Re: [computer-go] Dynamic komi at high handicaps
One last rumination on dynamic komi: The main objection against introducing dynamic komi is that it ignores the true goal of winning by half a point. The power of the win/loss step function as scoring function underscores the validity of this critique. And yet, the current behaviour of mc bots, when either leading or trailing by a large margin, resembles random play. The simple reason for this is that either every move is a win or every move is a loss. So from the perspective of securing a win, every move is as good as any other move. Humans know how to handle these situations. They try to catch up from behind, or try to play safely while defending enough of a winning margin. For a bot, there are some numerical clues when it is missbehaving. When the calculated win rate is either very high or low and many move candidates have almost identical win rates, the bot is in coin toss country. A simple rule would be this: define a minimum value that has to separate the win rate of the 2 best move candidates. Do a normal search without komi. If the minimum difference is not reached, do a new a new search with some komi, but only allow the moves within the minimum value range from the best candidate. Repeat this with progressively higher komi until the two best candidates are sufficiently separated.(Or until the win rate is in a defined middle region) There can be some traps here, a group of moves can all accomplish the same critical goal. But I'm sure this can be handled. The main idea is to look for a less ambitious gloal when the true goal cannot be reached. (Or a more ambitious goal when it is allready satisfied). By only allowing moves that are in a statistical tie in the 0 komi search, it can be assured that short term gain doesn't compromise the long term goal. Stefan___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Human vs Computer in IEEE conference
Forthcoming human-vs-computer games in go: http://www.althofer.de/ieee-go-0.jpg http://www.althofer.de/ieee-go-1.jpg http://www.althofer.de/ieee-go-2.jpg http://www.althofer.de/ieee-go-3.jpg http://oase.nutn.edu.tw/FUZZ_IEEE_2009/result.htm Ingo. -- GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT! Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01 ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Dynamic komi at high handicaps
One must decide if the goal is to improve the program or to improve it's playing behavior when it's in a dead won or dead lost positions. It's my belief that you can probably cannot improve the playing strength soley with komi manipulation, but at a slight decrease in playing strength you can probably improve the behavior, as measured by a willingness to fight for space that is technically not relevant to the goal of winning the game.And only then if this is done carefully. However I believe there are better ways, such a pre-ordering the moves. Even if this can prove to be a gain, you are really working very hard to find something that only kicks in when the game is already decided - how to play when the game is already won or already lost.But only the case when the game is lost is this very interesting from the standpoint of making the program stronger. And even this case is not THAT interesting, because if you are losing, on average you are losing to stronger players. So you are working hard on an algorithm to beat stronger players when you are in a dead lost game? How much sense does that make? So the only realistic pay-off here is how to salvage lost games against players that are relatively close in strength since you can expect not to be in this situation very often agaist really weak players.So you are hoping to bamboozle players who are not not weaker than you - in situations where you have been bamboozled (since you are losing, you are the one being outplayed.) That is why I believe that at best you are looking at only a very minor improvement.If I were working on this problem I would be focused only on the playing style, not the playing strength. If you want more than the most minor playing strength improvement out of this algorithm, you have to start using it BEFORE the loss is clear, but then you are no longer playing for the win when you lower your goals, you are playing for the loss. - Don 2009/8/19 Stefan Kaitschick stefan.kaitsch...@hamburg.de One last rumination on dynamic komi: The main objection against introducing dynamic komi is that it ignores the true goal of winning by half a point. The power of the win/loss step function as scoring function underscores the validity of this critique. And yet, the current behaviour of mc bots, when either leading or trailing by a large margin, resembles random play. The simple reason for this is that either every move is a win or every move is a loss. So from the perspective of securing a win, every move is as good as any other move. Humans know how to handle these situations. They try to catch up from behind, or try to play safely while defending enough of a winning margin. For a bot, there are some numerical clues when it is missbehaving. When the calculated win rate is either very high or low and many move candidates have almost identical win rates, the bot is in coin toss country. A simple rule would be this: define a minimum value that has to separate the win rate of the 2 best move candidates. Do a normal search without komi. If the minimum difference is not reached, do a new a new search with some komi, but only allow the moves within the minimum value range from the best candidate. Repeat this with progressively higher komi until the two best candidates are sufficiently separated.(Or until the win rate is in a defined middle region) There can be some traps here, a group of moves can all accomplish the same critical goal. But I'm sure this can be handled. The main idea is to look for a less ambitious gloal when the true goal cannot be reached. (Or a more ambitious goal when it is allready satisfied). By only allowing moves that are in a statistical tie in the 0 komi search, it can be assured that short term gain doesn't compromise the long term goal. Stefan ___ 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] Dynamic komi at high handicaps
Don, what you write is certainly true for even games, but I think the problem is a real one in high handicap games with the computer as white. I use a hack to make Valkyria continue playing the opening in handicap games as white. It is forbidden to resign in the opening and early middle game because it would if it could. To rephrase your argument for even games, the problem situation should never occur because the losing player *should out of courtesy* resign long before the evalutaion become so skewed. But this does not apply to h9 games on 19x19 for example. And if I am not mistaken strong heavy playouts evaluates such positions very pessimistically, and thus we have a problem to solve, which grows with increasing playing strength. Still stronger programs will discriminate between bad and good moves even with extreme scores, so I think the dimensions of this problem is exaggerated. -Magnus Quoting Don Dailey dailey@gmail.com: One must decide if the goal is to improve the program or to improve it's playing behavior when it's in a dead won or dead lost positions. It's my belief that you can probably cannot improve the playing strength soley with komi manipulation, but at a slight decrease in playing strength you can probably improve the behavior, as measured by a willingness to fight for space that is technically not relevant to the goal of winning the game.And only then if this is done carefully. However I believe there are better ways, such a pre-ordering the moves. Even if this can prove to be a gain, you are really working very hard to find something that only kicks in when the game is already decided - how to play when the game is already won or already lost.But only the case when the game is lost is this very interesting from the standpoint of making the program stronger. And even this case is not THAT interesting, because if you are losing, on average you are losing to stronger players. So you are working hard on an algorithm to beat stronger players when you are in a dead lost game? How much sense does that make? So the only realistic pay-off here is how to salvage lost games against players that are relatively close in strength since you can expect not to be in this situation very often agaist really weak players.So you are hoping to bamboozle players who are not not weaker than you - in situations where you have been bamboozled (since you are losing, you are the one being outplayed.) That is why I believe that at best you are looking at only a very minor improvement.If I were working on this problem I would be focused only on the playing style, not the playing strength. If you want more than the most minor playing strength improvement out of this algorithm, you have to start using it BEFORE the loss is clear, but then you are no longer playing for the win when you lower your goals, you are playing for the loss. - Don 2009/8/19 Stefan Kaitschick stefan.kaitsch...@hamburg.de One last rumination on dynamic komi: The main objection against introducing dynamic komi is that it ignores the true goal of winning by half a point. The power of the win/loss step function as scoring function underscores the validity of this critique. And yet, the current behaviour of mc bots, when either leading or trailing by a large margin, resembles random play. The simple reason for this is that either every move is a win or every move is a loss. So from the perspective of securing a win, every move is as good as any other move. Humans know how to handle these situations. They try to catch up from behind, or try to play safely while defending enough of a winning margin. For a bot, there are some numerical clues when it is missbehaving. When the calculated win rate is either very high or low and many move candidates have almost identical win rates, the bot is in coin toss country. A simple rule would be this: define a minimum value that has to separate the win rate of the 2 best move candidates. Do a normal search without komi. If the minimum difference is not reached, do a new a new search with some komi, but only allow the moves within the minimum value range from the best candidate. Repeat this with progressively higher komi until the two best candidates are sufficiently separated.(Or until the win rate is in a defined middle region) There can be some traps here, a group of moves can all accomplish the same critical goal. But I'm sure this can be handled. The main idea is to look for a less ambitious gloal when the true goal cannot be reached. (Or a more ambitious goal when it is allready satisfied). By only allowing moves that are in a statistical tie in the 0 komi search, it can be assured that short term gain doesn't compromise the long term goal. Stefan ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ -- Magnus Persson
Re: [computer-go] Dynamic komi at high handicaps
Consider the game when computer is black, with 7 stones against a very strong human opponent. Computer thinks every move is a winning move; it plays randomly; a half-point win is as good as a 70-point win. Pro gains ground as computer makes slack moves, taking slightly less than its full due. At some point, the computer, being the weaker player, makes one slack move too many and loses the game. Rinse and repeat. At some point, it dawns on the programmer: must attack to win handicap games. Must be a little bit greedy, to slow down the process of attrition. Dynamic komi models something real: the significant advantage of the computer in a handicap game. It tries to preserve as much of that advantage as possible. I don't know if it will work for computer vs human games. I do know that a similar idea helped me defeat a human player and reduce my handicap by 3 stones. Not having the patience for thousands of 30-minute games to achieve statistically valid results, I settled for trouncing my opponent three games in a row by a large margin, then doing it again with a smaller handicap for three more games. I can't win by 70 stones in a 7 stone game, but 20 or 30 was enough to prove my point. If I made random plays under the assumption that I still had a half-point win, my opponent's predictive powers would be superior to mine - that's why he gives me a handicap and not vice versa. I can't really be sure that my prediction of a 22.5 point win is exact to the last decimal point - but if it should be within 5 or 10 or even 20, I'm perfectly happy. It's nice that statistics of a series of one-bit values are so useful, but when a significant fraction of those one-bit values are 100% wrong, that introduces a bit of noise to one's estimates. One hopes that they balance evenly, but perhaps they do not. Terry McIntyre terrymcint...@yahoo.com “We hang the petty thieves and appoint the great ones to public office.” -- Aesop From: Don Dailey dailey@gmail.com To: computer-go computer-go@computer-go.org Sent: Wednesday, August 19, 2009 6:03:50 AM Subject: Re: [computer-go] Dynamic komi at high handicaps One must decide if the goal is to improve the program or to improve it's playing behavior when it's in a dead won or dead lost positions. It's my belief that you can probably cannot improve the playing strength soley with komi manipulation, but at a slight decrease in playing strength you can probably improve the behavior, as measured by a willingness to fight for space that is technically not relevant to the goal of winning the game.And only then if this is done carefully. However I believe there are better ways, such a pre-ordering the moves. Even if this can prove to be a gain, you are really working very hard to find something that only kicks in when the game is already decided - how to play when the game is already won or already lost.But only the case when the game is lost is this very interesting from the standpoint of making the program stronger. And even this case is not THAT interesting, because if you are losing, on average you are losing to stronger players. So you are working hard on an algorithm to beat stronger players when you are in a dead lost game? How much sense does that make? So the only realistic pay-off here is how to salvage lost games against players that are relatively close in strength since you can expect not to be in this situation very often agaist really weak players.So you are hoping to bamboozle players who are not not weaker than you - in situations where you have been bamboozled (since you are losing, you are the one being outplayed.) That is why I believe that at best you are looking at only a very minor improvement.If I were working on this problem I would be focused only on the playing style, not the playing strength. If you want more than the most minor playing strength improvement out of this algorithm, you have to start using it BEFORE the loss is clear, but then you are no longer playing for the win when you lower your goals, you are playing for the loss. - Don 2009/8/19 Stefan Kaitschick stefan.kaitsch...@hamburg.de One last rumination on dynamic komi: The main objection against introducing dynamic komi is that it ignores the true goal of winning by half a point. The power of the win/loss step function as scoring function underscores the validity of this critique. And yet, the current behaviour of mc bots, when either leading or trailing by a large margin, resembles random play. The simple reason for this is that either every move is a win or every move is a loss. So from the perspective of securing a win, every move is as good as any other move. Humans know how to handle these situations. They try to catch up from behind, or try to play safely while defending enough of a winning margin. For a bot, there are
Re: [computer-go] Dynamic komi at high handicaps
On Wed, Aug 19, 2009 at 9:39 AM, Magnus Persson magnus.pers...@phmp.sewrote: Don, what you write is certainly true for even games, but I think the problem is a real one in high handicap games with the computer as white. I use a hack to make Valkyria continue playing the opening in handicap games as white. It is forbidden to resign in the opening and early middle game because it would if it could. In handicap games the situation is different. You have roughly even chances whether taking the handicap or giving it. I think this illustrates that fundamentally this is an opponent modeling issue.And I really like the idea that someone had of throwing in occasional pass moves for the player who is presumed weaker. There is an analogy in computer chess - it's called the null move heuristic. If it is white to move, you can measure the potential of blacks positions by playing a pass move for white (called the null move) and the a reduced depth search.Whatever score is returned can be considered a lower bound - since you as white skipped one of your moves. With Go it is a little different. If you are trying to beat a much stronger player but you have been given a nice advantage due to handicap, then the playouts will see you easily winning the game and you will play these random looking moves. However, if the computer throws in some pass moves for itself in the playouts, it will play more focused - it will be challenged to find strategies that work in the presense of it's own sloppy play. In other words the computer will stop this anything works attitude and it will focus on robust strategies that give it some room for error. It should be able to find these more robust strategies because it knows it is comfortably ahead. I don't know if this will actually work, it's only at the idea stage as far as I know - but it's something that seems more consistent with the actual problem.Komi manipulation changes the goal which is very dangerous but this ideas does not change the goal, just how it is achieved. To rephrase your argument for even games, the problem situation should never occur because the losing player *should out of courtesy* resign long before the evalutaion become so skewed. That's not correct, because with handicap games the premise is different. My reasoning is based on the well known fact that you will not often get outplayed by signficantly weaker opponents and you will not often outplay signficantly stronger opponents. But this does not apply to handicap games because nobody was outplayed - you started from a game that is a dead win for one side. In a handicap game, it's not only likely, it is CERTAIN that you will find youself in a dead won game against a much stronger opponent.In even games this is going to be a rare occurance. But this does not apply to h9 games on 19x19 for example. And if I am not mistaken strong heavy playouts evaluates such positions very pessimistically, and thus we have a problem to solve, which grows with increasing playing strength. Still stronger programs will discriminate between bad and good moves even with extreme scores, so I think the dimensions of this problem is exaggerated. Yes, it's a problem. And likewise with komi manipulation, the stronger the program is the more likely a small komi change will wildly change the score, from dead won to dead lost or visa versa. Imagine a program so strong that it always plays random moves when losing, and when winning it randomly plays any move that does not lose. It should be obvious that in a winning position, it is going to play a winning move with certainty, but if you adjust komi to make it play better it will play a random move - which could be a losing move. This thought experiment consistitues a kind of proof that the idea at it's most fundamental level is wrong. This can be salvaged by doing multiple searches with different komi's and only playing moves they have in common. I think this all gets complicated (and interesting) becasue we tend to think in two different ways about playing games, one way is all about correctness, finding the best move in the game theoretic sense and the other is how to improve your practical winning chances in the face of fallible opposition (such as blowing smoke in his face.) So it's rather hard to make any kind of proof that something like this is better or not better - it all has to be emprical. - Don -Magnus Quoting Don Dailey dailey@gmail.com: One must decide if the goal is to improve the program or to improve it's playing behavior when it's in a dead won or dead lost positions. It's my belief that you can probably cannot improve the playing strength soley with komi manipulation, but at a slight decrease in playing strength you can probably improve the behavior, as measured by a willingness to fight for space that is technically not relevant to the goal of
Re: [computer-go] Dynamic komi at high handicaps
On Wed, Aug 19, 2009 at 07:27:00AM -0700, terry mcintyre wrote: Consider the game when computer is black, with 7 stones against a very strong human opponent. Computer thinks every move is a winning move; it plays randomly; a half-point win is as good as a 70-point win. Didn't this game actually happen? Didn't MoGo *beat* a pro with 7 stones? Did it play randomly? Don't the monte carlo bots frequently win as White when giving handicap stones on KGS? I think we need some real statistical evidence that this problem is even worth talking about, aside from stylistic issues. I'm not the first to say this, but I think it bears repeating. -Jeff ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Dynamic komi at high handicaps
zen wins many more of its even games with no handicap than it does with even, say, an even 2 stone handicap as either black or white. i haven't compiled numbers for it (i'm not zen's maintainer), but i watched it happen over the course of about 50 games one day. it was pretty consistently worse with any kind of handicap on the board, the more handicap the worse. fix the handicap problem and it would likely rise a stone in strength. s. On Wed, Aug 19, 2009 at 12:15 PM, Jeff Nowakowskij...@dilacero.org wrote: On Wed, Aug 19, 2009 at 07:27:00AM -0700, terry mcintyre wrote: Consider the game when computer is black, with 7 stones against a very strong human opponent. Computer thinks every move is a winning move; it plays randomly; a half-point win is as good as a 70-point win. Didn't this game actually happen? Didn't MoGo *beat* a pro with 7 stones? Did it play randomly? Don't the monte carlo bots frequently win as White when giving handicap stones on KGS? I think we need some real statistical evidence that this problem is even worth talking about, aside from stylistic issues. I'm not the first to say this, but I think it bears repeating. -Jeff ___ 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] Dynamic komi at high handicaps
2009/8/19 terry mcintyre terrymcint...@yahoo.com Consider the game when computer is black, with 7 stones against a very strong human opponent. Computer thinks every move is a winning move; it plays randomly; a half-point win is as good as a 70-point win. Pro gains ground as computer makes slack moves, taking slightly less than its full due. At some point, the computer, being the weaker player, makes one slack move too many and loses the game. Rinse and repeat. All you are doing here is repeating the idealized scenario that illustrates the rationale for this idea. That is a fine way to describe how this has potential to be a solution to the problem, but it doesn't explain why it has not been made to work and it does not address the fact that your scenario is highly idealized - not all positions work that way with everything so cut and dried. I can do exactly the same thing and I have been, constructing scenarios that will fail with any heuristic you propose. But that does not prove or disprove anything. So I propose that we need to start thinking about why this has been so resistant to success and consider the possibility that it doesn't work, or that we are not properly addressing the reason why it doesn't work.And to do this YOU have to be the one that constructs counter-examples. At some point, it dawns on the programmer: must attack to win handicap games. Must be a little bit greedy, to slow down the process of attrition. The attack part I agree with, the greed I do not. Dynamic komi models something real: the significant advantage of the computer in a handicap game. It tries to preserve as much of that advantage as possible. I think the problem and what I consider your misconception is revealed here. You use the word advantage incorrectly. Grabbing up points on the board is a different concept than having or not having an advantage Your solution is to suddenly switch to an inferior definition of advantage, one that is clearly broken, otherwise we would be using point count instead of win count in our programs. So I don't see this at all as trying to preserve the advantage, I see it as giving it away. I really believe the secret has to be in the playouts - we use those to estimate our chances. If you change komi the playouts try to estmate the chances that you will win at that NEW KOMI, not at some other komi. I don't know if it will work for computer vs human games. I do know that a similar idea helped me defeat a human player and reduce my handicap by 3 stones. Not having the patience for thousands of 30-minute games to achieve statistically valid results, I settled for trouncing my opponent three games in a row by a large margin, then doing it again with a smaller handicap for three more games. I can't win by 70 stones in a 7 stone game, but 20 or 30 was enough to prove my point. If I made random plays under the assumption that I still had a half-point win, my opponent's predictive powers would be superior to mine - that's why he gives me a handicap and not vice versa. That trick helped you due to human psychology. Humans have a tendancy to rise to the occaions and computers do not know how to try harder like we do. In computer chess one of the strengths of the old programs was that when they were losing they did not become disheartened like human players often do. Once I win a pawn or two or a piece you can sometimes feel your oppoent resign even if he doesn't say the words right away. You also assume the computer program is being sloppy which could not be farther from the truth. If the comptuer plays a random looking move it's only because the move has no affect on the winning chances that are within the computers ability estimate.And the move is only random because you define it to be so. If you try the same thing, you are just being cocky and you are letting your guard down. Us humans must always be vigilant and keep our interest in the game as high as possible and the best way to do this is to continue to fight - what is called the follow through in baseball. In tennis doubles, after breaking the opponent serve we used to say, now that we have them down, let's kick them. You almost have to play like you are losing in order to win if you are human. But computers always play at full throttle. I can't really be sure that my prediction of a 22.5 point win is exact to the last decimal point - but if it should be within 5 or 10 or even 20, I'm perfectly happy. It's nice that statistics of a series of one-bit values are so useful, but when a significant fraction of those one-bit values are 100% wrong, that introduces a bit of noise to one's estimates. One hopes that they balance evenly, but perhaps they do not. Terry McIntyre terrymcint...@yahoo.com “We hang the petty thieves and appoint the great ones to public office.” -- Aesop -- *From:* Don Dailey
[computer-go] (no subject)
Jeff Nowakowski wrote: On Wed, Aug 19, 2009 at 07:27:00AM -0700, terry mcintyre wrote: Consider the game when computer is black, with 7 stones against a very strong human opponent. ... Didn't this game actually happen? Didn't MoGo *beat* a pro with 7 stones? It was long ago: in February 2009, and it was only the first game in a series of 6 games. All other five games in that event were won by the humans. Later, only one more bot-win against a low pro at h7. http://www.computer-go.info/h-c/index.html Without special techniques the h7 wall will stand for a long time. Ingo. PS: Once again I would like to mention my report on Laziness of Monte Carlo, at http://www.althofer.de/mc-laziness.pdf In the meantime, a student has found the same phenomenon in UCT search (instead of basic MC). Also in discrete online optimization (so outside of combinatorial games) it has been observed by another Ph.D. student of mine: porcedures on Monte Carlo basis are stronger when they have the impression that the situation is tense. -- Neu: GMX Doppel-FLAT mit Internet-Flatrate + Telefon-Flatrate für nur 19,99 Euro/mtl.!* http://portal.gmx.net/de/go/dsl02 ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] (no subject)
PS: Once again I would like to mention my report on Laziness of Monte Carlo, at http://www.althofer.de/mc-laziness.pdf In the meantime, a student has found the same phenomenon in UCT search (instead of basic MC). Also in discrete online optimization (so outside of combinatorial games) it has been observed by another Ph.D. student of mine: porcedures on Monte Carlo basis are stronger when they have the impression that the situation is tense. Laziness is something we all agree on. This is not in dispute. But how do you create the required tension in a way that produces a program that plays the game better? I don't mean selected positions, but the entire game. - Don -- Neu: GMX Doppel-FLAT mit Internet-Flatrate + Telefon-Flatrate für nur 19,99 Euro/mtl.!* http://portal.gmx.net/de/go/dsl02 ___ 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] Dynamic komi
Don wrote: But how do you create the required tension in a way that produces a program that plays the game better? At least in high handicap go on 19x19 (with the dynamic bot being the stronger player) it seems to work when the bot is kept in some 35-45 % corridor, as long as it is clearly behind. Ingo. -- GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT! Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01 ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Laziness
Speaking of laziness, I have been intending to post a study concerning capturing races, but I haven't gotten around to it. So is it surprising that MC is lazy, given that MC programmers are lazy? :-) Ingo's Double Step Race is a simplified model of capturing race. My model was more complex, and I solved it recursively rather than via simulations. I promise a full post at some point. For now, here are the overall conclusions: 1) It is possible to be an outright underdog in a race played out by an MC process even if you win by force under alternating play. 2) The longer the race, the closer to even it appears, even if it is lopsided in alternating play (e.g., 4 moves vs 7). 3) If a position features multiple races, then the chance that MC will play all correctly is very small. Please consider race here in a general sense: you must reach your goal before the opponent reaches his goal, where the goals are incompatible. Semeai is a special case. My conclusion is the same as Gian-Carlo Pascutto's: I am convinced that the phenomenon of laziness is real, and that it hurts practical strength. To this I would add that laziness is not just a problem in handicap games. We need to elevate the discussion about laziness beyond the question of how to win when given 7 stones. I could not care less. The problem is that we need so many stones. This comes down to the difference between alternating play and random play. It is fundamental to the whole framework that MC will give high scores to positions that are favorable in alternating play. Unfortunately, there are many dead losing situations that have a reasonable chance of working in random play. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Bulky nakade shapes (was: Mercy rule position)
Fuego has no trouble with the mercy rule here - I guess our threshold is high enough. However, it has no clue about how to play out the nakade shape. So it starts out with 57% wins for White, and it needs maybe 30K simulations until the search pushes it below 50%. Then the score keeps dropping continually. It is actually a nice example of how search can fix bad playouts in Fuego. Fuego has an effective rule for stretched nakade shapes, such as 3 in a row, T or +. It simply moves the single-stone selfataries to the adjacent point. However, it cannot handle the bulky shapes such as in Brian's example - it just plays the first move randomly, then usually a Mogo-style pattern matches, which makes it likely that two eyes will be created. How do others handle such cases in the playouts? Martin ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Bulky nakade shapes
Following Mogo, Pebbles uses the 3-in-a-row modification that automatically plays in the center of 3-point eyeshapes. Mogo's rule guarantees that an opponent will not be able to convert a 3-point eyespace into two eyes. The downside of Mogo's rule is that it wastes a *lot* of moves when it generates such plays where the opponent strings have other eyes. Following Fuego, Pebbles has the upgrade self atari rule that moves attacker plays to the center of 3-point eyes. Fuego's clump correction rule upgrades the defender's plays. Fuego's rules triples the chance of making a correct play when a 3-point eyespace exists, but does not guarantee that any play will be made. The rules do guarantee that the best move will be made *within* a 3-point eyespace. This allows RAVE to accelerate discovery of life-giving shape. The only downside that I have noticed is that straight-four eyespaces are more likely to die in the playouts, because attacker moves are upgraded to center points, and pattern replies are not guaranteed to be made on the vital point. When Mogo's rule was first implemented, it made a huge increase in strength. But when Fuego's rules went in, then Mogo's rule started being less important, and now it has even become a disadvantage. My impression is that Fuego's rules covers a lot of situations, and the upgraded move is almost always an improvement. Perhaps Mogo's rule would be beneficial if the number of wasted moves could be cut down. I have not looked into that. I have implemented a rule that plays on the vital point of nakade shapes after captures. That rule has never made much effect on playing strength, but it looks great in particular cases. It should work out to be a low-cost positive. Note that all Pebbles testing is on 9x9. I expect situational rules to be relatively more important on larger boards. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Laziness
On Wed, Aug 19, 2009 at 2:11 PM, Brian Sheppard sheppar...@aol.com wrote: My conclusion is the same as Gian-Carlo Pascutto's: I am convinced that the phenomenon of laziness is real, and that it hurts practical strength. Unfortunately this is not that point that is in question - I think we all agree on this. - Don ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Bulky nakade shapes
Following Mogo, Pebbles uses the 3-in-a-row modification that automatically plays in the center of 3-point eyeshapes. Fuego's rules triples the chance of making a correct play when a 3-point eyespace exists, but does not guarantee that any play will be made. The rules do guarantee that the best move will be made *within* a 3-point eyespace. This allows RAVE to accelerate discovery of life-giving shape. The only downside that I have noticed is that straight-four eyespaces are more likely to die in the playouts, because attacker moves are upgraded to center points, and pattern replies are not guaranteed to be made on the vital point. Interesting. I looked at some playouts from your sample position, and I also saw similar problems, e.g. when O's eye space becomes X.. then O will live with high likelihood in Fuego playouts, since O will soon capture the X stone by the capturing rule, but X has no rule to play in the center and kill. When Mogo's rule was first implemented, it made a huge increase in strength. But when Fuego's rules went in, then Mogo's rule started being less important, and now it has even become a disadvantage. Markus implemented Mogo's nakade rule after we had ours, and could not get it to work. So it remains switched off in Fuego by default. It is very interesting to me that you use the clump correction rule. I could never get that to work in Fuego, either. My impression is that Fuego's rules covers a lot of situations, and the upgraded move is almost always an improvement. Perhaps Mogo's rule would be beneficial if the number of wasted moves could be cut down. I have not looked into that. I have implemented a rule that plays on the vital point of nakade shapes after captures. That rule has never made much effect on playing strength, but it looks great in particular cases. It should work out to be a low-cost positive. Yes, I was thinking of trying that. Martin ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Dynamic komi
I'm curious to find out what is meant by lazy. If, as I am led to believe by your report, Monte Carlo strategies applied to Double Step Races are lazy, yet they converge to perfect play, then I'm not sure why we are meant to worry. I certainly understand that the strategies can converge faster in some cases than in others. I would expect that this could be used to one's advantage, by using Monte Carlo to evaluate a related game, that has similar correct moves, but for which the Monte Carlo evaluation is more accurate. But it isn't clear how one is supposed to transform games like his 6-vs-6 into 6-vs-5, without already knowing enough to completely solve the game! Specifically, is there a way to do this that doesn't _also_ convert 6-vs-5 into 6-vs-4? Switching back to go, I would like to point out that it seems to me to be a serious mistake to apply UCT to any game other than the one at hand. If you do some dynamic fiddling with komi, you really ought to only do it in playouts, and you should ensure that as the length of the playout decreases, the amount of fiddling approaches zero. That way, once UCT has expanded a deep line of play, it will be working with accurate win/loss values, instead of some fantasy based on a globally-applied dynamic komi. This seems very much in line with what others have already suggested regarding the insertion of passes or other ways of degrading the play in the playouts. I suspect that adjusting komi in the above manner may be an even better solution. Weston On Wed, Aug 19, 2009 at 1:45 PM, Ingo Althöfer3-hirn-ver...@gmx.de wrote: Don wrote: But how do you create the required tension in a way that produces a program that plays the game better? At least in high handicap go on 19x19 (with the dynamic bot being the stronger player) it seems to work when the bot is kept in some 35-45 % corridor, as long as it is clearly behind. Ingo. -- GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT! Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01 ___ 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/
