -Original Message- > From: [EMAIL PROTECTED] [mailto:computer-go- > [EMAIL PROTECTED] On Behalf Of Bob Hearn > Sent: Tuesday, November 18, 2008 10:13 AM > To: computer-go > Subject: Re: [computer-go] Re: Opportunity to promote ... > > > On Nov 17, 2008, at 11:34 PM, Ingo Althöfer wrote: > > > Dear Bob Hearn, > > > > it is not what you have been looking for, but nevertheless > > I want to ask you if the title of your talk > > "Games Computers Can't Play" is still up-to-date. > > > > I would accept something like > > "Games Computers Could not play well before 2003", > > but Monte Carlo has changed our world. > > Dear Ingo, > > First, the title is deliberately provocative. Also, though, the talk > is not just about go: some of it is about formally undecidable games, > that computers provably can't play well (and of course, that humans > can't either!). Surprisingly, some of these games can be played with > finite physical resources (unlike, say, an infinite Turing-machine > tape). One real game that is a good candidate for being undecidable > (though it hasn't been proven) is Rengo Kriegspiel: team blindfold go. > I've played this a few times. > > Bob > > - > Robert A. Hearn > Neukom Institute for Computational Science, Dartmouth College > [EMAIL PROTECTED] > http://www.dartmouth.edu/~rah/ > > > ___ > 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/
