"entropi","bent four in the corner","2009-01-16 01:00:40","Hi, \ It may be a novice question but I could not find a satisfying answer anywhere. \ \ First what I know about bent four in the corner: \ The rule \"bent four in the corner is dead\", refers to situations where in fact the shape \"bent four in the corner\" itself has not yet occured. The defending player cannot do anything locally. Therefore the attacking player does not hurry and removes all ko threats and fills the outside liberties before starting the actual fight in the corner. Since the defender has no more ko threats left on the board, the attacker will win. Therefore, we can say that the bent for in the corner is doomed to die. \ \ [U]Now my first questions is:[/U] \ Since removing all of the ko threats will cost the attacker an unforeseeable number of points (at least under japanese rules), and the defender may \"pass\" each time (i.e. no damage to his own territory), why do the rules give the benefit to the attacker? \ \ After all a \"ko threat\" itself is also not a well defined concept. A move that is perceived as a ko threat by my 5kyu eyes, may be ignorable by a dan. \ \ I don't mention the unremovable ko threats (like a seki for example). \ \ [U]And my second questions is:[/U] \ There may be positions where the final [B]shape[/B] is not \"bent four in the corner\" but the [B]situation[/B] is exactly the same. The attacker waits until the end, then removes all the ko threats and attacks. The defender is then helpless. Is the rule still applicable? \ \ \ Sorry, if the questions are already answered somewhere in forums. I could not find a simple and satisfying answer and I am confused about that basic thing. \ \ Thanks in advance." "sol.ch","","2009-01-16 01:11:47","[QUOTE]Since removing all of the ko threats will cost the attacker an unforeseeable number of points (at least under japanese rules), and the defender may \"pass\" each time (i.e. no damage to his own territory), why do the rules give the benefit to the attacker?[/QUOTE]\ \ Under Japanese rule, bent 4 is just considered dead (except for extreme circumstances; see [URL=warp.povusers.org/go/UndeadBentFours/analysis.html]here[/URL]). It is under the Chinese rules where you have to remove all ko threats and ensure you kill it \"manually\".\ \ [QUOTE]There may be positions where the final shape is not \"bent four in the corner\" but the situation is exactly the same. The attacker waits until the end, then removes all the ko threats and attacks. The defender is then helpless. Is the rule still applicable?[/QUOTE]\ \ I'm not quite sure what you mean by this question." "entropi","","2009-01-16 02:07:13","[QUOTE=sol.ch]Under Japanese rule, bent 4 is just considered dead (except for extreme circumstances; see [URL=warp.povusers.org/go/UndeadBentFours/analysis.html]here[/URL]). It is under the Chinese rules where you have to remove all ko threats and ensure you kill it \"manually\".[/QUOTE] \ \ This is exactly what I am confused about. The japanese rules say to the attacker \"[I]you don't have to remove all of the ko threats[/I]\" which probably saves him a good number of points. Without this rule, the attacker would in fact have to remove the ko threats before capturing. Isn't it unfair? \ \ Let alone the situation where there are non-removable ko threats, like some seki situation. \ \ \ [QUOTE]I'm not quite sure what you mean by this question.[/QUOTE] \ \ In one of my recent games there was a similar situation where the shape in question was neither a \"bent four\" nor in the corner. I could not make a reasonable local move. My opponent could capture a stone in a ko situation and put my whole group into atari. My only chance to survive was winning the ko. But of course my opponent did not take the ko immediately and waited until the end of the game (which did never come because I resigned at some point). \ \ It was not \"bent four in the corner\" because it was not even in the corner :). But the principle was exactly the same. Should we have considered my group dead, or should my opponent have removed all ko threats \"manually\" first, if we were to play until the end?" "flOvermind","","2009-01-16 02:44:35","I think you're confusing something here. There is no explicit rule that says \"bent four is dead\". \ \ But: \ \ Under Japanese rules, life and death disputes are settled in a separate phase after the two passes. The rules state that in this phase, the only valid ko threat is a pass. Basically, you are playing it out \"hypothetically\", and after having decided the status of the groups, you restore the original position, so none of the point-losing moves will matter. \ \ So the procedure for killing the bent four does not involve removing any ko threat. You just pass. When the opponent disagrees with your claim that the group is dead, you kill it in hypothetical play. \ \ \ Most of the time, the \"pass for ko threat\" rule implies that bent four is dead. But that's not always the case. Consider this situation: \ \ [go]$$ Seki. \ $$ +-------------+ \ $$ | X X O . X . | \ $$ | . X O O O X | \ $$ | O X X X O . | \ $$ | . O . X O O | \ $$ +-------------+ \ [/go]" "xela","","2009-01-16 03:24:37","[QUOTE=flOvermind]I think you're confusing something here. There is no explicit rule that says \"bent four is dead\".[/QUOTE] \ \ I believe there used to be such a rule in the old Japanese rules (before the 1989 version of the rules was published), and so a lot of old books still explain it this way. And I think the current World Amateur rules do the same thing." "Tommie","coffee <= 15:45 ?","2009-01-16 04:19:08","\"Entropi\", \ \ show me your imagined other positions over a coffee \ and we discuss them. Was it on a real board or online? \ \ What you describe, sounds like a stage(ed)-ko, e.g. \ [url]http://senseis.xmp.net/?TwoStageKoExamples[/url] or \ better this: \ [url]http://senseis.xmp.net/#diag3[/url] ? \ How big was the group, how big where the moves elsewhere , who was behind at that stage, etc.? \ \ Sol.Ch's link explains (un-)dead bent-4 very well." "entropi","","2009-01-16 05:04:00","[QUOTE=flOvermind]I think you're confusing something here. There is no explicit rule that says \"bent four is dead\". \ \ But: \ \ Under Japanese rules, life and death disputes are settled in a separate phase after the two passes. The rules state that in this phase, the only valid ko threat is a pass. [/QUOTE] \ \ Thanks! That exactly answers the second question, which was silly anyway. Of course a generic rule should not refer to any specific shape :o \ \ But the answer to the first question is still not very clear to me. I understand perfectly what you mean. The ko threats are not removed actually and everything happens \"hypothetically\" as in a dispute case. \ \ But my question was, isn't the rule (after two passes in dispute case the only ko threat being pass) unfair against the defender of the bent-four group? Or put in another way, why is there such a rule? \ \ [QUOTE=Tommie] \ show me your imagined other positions over a coffee \ and we discuss them. Was it on a real board or online?[/QUOTE] \ \ Tommie, It was online but I dont have the sgf file and it's very difficult to regenerate. \ In fact it would be very nice because it is highly possible that we were missing something in such a complicated position. I will try to remember it for discussing." "Tommie","","2009-01-16 06:38:41","[QUOTE=entropi]... It was online but I dont have the sgf file and it's very difficult to regenerate.[/QUOTE] \ \ Which server?" "entropi","","2009-01-16 06:50:08","[QUOTE=Tommie]Which server?[/QUOTE] \ igs" "Kirby","","2009-01-16 06:52:48","[quote=entropi]igs[/quote]\ \ I haven't used IGS in awhile, but if I recall you can access your game archives here:\ [url]http://my.pandanet.co.jp/index-e.htm[/url]" "shapenaji","","2009-01-16 07:19:00","bent-4 is really one of those situations that makes chinese/AGA rules make a great deal more sense. \ \ by Chinese rules, removing removable ko threats costs a player nothing. And, as a result, the rule is to play it out. If there's an unremoveable ko threat... well, its seki, and if you play it out, and you don't realize that its seki, you'll take a loss elsewhere on the board. \ \ Artificial rules for \"special\" positions are silly." "Bill Spight","","2009-01-16 11:07:02","[QUOTE=entropi]But the answer to the first question is still not very clear to me. I understand perfectly what you mean. The ko threats are not removed actually and everything happens \"hypothetically\" as in a dispute case.\ \ But my question was, isn't the rule (after two passes in dispute case the only ko threat being pass) unfair against the defender of the bent-four group? Or put in another way, why is there such a rule?[/QUOTE]\ \ The reason is lost in the mists of time. Japanese rules have apparent quirks which nobody knows the origin of. (Such as the rule that there are no points in seki.)\ \ Consider an ordinary dead group inside your territory. At the end of the game you do not have to make the moves that are necessary to capture them. Is that unfair to the player with the dead group? You might think so, but those are the rules. Different rules would make a different game. (Note: There is no penalty for capturing dead stones under area rules, so this rule makes the Japanese rules more like area rules than a rule that said that all stones left at the end of the game are alive. Are area rules unfair?)\ \ The same applies to removing ko threats in the case of a bent four. Removing those threats is part of capturing dead stones, and so those plays are not needed. There is nothing special about bent four in this regard. Dead stones are simply removed without being captured. (Note that the result is the same under area rules, except in the case of unremovable ko threats, and perhaps other rare situations.)\ \ As for the pass for ko rule, that is an innovation in the Japanese 1989 rules. It is a clever way of decoupling kos from threats. Whether it is too clever is another question." "entropi","","2009-01-17 10:54:08","[QUOTE=Bill Spight]\ Consider an ordinary dead group inside your territory. At the end of the game you do not have to make the moves that are necessary to capture them. Is that unfair to the player with the dead group? You might think so, but those are the rules. Different rules would make a different game. (Note: There is no penalty for capturing dead stones under area rules, so this rule makes the Japanese rules more like area rules than a rule that said that all stones left at the end of the game are alive. Are area rules unfair?)\ [/QUOTE]\ \ \ But there is an important difference between\ - \"you dont have to capture dead stones\" and\ - \"you don't have to remove all ko threats\".\ \ The difference is, in case of a doubt whether a group of stones is dead or not, there is the dispute procedure for resolving it. This means that whatever the result is, the players are fully responsible of it. The rules do not favor any of the players.\ \ However, in case of removing \"all\" ko threats there is no procedure for deciding whether a move is a real ko threat or not.\ \ I mean, consider the situation where my opponent has bent four in the corner and I have the right to claim it dead according to the japanese rules. If I had to remove all ko threats (if there were no such rule), the number of my \"self-damaging moves\" would depend on my judgment of whether a move is a real ko threat or not. This might sound like a tiny detail but I think it is principally important.\ \ Anyway, in fact you are right the rules are not unfair, or at least the term \"unfair\" is not appropriate, because the same rules apply to both players. I just wanted to learn the logic behind this rule." "Bill Spight","","2009-01-17 12:52:59","[QUOTE=entropi]But there is an important difference between \ - \"you dont have to capture dead stones\" and \ - \"you don't have to remove all ko threats\".[/QUOTE] \ \ If you can make a position a ko but your opponent cannot, removing the opponent's ko threats is a part of capturing the opponent's stones in that position. \ \ [QUOTE]The difference is, in case of a doubt whether a group of stones is dead or not, there is the dispute procedure for resolving it. This means that whatever the result is, the players are fully responsible of it. The rules do not favor any of the players. \ \ However, in case of removing \"all\" ko threats there is no procedure for deciding whether a move is a real ko threat or not.[/QUOTE] \ \ So? The players are still fully responsible for the result. \ \ [QUOTE]I mean, consider the situation where my opponent has bent four in the corner and I have the right to claim it dead according to the japanese rules. If I had to remove all ko threats (if there were no such rule), the number of my \"self-damaging moves\" would depend on my judgment of whether a move is a real ko threat or not. This might sound like a tiny detail but I think it is principally important.[/QUOTE] \ \ Under Japanese rules you do not have to remove the ko threats, so that point is moot. That fact that in nearly all cases the results are the same, regardless of the form of scoring, is a happy circumstance. (I do not say happy accident or coincidence, because I do not think that it is coincidental. :)) \ \ [QUOTE]Anyway, in fact you are right the rules are not unfair, or at least the term \"unfair\" is not appropriate, because the same rules apply to both players. I just wanted to learn the logic behind this rule.[/QUOTE] \ \ The logic of the Japanese '89 rules is, IMO, a bit strained. In the Japanese '49 rules there was no logic at all, just tradition. We can speculate about the reasons for the tradition, but nobody really knows what they were." "HermanHiddema","","2009-01-17 15:07:11","The exact evolution of the rules is unknown. So the following is pure speculation. All of this might have happened at an early stage, somewhere in the mists of time, before such changes were recorded. Or it might now have. But it is how it might have happened that Japanese rules are the way they are:\ \ When Go was first invented, the rules were \"stone counting\", where only the number of stones you have on the board counts toward your score. This was quickly found to be tedious, because at the end of the game, both players have to play some 50 extra moves to fill their territory.\ \ So a way of speeding it up was devised: Count the stones on the board, and also the territory that is surrounded, but substract two points for each group (because you couldn't fill those, you need two eyes). Also, count no points in seki, because you can't fill without dying. (This is now called \"group tax\")\ \ This, however still means you have to count every point on the board. An even quicker way to count was thus devised: Since the players have the same number of stones on the board (minus prisoners), lets just count the territory (minus prisoners). Now, we only need to count about 50 points each, and we can arrange the territory in easy shapes to count. (also, lets forget about that 2 point penalty for each group for each group, but still not give points in seki)\ \ So now you have something approaching Japanese rules.\ \ The problem with counting territory, of course, is that you can no longer capture dead stones inside your territory without losing points. So rules came into effect that you didn't have to, that group that were dead were taken of the board wihout playing. Similarly, bent four became a problem, because removing the ko threats now costs points. So again, the rules became such that there were no ko threats after the end of the game, and bent four was dead.\ \ So all in all, it may have been a trade-off. Where making most games were much easier to count was deemed more important than the rare game where territory rules caused trouble.\ \ (Note: where I say \"rules\" above, this does not actually refer to some official rules document, but more likely to an oral tradition of explaining the rules)" "sol.ch","","2009-01-17 15:53:17","So far based on the responses I've read, it seems that the Chinese rules are simply superior to Japanese. I will now be playing with that ruleset whenever possible so I don't have a headache when dealing with these special cases :)." "HermanHiddema","","2009-01-17 15:55:38","[QUOTE=sol.ch]So far based on the responses I've read, it seems that the Chinese rules are simply superior to Japanese. I will now be playing with that ruleset whenever possible so I don't have a headache when dealing with these special cases :).[/QUOTE] \ \ Have you ever done Chinese style counting? :)" "sol.ch","","2009-01-17 15:58:45","[QUOTE=HermanHiddema]Have you ever done Chinese style counting? :)[/QUOTE] \ \ Yes, and while it does take longer to do, it isn't any more difficult than Japanese style counting. \ \ And besides, I play mostly online anyways, so it's not a big deal for me :)." "HermanHiddema","","2009-01-17 16:03:28","[QUOTE=sol.ch]Yes, and while it does take longer to do, it isn't any more difficult than Japanese style counting. \ \ And besides, I play mostly online anyways, so it's not a big deal for me :).[/QUOTE] \ \ Yup, online it is no problem. And AGA style area scoring through territory counting is also a nice solution. \ \ Then, of course, there is the granularity issue, which I think is a plus for Japanese rules :)" "Bill Spight","","2009-01-17 16:20:49","[QUOTE=HermanHiddema]Yup, online it is no problem. And AGA style area scoring through territory counting is also a nice solution. \ \ Then, of course, there is the granularity issue, which I think is a plus for Japanese rules :)[/QUOTE] \ \ Which is one reason I like Button Go. :)" "Hicham","","2009-01-18 02:59:29","[quote=sol.ch]So far based on the responses I've read, it seems that the Chinese rules are simply superior to Japanese. I will now be playing with that ruleset whenever possible so I don't have a headache when dealing with these special cases :).[/quote] \ Hear, hear! I have been playing with chinese rules online for three years now. On KGS you do have to remind most people to fill dame. Once in a while you get an idiot who calls you a cheater for playing non-Japanese rules or people who just resign when you remind them the rules are chinese... But these are the exceptions luckily.\ \ I think I should go all the way and play with New Zealand rules which seem even better/clearer/more logical to me, but so far I have been put off by the need to explain the rules every time I play." "HermanHiddema","","2009-01-18 03:39:27","[QUOTE=Bill Spight]Which is one reason I like Button Go. :)[/QUOTE] \ \ Yes! Button Go is very good :)" "sol.ch","","2009-01-18 12:59:47","[QUOTE=HermanHiddema]Yes! Button Go is very good :)[/QUOTE]\ \ I have just read the article on senseis ([url]http://senseis.xmp.net/?ButtonGo)[/url], and it is a very interesting ruleset (I'm rather perplexed I hadn't run into it before). \ \ Has there been any tournaments (even local ones?) that has tried to use this ruleset? I think it would be an interesting experiment and it's certainly worth trying out if there hasn't been any." "Harleqin","","2009-01-18 22:57:59","[QUOTE=sol.ch] \ Has there been any tournaments (even local ones?) that has tried to use this ruleset? I think it would be an interesting experiment and it's certainly worth trying out if there hasn't been any.[/QUOTE] \ \ The World Mind Sports Games last year actually used a variant of that. Sadly, the original concept had been ... modified ... by some ... Go rules politics. Still, it is a very promising start." "HermanHiddema","","2009-01-19 02:57:37","[QUOTE=Harleqin]The World Mind Sports Games last year actually used a variant of that. Sadly, the original concept had been ... modified ... by some ... Go rules politics. Still, it is a very promising start.[/QUOTE] \ \ But in effect, the WMSG are button go rules, right? \ \ WMSG rules state: \ Komi is 6.5 \ If white passes first, black's score is reduced by 1 \ \ Which gives the same result as: \ Komi is 6.5 \ If white passes first, white's score is increased by 1 \ \ Which gives the same result as: \ Komi = 6.5 if black passes first \ Komi = 7.5 if white passes first \ \ Which gives the same result as: \ Komi = 7 \ The first player to pass gets half a point added to their score." "RobertJasiek","","2009-01-19 03:30:37","Hicham, I have been using NZ Rules on KGS for years now. \ \ ca. 95% no problems \ ca. 5% opponent does not know NZ Rules and I fill all or most 2-sided dame \ ca. 0% opponent resigns, escapes, call me cheater, etc. (some of them thereafter never play me again, especially the cheater cryers) \ \ Trying to hint to the opponent that NZ Rules are being using is no good idea though. Most feel insulted being told at all. They either know the rules anyway or rather prefer not being told. \ \ A few years ago, ca. 0% asked for the rules differences before 3rd move and I explained basic strategic effect. Noadays everybody just knows NZ Rules and they don't need to ask (unless they belong to 5% mentioned above). \ \ So overall, apart of complete rules (and strategic consequneces) ignorance by a few, there are no problems. \ \ The problem is rather KGS with its Japanese Rules autosetting default, which currently cannot be changed when accepting another player's game request." "HermanHiddema","","2009-01-19 03:38:25","[QUOTE=RobertJasiek] \ ca. 5% opponent does not know NZ Rules and I fill all or most 2-sided dame \ [/QUOTE] \ \ Any estimate on how many of these [i]have not noticed that the rules are NZ[/i]? (even though they might know them)." "RobertJasiek","","2009-01-19 03:41:40","Everything about rules and also Japanese rules and how to apply them: \ \ [url]http://home.snafu.de/jasiek/rules.html[/url] \ \ How bent-4 is treated depends on WHICH Japanese ruleset is being used. However, removal of ko threats is not a good explanation for several reasons, among which is: Some ko threats cannot be removed. \ \ Pre-1989 Japanese rules did not explain \"shapes\" with behaviour similar to bent-4. 1989 Japanese rules explain at least most practically occurring cases, But only my model 2003 Japanese Rules explain the thing for all positions in general. And then there are yet newer models that approach life definition completely afresh and so also simplify bent-4-like hidden kos: See e.g. the Simplified Japanese Rules." "RobertJasiek","","2009-01-19 03:47:23","Of the ca. 5% opponents not knowing / noticing(?) NZ Rules and I fill all or most 2-sided dame, maybe 20% just did not notice in time and 80% do not seem to be aware (or do not want to know) that rules other than Japanese / Korean exist at all (presumably those are players from that country). But the figures are wild guesses; I might be off by +-20%." "Hicham","","2009-01-19 07:12:18","As said I play with Chinese rules when not using automatch. \ \ When late endgame approaches, I tell my opponents 'Just a reminder, Chinese rules. Some don't notice\". I do this to not feel guilty when they pass instead of filling dame. \ \ Still I get occasianal problems with Japanese players who refuse (don't understand why) to fill dame. I tell them (if necessary in Japanese), and even offer undo. I go out of my way to keep it fair, but sometimes people refuse to cooperate and just give me points. I don't feel guilty anymore as people at my level (1k)should know better or at least pay attention to the rulesset they accept.\ Sometimes people ask me to explain the difference to Japanese rules. This is a good scenario for me, so I can spread the knowledge of superior rules (IMO of course:D).\ \ Maybe I will make the leap to NZ rules. I like the fact that suicide is allowed. I never understood why this was forbidden." "Zahlman","","2009-01-19 10:40:22","[QUOTE=entropi] \ [U]Now my first questions is:[/U] \ Since removing all of the ko threats will cost the attacker an unforeseeable number of points (at least under japanese rules), and the defender may \"pass\" each time (i.e. no damage to his own territory), why do the rules give the benefit to the attacker?[/quote] \ \ Removing the ko threat [b]would[/b] require a play in own territory, but under Chinese rules that would not cost a point (it would simply fail to take a point, if any dame remain). The Japanese rules don't require the attacker to lose points this way, precisely in order to keep the score the same as it would be with area counting. \ \ This only becomes arguably unfair in the situation where some ko threat [b]can't[/b] be removed, or e.g. if there are multiple such situations on the board, etc. \ \ [quote]After all a \"ko threat\" itself is also not a well defined concept. A move that is perceived as a ko threat by my 5kyu eyes, may be ignorable by a dan.[/quote] \ \ It's a very well-defined concept. At least, well-enough defined for our purposes. Under Chinese rules, you have all the time you need to cover even \"too small\" ko threats (since it would still be wrong to offer the threat's value in compensation when you don't have to; also see [url]http://senseis.xmp.net/?TertiaryKoThreat[/url] ), without losing points. \ \ The Japanese rules are meant to give the same results as the Chinese, under \"normal\" situations, while avoiding dragging things out. This is in keeping with the general philosophy of the rules (and see also Zen Buddhist ideas about the value of \"emptiness\" and the avoidance of \"useless things\"); which is presumably also why territory scoring is preferred (notwithstanding that the professionals have discovered that you really need to fill the dame *anyway* to avoid disputes). \ \ As for estimating ko threats and fighting ko in general, just try some exercises. Sol.ch recently posted a good one in the Game Analysis forum. \ \ [quote]I don't mention the unremovable ko threats (like a seki for example).[/quote] \ \ Of course, using a seki as a ko threat offers to lose points. But it can be correct play under such situations in the Chinese rules. ;) \ \ [quote][U]And my second questions is:[/U] \ There may be positions where the final [B]shape[/B] is not \"bent four in the corner\" but the [B]situation[/B] is exactly the same. The attacker waits until the end, then removes all the ko threats and attacks. The defender is then helpless. Is the rule still applicable?[/quote] \ \ Such positions are extremely rare (IIRC one is commonly referred to as \"Nakamura's position\"; that it could be named after a player - and a strong amateur at that - should give some idea of how strange it is for this to come up), but AFAIK, the jury is still out on them. :)" "RobertJasiek","","2009-01-19 11:44:27","Not that rare. Every semeai with a hidden ko does it and you can have it by a simple recapture that the opponent then needs to approach as a ko." "HermanHiddema","","2009-01-19 12:03:50","[QUOTE=RobertJasiek]Not that rare. Every semeai with a hidden ko does it and you can have it by a simple recapture that the opponent then needs to approach as a ko.[/QUOTE] \ \ I don't think this applies. The question is about positions which one player can leave until the end of the game, and which he can then win because the only valid ko-threat after game end is a pass. So position where there is a potential ko, but only one player can start it, [i]and[/i] he is the first to capture the ko. Any position that becomes ko during the game does not apply." "RobertJasiek","","2009-01-19 14:39:09","If creating diagrams here were easier, I would give an example. \ \ If rules analyse the status of each string separately, bent-4 does not need to rely on pass being the only ko threat. Almost all shapes don't need. Forget about the 1989 pass-for-ko-rule; it's only necessiety is for about just 2 precedents that do behave differently with a pass-for-ko-rule compared to a basic ko rule applied also during the hypothetical analysis. a) triple ko with external ko mouth, b) double-ko death + direct fighting ko elsewhere. The generality of the pass-for-ko-rule(s) is just pretence to hide these precedents. Everybody swallowed the trick until I revealed it with J2003." "HermanHiddema","","2009-01-19 14:45:50","[QUOTE=RobertJasiek]If creating diagrams here were easier, I would give an example.\ \ If rules analyse the status of each string separately, bent-4 does not need to rely on pass being the only ko threat. Almost all shapes don't need. Forget about the 1989 pass-for-ko-rule; it's only necessiety is for about just 2 precedents that do behave differently with a pass-for-ko-rule compared to a basic ko rule applied also during the hypothetical analysis. a) triple ko with external ko mouth, b) double-ko death + direct fighting ko elsewhere. The generality of the pass-for-ko-rule(s) is just pretence to hide these precedents. Everybody swallowed the trick until I revealed it with J2003.[/QUOTE]\ \ Creating diagrams here is no harder than it is on sensei's library, and is pretty easy I would say.\ \ You are claiming that situations like bent four, where under Chinese (and similar) rules one player can almost always remove all the ko-threats, but under Japanese he can not (hence this pass-for-ko rule), are not rare?" "Harleqin","","2009-01-19 15:10:22","I have personally witnessed a situation with a hidden ko that was not a bent-four come up in a casual game. Only one player could start it, and he would have been the first to capture. None of the other players present (several dan players with years of experience) could apply the rules that were supposed to be in effect (\"usual japanese rules\")." "HermanHiddema","","2009-01-19 15:26:03","[QUOTE=Harleqin]I have personally witnessed a situation with a hidden ko that was not a bent-four come up in a casual game. Only one player could start it, and he would have been the first to capture. None of the other players present (several dan players with years of experience) could apply the rules that were supposed to be in effect (\"usual japanese rules\").[/QUOTE] \ \ Yes, they exist, eg here's one: \ \ \ [go]$$W position at the end of the game \ $$ ------------------ \ $$ | . X . X . X O X . \ $$ | O O X X X X O X . \ $$ | . O O O O O O X . \ $$ | X X O X X X X X . \ $$ | X O O X . . . . . \ $$ | O O X . . . . . . \ $$ | X X . X . . . . . \ $$ | . . . . . . . . . \ $$ | . . . . . . . . . \ [/go] \ \ \ [go]$$W W1 captures, B2 plays the vital point \ $$ ------------------ \ $$ | 3 X . X . X O X . \ $$ | O O X X X X O X . \ $$ | 1 O O O O O O X . \ $$ | 2 . O X X X X X . \ $$ | . O O X . . . . . \ $$ | O O X . . . . . . \ $$ | X X . X . . . . . \ $$ | . . . . . . . . . \ $$ | . . . . . . . . . \ [/go]" "RobertJasiek","","2009-01-19 15:41:21","HermanHiddema, I have referred to frequency of hidden kos not of hidden kos + all ko threats removable." "HermanHiddema","","2009-01-19 15:49:10","[QUOTE=RobertJasiek]HermanHiddema, I have referred to frequency of hidden kos not of hidden kos + all ko threats removable.[/QUOTE] \ \ Well, the orignal question by entropi was: \ [QUOTE=entropi]And my second questions is: \ There may be positions where the final shape is not \"bent four in the corner\" but the situation is exactly the same. The attacker waits until the end, then removes all the ko threats and attacks. The defender is then helpless. Is the rule still applicable? \ [/QUOTE] \ \ And the response by Zahlman was: \ [QUOTE=Zahlman] \ Such positions are extremely rare (IIRC one is commonly referred to as \"Nakamura's position\"; that it could be named after a player - and a strong amateur at that - should give some idea of how strange it is for this to come up), but AFAIK, the jury is still out on them. \ [/QUOTE] \ \ To which you replied: \ [QUOTE=RobertJasiek] \ Not that rare. Every semeai with a hidden ko does it and you can have it by a simple recapture that the opponent then needs to approach as a ko.[/QUOTE] \ \ It seems quite clear to me that the original question refers to situations where all ko threats are removable. \ \ So as I read it, your original reply did claim that such positions were rare (So maybe you misread the original question?)" "RobertJasiek","","2009-01-19 23:04:35","Rather I read Zahlman's statement as referring to hidden kos but not to all ko threats removable. Whatever. My estimate implies that also the more specific all ko threats removable is not that rare because all ko threats removable is much more likely than not all ko threats removable." "HermanHiddema","","2009-01-20 01:02:45","[QUOTE=RobertJasiek]Rather I read Zahlman's statement as referring to hidden kos but not to all ko threats removable. Whatever. My estimate implies that also the more specific all ko threats removable is not that rare because all ko threats removable is much more likely than not all ko threats removable.[/QUOTE] \ \ Yes, I think hidden ko with [i]not[/i] all ko threats removable is [i]extremely[/i] rare. \ \ It all depend on your definition of [i]rare[/i], of course. But I, for one, have in my entire playing career not ever had a hidden ko other than bent four on the board. That spans at least 1000 serious games (discounting the thousands if quick/lightning games). \ \ Harleqin, above, mentions that he has seen one, but not in his own game. (I haven't even seen one happen at any tournament I have visited) \ \ \ Have you ever had a hidden ko other than bent four come up in one of your games? \ \ I would be surprised if hidden ko other than bent four came up in much more than 1 in 10,000 games (Perhaps we should call it the \"ten-thousand game ko\" ;))" "RobertJasiek","","2009-01-20 01:25:17","Unremovable ko threats are pretty frequent: Self-atari in sekis. But less frequent than no unremovable ko threats at all, OC. \ \ I have semeais with hidden kos in about - guessed very roughly - every 10th game. Usually it is a capturing-2-stones, ko, or pre-ko shape that the opponent then - all in hypothetical play - needs to approach as a ko because of an eye or other kos of the defender in the same semeai. \ \ This ordinary case has been neglected by most because the Japanese rules made such an extraordinary fuss about bent-4. The bent-4-related rules were bad as exceptions / precedents - but they were even much worse as misleading teaching." "HermanHiddema","","2009-01-20 02:18:14","[QUOTE=RobertJasiek]Unremovable ko threats are pretty frequent: Self-atari in sekis. But less frequent than no unremovable ko threats at all, OC. \ \ I have semeais with hidden kos in about - guessed very roughly - every 10th game. Usually it is a capturing-2-stones, ko, or pre-ko shape that the opponent then - all in hypothetical play - needs to approach as a ko because of an eye or other kos of the defender in the same semeai. \ \ This ordinary case has been neglected by most because the Japanese rules made such an extraordinary fuss about bent-4. The bent-4-related rules were bad as exceptions / precedents - but they were even much worse as misleading teaching.[/QUOTE] \ \ OK, I'll try to be as specific as possible: \ \ In how many of your games do you have a situation where: \ [LIST] \ [*]You can start a certain ko \ [*]The opponent can neither start the ko, nor remove it \ [*]When you start the ko, you capture first \ [*]You can leave the situation until the end of the game \ [*]The situation is [i]not[/i] bent four in the corner. \ [/LIST] \ \ As I said, I have never had such a situation in any game of mine. If they truly occur in one out of every 10 games of yours, then surely you have plenty of game records of them that you can show?" "entropi","","2009-01-20 03:23:24","[QUOTE=HermanHiddema][LIST] \ [*]You can start a certain ko \ [*]The opponent can neither start the ko, nor remove it \ [*]When you start the ko, you capture first \ [*]You can leave the situation until the end of the game \ [*]The situation is [i]not[/i] bent four in the corner. \ [/LIST] \ [/QUOTE] \ \ This is the game I was initially mentioning. \ It actually fulfills the 5 requirements, but I now realize that white is in fact stone dead :) \ While playing the fast game I thought it is a similar situation as bent four, but apparently it is not because white has no way to kill black because of the \"internal\" eye. \ \ Sorry about initiating a long discussion based on a \"wrong\" question :o \ \ [go]$$From my game (simplified the surrounding positions) \ $$ | . . . . . . . \ $$ | X X X . . . . \ $$ | O O X . . . . \ $$ | . O X X X . . \ $$ | X O O O X X . \ $$ | X X X O O X . \ $$ | . X O . O X . \ $$ ---------------[/go]" "HermanHiddema","","2009-01-20 03:31:32","[QUOTE=entropi]This is the first time I post a diagram. I hope it will work:\ \ [go]$$From my game (simplified the surrounding positions)\ $$ | . . . . . . . . .\ $$ | . . . . . . . . .\ $$ | . . . . . . . . .\ $$ | X X X . . . . . .\ $$ | O O X . . . . . .\ $$ | . O X X X . . . .\ $$ | X O O O X X . . .\ $$ | X X X O O X . . .\ $$ | . X O . O X . . .\ $$ ------------------[/go][/QUOTE]\ \ In this situation, white is dead. White has no way to ever win this ko, because white can never capture the corner.\ \ [go]$$Altered position\ $$ | . . . . . . . . .\ $$ | . . . . . . . . .\ $$ | X X X . . . . . .\ $$ | O O X . . . . . .\ $$ | . O X . . . . . .\ $$ | O O X X X . . . .\ $$ | a O O O X X . . .\ $$ | X X X O O X . . .\ $$ | . X O . O X . . .\ $$ ------------------[/go]\ \ If the situation was like this (note: I have given white an eye as well now), then white would have a way to win this ko (approach at a, black captures, white must win the ko). But white really doesn't want to approach at a, because he his filling his own liberty.\ \ The same goes for black. black can only win this ko by:\ Taking the ko, then playing at a (white will then take the ko), and then black has to win the ko.\ \ So in this situation, if one player starts the ko, the other gets to take the ko first, so both players would rather see the other player start the ko. Such a situation is known as Ten Thousand Year Ko (see: [url]http://senseis.xmp.net/?TenThousandYearKo[/url])\ \ As long as neither player has enough ko threats to win this ko, the situation will probably remain on the board. Alternatively, in this position, white can choose to make it seki (by filing the ko), and black can do the same (by taking, then filling the ko).\ \ If it is still there at the end, and still neither player wants to play the ko, it will become seki." "RobertJasiek","","2009-01-20 03:39:10","HermanHiddema, I lack time to go through my game records for statistics, sorry. Maybe next year. \ \ Bent-4 at the game end occurs only in roughly every 500th of my games, or possibly less frequently. Other semeais with hidden kos are much more frequent in my games." "entropi","","2009-01-20 03:40:56","[QUOTE=HermanHiddema]In this situation, white is dead. White has no way to ever win this ko, because white can never capture the corner. \ [/QUOTE] \ \ Sure! I realized it after sending the message :)" "HermanHiddema","","2009-01-20 04:09:39","[QUOTE=RobertJasiek]HermanHiddema, I lack time to go through my game records for statistics, sorry. Maybe next year. \ \ Bent-4 at the game end occurs only in roughly every 500th of my games, or possibly less frequently. Other semeais with hidden kos are much more frequent in my games.[/QUOTE] \ \ Oh, I don't particularly want statistics (though that might be nice), I'm just asking for an example or two. As I've never had such a position in my games, other than bent four, I'm interested in seeing how some of them might form." "RobertJasiek","","2009-01-20 05:37:35","Basic semeai. One side dead. The dead side has one of the following: \ \ a) eye + open ko \ \ b) eye + several open kos \ \ c) 2 or more open kos \ \ So far the kos are not hidden. The following have hidden kos: \ \ d) eye + the attacker can atari a stone that is then open for the defender \ \ e) (d) combined with another open ko or another \"the attacker can atari a stone that is then open for the defender\" \ \ f) like (c) but at least 2 \"the attacker can atari a stone that is then open for the defender\" \ \ g) similar as before but now a hidden ko is created as follows: (approach and) capture an opposing string by means of the attacker's single stone that the defender immediately recaptures and that then the attacker has to atari making a ko. \ \ etc." "HermanHiddema","","2009-01-20 06:22:48","[QUOTE=RobertJasiek]Basic semeai. One side dead. The dead side has one of the following:\ \ a) eye + open ko\ \ b) eye + several open kos\ \ c) 2 or more open kos\ \ So far the kos are not hidden. The following have hidden kos:\ \ d) eye + the attacker can atari a stone that is then open for the defender\ \ e) (d) combined with another open ko or another \"the attacker can atari a stone that is then open for the defender\"\ \ f) like (c) but at least 2 \"the attacker can atari a stone that is then open for the defender\"\ \ g) similar as before but now a hidden ko is created as follows: (approach and) capture an opposing string by means of the attacker's single stone that the defender immediately recaptures and that then the attacker has to atari making a ko.\ \ etc.[/QUOTE]\ \ Ok, so correct me if I'm wrong, those are:\ \ a)\ [go]$$B eye + open ko\ $$ ------------------\ $$ | X . X . X O X . .\ $$ | O X X X X O X . .\ $$ | O O O O O O X . .\ $$ | . X . X . . X . .\ $$ | . X . X . X . . .\ [/go]\ \ b)\ [go]$$B eye + several open kos\ $$ ------------------\ $$ | X . X . X . X O X\ $$ | O X X X X X O O X\ $$ | O O O O O O O X X\ $$ | . X . X . X . X .\ $$ | . X . X . X . X .\ [/go]\ \ c)\ [go]$$B 2 or more open kos\ $$ ------------------\ $$ | X . X . X O X . .\ $$ | O X X X O O X . .\ $$ | O O O O O X X . .\ $$ | . X . X . X . . .\ $$ | . X . X . X . . .\ [/go]\ \ d)\ [go]$$B eye + attacker can atari\ $$ ------------------\ $$ | X . X . X . O X .\ $$ | X X X X O O O X .\ $$ | O O O O O X X X .\ $$ | . X . X . X . . .\ $$ | . X . X . . X . .\ [/go]\ \ e)\ [go]$$B eye + attacker can atari + open ko\ $$ -------------------\ $$ | X . X . X . X . O X\ $$ | O X X X X X O O O X\ $$ | O O O O O O O X X X\ $$ | . X . . . . X . . .\ $$ | . X . X X . X . . .\ [/go]\ \ f)\ [go]$$B 2 times attacker can atari\ $$ -------------------\ $$ | . X . X . X . O X\ $$ | O O X X X O O O X\ $$ | . O O O O O X X X\ $$ | X . . X . X . . .\ $$ | . X X X . X . . .\ [/go]\ \ g)\ [go]$$B capture-recapture-atari\ $$ ------------------\ $$ | X . X . X X O X .\ $$ | X X X X O O O X .\ $$ | O O O O O . X X .\ $$ | . X . X . X . . .\ $$ | . X . X . . X . .\ [/go]\ \ Yes?" "RobertJasiek","","2009-01-20 07:00:13","Yes, except that the attacker needs some more liberties in some examples so that they would likely exist until the end of alternation. In the last example, the string on atari or to become being ataried can also be bigger. \ \ EDIT: The attacker can have two eyes on the outside to make it clearer why the won semeai would still exist at the game end." "HermanHiddema","","2009-01-20 07:08:53","[QUOTE=RobertJasiek]Yes, except that the attacker needs some more liberties in some examples so that they would likely exist until the end of alternation. In the last example, the string on atari or to become being ataried can also be bigger. \ \ EDIT: The attacker can have two eyes on the outside to make it clearer why the won semeai would still exist at the game end.[/QUOTE] \ \ \ Like this? \ \ \ a) \ [go]$$B eye + open ko \ $$ ------------------ \ $$ | X . X . X O X . . \ $$ | O X X X X O X . . \ $$ | O O O O O O X . . \ $$ | . O . O X X X . . \ $$ | O O O X . . . . . \ $$ | X X X X . . . . . \ [/go] \ \ b) \ [go]$$B eye + several open kos \ $$ ------------------ \ $$ | X . X . X . X O X \ $$ | O X X X X X O O X \ $$ | O O O O O O O X X \ $$ | . O . O X X X X . \ $$ | O O O X . . . . . \ $$ | X X X X . . . . . \ [/go] \ \ c) \ [go]$$B 2 or more open kos \ $$ ------------------ \ $$ | X . X . X O X . . \ $$ | O X X X O O X . . \ $$ | O O O O O X X . . \ $$ | . O . O X X . . . \ $$ | O O O X . . . . . \ $$ | X X X X . X . . . \ [/go] \ \ d) \ [go]$$B eye + attacker can atari \ $$ ------------------ \ $$ | X . X . X . O X . \ $$ | X X X X O O O X . \ $$ | O O O O O X X X . \ $$ | . O . O X X . . . \ $$ | O O O X . . . . . \ $$ | X X X X . X . . . \ [/go] \ \ e) \ [go]$$B eye + attacker can atari + open ko \ $$ ------------------- \ $$ | X . X . X . X . O X \ $$ | O X X X X X O O O X \ $$ | O O O O O O O X X X \ $$ | . O . O X X X . . . \ $$ | O O O X . . . . X . \ $$ | X X X X . X . . . . \ [/go] \ \ f) \ [go]$$B 2 times attacker can atari \ $$ ------------------- \ $$ | . X . X . X . O X \ $$ | O O X X X O O O X \ $$ | . O O O O O X X X \ $$ | O O . O X X . . . \ $$ | X O O X . . . X . \ $$ | X X X X . X . . . \ [/go] \ \ g) \ [go]$$B capture-recapture-atari \ $$ ------------------ \ $$ | X . X . X X O X . \ $$ | X X X X O O O X . \ $$ | O O O O O X X X . \ $$ | . O . O X X . . . \ $$ | O O O X . . . . . \ $$ | X X X X . X . . . \ [/go]" "RobertJasiek","","2009-01-20 07:17:52","Yes. - One can also imagine your earlier examples though but it is much less likely then that they would persist to the game end." "HermanHiddema","","2009-01-20 07:27:17","[QUOTE=RobertJasiek]Yes. - One can also imagine your earlier examples though but it is much less likely then that they would persist to the game end.[/QUOTE] \ \ Yes, the others would probably be played out before game end. \ \ So can you explain why these situations are similar to this one: \ \ [go]$$B bent four \ $$ ------------ \ $$ | O O O . X O \ $$ | . X X X X O \ $$ | X X O O O O \ $$ | O O O . . . \ [/go] \ \ It seems to me that in none of those cases, there is any need for white to remove ko threats in order to kill black. They might lead to moonshine life if there is some infinite supply of ko-threats for black on the board (double-ko seki or some such thing), but other than that, white can kill them locally without problem." "RobertJasiek","","2009-01-20 09:04:26","Can you explain why ko threats need be removed for bent-4? :) If the bent-4 string's status are analysed on a per string (or per group for whichever suitable definition of group) basis, then ko threats and their effect elsewhere on the board are immaterial(!) for the possibility to create a two-eye-formation on at least one intersection of the currently studied bent-4 string. \ \ Since for bent-4 or every other (reasonably ordinary) hidden ko example diagram ko threats are immaterial, both have the same principle behaviour." "HermanHiddema","","2009-01-20 09:16:13","[QUOTE=RobertJasiek]Can you explain why ko threats need be removed for bent-4? :) If the bent-4 string's status are analysed on a per string (or per group for whichever suitable definition of group) basis, then ko threats and their effect elsewhere on the board are immaterial(!) for the possibility to create a two-eye-formation on at least one intersection of the currently studied bent-4 string. \ \ Since for bent-4 or every other (reasonably ordinary) hidden ko example diagram ko threats are immaterial, both have the same principle behaviour.[/QUOTE] \ \ Since two-eye-formation is not part of the definition of current official rulesets, that possibility is meaningless. \ \ I am not arguing that there are is no way to resolve such situations with hypothetical play. I am not denying that the rules could be written in such a way that these cases would be identical. \ \ I am arguing that under the [i]current[/i] rule sets, there are situations that behave identical to bent four in the corner, and that these positions are not examples (the example I gave in response to Harleqin earlier is)." "RobertJasiek","","2009-01-20 10:31:37","Two-eye-formation implies uncapturable implies J1989-style alive of the inner bent-4 string(s). Hence two-eye-formation is very much applicable here although not part of J1989 as a term. For WAGC Rules, see the other thread." "HermanHiddema","","2009-01-20 11:12:00","[QUOTE=RobertJasiek]Two-eye-formation implies uncapturable implies J1989-style alive of the inner bent-4 string(s). Hence two-eye-formation is very much applicable here although not part of J1989 as a term. For WAGC Rules, see the other thread.[/QUOTE] \ \ What is relevant here is current rules, ie current Japanese, Chinese, AGA, etc. Since two-eye-formation and uncapturable are not part of their definition, they do not apply. \ \ Under J1989, the bent four depends on pass-ko-threat-only for both finite and infinite number of ko threats on the board. The other positions only depend on pass-ko-threat-only in the case of infinite unremovable ko threats. \ \ Also, under Chinese (and AGA, NZ, etc) rules, the bent four situation is ko (and can go either way if there are unremovable ko threaths), while the other positions are dead." "RobertJasiek","","2009-01-21 02:15:08","To simplify discussion a bit, I answer only for Japanese style rules because that was the original concern in this thread. \ \ Current Japanese rulesets are: \ 1) the Japanese 1989 Rules (with some not translated and therefore mostly unknown to us amendments) \ 2) the World Amateur Go Championship [1979] Rules \ 3) various Verbal Japanese Rules(ets), e.g., the Verbal European-Japanese Rules \ 4) various go server Japanese rules, e.g., the KGS-Japanese Rules \ \ For each such official Japanese style ruleset, the answer will differ. \ \ *********************************************************** \ \ The Japanese 1989 Rules \ ----------------------- \ \ Apart from the in-seki versus not in-seki classification, they use two types of life: uncapturable and capturable. As everybody knows, they do not use explicit terms for either. Instead they speak of \"if they cannot be captured by the opponent\" and \"if capturing them would enable a new stone to be played that the opponent could not capture\". \ \ If it is your aim to ensure that nobody understands what you are saying then never use the abbreviations \"uncapturable\" and \"capturable\" but always write down the complete phrases \"if they cannot be captured by the opponent\" and \"if capturing them would enable a new stone to be played that the opponent could not capture\" while each time ensuring that you are using either with the function of a term or fixed concept. Stupid! It is much easier to use the terms \"uncapturable\" and \"capturable\". \ \ The next thing is to interpret these terms. See my commentary webpages. In particular, \"can\" and \"could\" have to be explained, \"enable\" is ambiguous, \"enable\" is a flawed concept that does not always produce the intended results, capturable is incomplete because it covers only \"capturable-1\" and overlooks the possibility of \"capturable-2\". IOW, a profound application of the official rules cannot be done without also referring to my commentaries on them. \ \ Furthermore, the official pass-for-ko-rule is ambiguous and incomplete. Again, a profound application of the official rules cannot be done without also referring to my commentaries on them. In particular, the official pass-for-a-specific-ko-rule does not work as intended in some positions. For this reason, it should be the generic ko-pass of the Japanese 2003 Rules. (For bent-4, usually it does not make a difference though.) \ \ Let me repeat that for bent-4 (the usual bent-4 - not a bent-4 with adjacent seki shape or some such; that would need fresh study) ko threats elsewhere on the board are IMMATERIAL. Try it out for yourself! The ko threat elsewhere does not prevent a capturable-1 stone in the local-1 (i.e., under the currently considered initial inner bent-4 string). So what you might want to call a ko threat is none, except technically as being a play made in between ko-capture and ko-dissolution. Let me also repeat that the plain ordinary basic ko rule creates the same behaviour. (Or superko.) \ \ So let me repeat that, under J1989, the bent-4 does NEITHER depend on pass-ko-threat-only, NOR on \"the\" number of ko threats on the board being finite, NOR on \"the\" number of ko threats being infinite. The behaviour of bent-4 is exactly the same regardless of the number of ko threats elsewhere. It is (apart from the exact move type of the pass) exactly the same for the basic ko rule, for the superko rule, for the pass-for-a-particular-ko-rule, and for the generic-pass-for-ko-rule. \ \ *********************************************************** \ \ The World Amateur Go Championship [1979] Rules \ ---------------------------------------------- \ \ Life in the rules is too ambiguous. It cannot be reasonably applied without the aid of commentaries and models. \ \ One model defines independently-alive via two-eye-formation and scores only in between independently-alive strings of the same player. \ \ The other model defines uncapturable, capturable-1, and capturable-2, then distinguishes independently-alive from seki-alive, then scores in between independently-alive strings of the same player and does not score in between seki-alive. \ \ Both models are equivalent. This has been mathematically proven. \ \ Therefore two-eye-formation is as good an explanation as the three (un)capturability types. Even more, since J2003 uses the same terminology, two-eye-formation is also a good term for J1989 explanation. \ \ In particular, the attacker of a bent-4 can force creation of a two-eye-formation of his on at least one intersection of the initial inner string(s) stones. Practically speaking, the whole bent-4 can be covered by such a two-eye-formation. REGARDLESS of ko threats elsewhere or the exact kind of (hypothetical) ko rule. \ \ *********************************************************** \ \ Verbal Japanese Rules \ --------------------- \ \ If they claim to refer to official written rulesets, see above. If they claim to refer to \"common go theory\", then two-eye-formation is even a much better background explanation than (un)capturable because almost all players explain independent life by means of (the ability to create) two eyes to themselves. \ \ *********************************************************** \ \ Go Server Japanese Rules \ ------------------------ \ \ Similar to Verbal Japanese Rules, but with some server-specific exceptions." "RobertJasiek","","2009-01-21 02:20:04","Addition: Do not confuse bent-4 during the alternation with bent-4 after the game stop!" "HermanHiddema","","2009-01-21 02:27:48","[QUOTE=RobertJasiek]Addition: Do not confuse bent-4 during the alternation with bent-4 after the game stop![/QUOTE] \ \ Well of course. The whole \"bent four in the corner is dead\" thing is about unremovable bent fours that remain until the end. Bent four during the game is a simple (throw-in) ko." "HermanHiddema","","2009-01-21 02:40:42","i will reply to your other post this evening (don't have the time right now). \ \ I do have an unrelated question for you, now that we are discussing rules anyway. \ \ Have you ever come across the following position, or anything similar? And can you give your assessment as to the status of each of the groups on the board in various rule sets? \ \ [go]$$B 4x6 position \ $$ ------------- \ $$ | . O O X . . | \ $$ | X X O X X O | \ $$ | X X O X O O | \ $$ | . O O X X . | \ $$ ------------- \ [/go] \ \ This position was brought to my attention by Erik v/d Werf (the guy that solved Go on 5x5) who came across it during research into Go on 6x4 (although the basic idea can easily be transferred to larger boards, and is in fact not unlikely to come up in a game)." "RobertJasiek","","2009-01-21 04:52:46","I have not seen such in actual play, if you mean the same nakade size. I can solve it under various rulesets but by experience a thorough J2003 analysis would take between 30 and 200 minutes; more than I currently have. Beware of single regular passes! They make such things much more complex than one hopes." "RobertJasiek","","2009-01-21 05:11:17","Simplified Japanese Rules: part of the analysis beginning from its first play: \ \ 1) no plays -> no string is independently alive -> score = 0 \ \ 2) Black e4 -> bigger eye beats smaller eye; Black does not choose this \ \ 3) White ko throwin -> White does not have any ko threat play; White does not choose this \ \ 4) Black a1, a4, a2, e4, f4, a3, f1, b2,... both create a two-eye-formation -> the initial strings c1 and d1 are independently alive -> score = 9 - 10 = -1; Black chooses (1) instead of (4) \ \ Conclusion: successive passes are perfect analysis play under the Simplified Japanese Rules." "HermanHiddema","","2009-01-21 05:17:13","Here is an example position that could result from actual play:\ \ [go]$$W\ $$ -------------------------\ $$ | . . X O X W . X X . W X .\ $$ | 3 2 X O X W W W W W W X .\ $$ | 1 5 X O X X X X X X X X .\ $$ | . X X O X . . . . . . . .\ $$ | X X O O X . . . . . . . .\ $$ | O O O O X . . . . . . . .\ [/go]\ \ The upper left is the result of a standard \"walkie-talkie seven\" invasion as shown.\ \ The question in this case is whether the marked white group along the upper edge is dead or alive (in various rulesets). There are no ko-threats on the rest of the (not shown) board." "RobertJasiek","","2009-01-21 06:08:04","Correction of my rules case study: \ \ WAGC Rules applied to bent-4: The precendental ruling for bent-4 applies: \"A bent four in the corner is independently dead, regardless of the rest of the board.\" This overrides life and death rules paragraphs. \ \ Verbal Japanese Rules: We have to distinguish such that assume a similar precedental ruling for bent-4 and such that do not. If we consider such that do, then bent-4 is treated differently from all other hidden ko shapes: The precedent applies to bent-4 but (presumably) not to all the other shapes. (Another theory would be to generalize the precedent.)" "RobertJasiek","","2009-01-21 06:11:37","Applying J2003 to your new example is relatively easy: The local-2 of the marked white string does not extend to or beyond White's big uncapturable string. Therefore the marked string is dead. \ \ EDIT: Wait. You still have a dame there? This makes things more difficult..." "HermanHiddema","","2009-01-21 06:22:05","[QUOTE=RobertJasiek]Applying J2003 to your new example is relatively easy: The local-2 of the marked white string does not extend to or beyond White's big uncapturable string. Therefore the marked string is dead. \ \ EDIT: Wait. You still have a dame there? This makes things more difficult...[/QUOTE] \ \ The dame is an artefact, you can consider the rest of the board dame free and ko-threat free. I'll change the diagram." "Zahlman","","2009-01-21 15:58:00","This is what happens when I try to give a simple overview that will be understandable to a weaker player... x.x"