(add your issue at the end of this page with an appropriate headline)
Newbie question: What is the advantage of knowing a joseki? is it so that you can try to force the stones into a settled position while retaining sente? or is there another advantage?
Illume (25kyu): I suppose the major advantage is to recognize common continuations as early as possible. For example you might be in a position where you'd like to make life in the corner and connect two groups of stones. If you study similar positions you will know how easy or difficult it is to achieve both goals.
Scartol: Go is all about patterns. Recognizing life patterns, death patterns, fuseki patterns, yose patterns. Once you are able to recognize patterns, you can consider their context in your games and adapt your play appropriately. Knowing how to use joseki is useful to me because it allows me to approach difficult situations (like the sansan invasion) without fear. I know there is a mutually beneficial way to resolve the situation, and we can proceed along the path until there's a chance to alter the pattern.
Andre Engels: Basically, learning joseki is a short-cut to playing the best moves. If you know the joseki, you can then see the outcome immediately, and then quickly go through the moves until a place where either the opponent plays something different, or you prefer to do something different. Joseki contain a lot of knowledge from many players, which certainly beats having to think things out every time by oneself - some joseki sequences contain tesuji that even a top player would have a hard time thinking out behind the board. Also, when you know how the joseki goes on, you know whether it will be good to play it in a certain situation.
Velobici: Joseki are based upon the best play for both players viewing the corner in isolation from the rest of the board. Therein lie two dangers for folks studying joseki. The first danger is that the joseki is the best play....meaning that both players are pushing their moves to the limit. This leaves no room for error. If the other person deviates from the joseki and you dont know why their move is wrong so that you can choose the appropriate response, they may be able to convert your previous maximum effort moves into mistakes (because you failed to support the earlier moves correctly). The second danger is that the joseki ignores the rest of the board. You can play the joseki perfectly and still get a bad result because the joseki does not support your stones on the rest of the board. A simple example is chosing the wrong 3-4 point block when another invades at the 3-3 point below your 4-4 point. Many hamete are based upon deviating from the joseki and punishing the other players failure to push you for deviating.
Dieter: I agree with the second but strongly disagree with the first. Whether or not you are consciously playing joseki, a Go player should always go for the best move he can think of. If he gets punished for it, too bad. This is not something typical of joseki. Secondly, sometimes mistakes in joseki lead to disaster, sometimes they incur a 2 point loss. The fact that people are not aware of these differences is a real danger of learning joseki. "Aha! This is not joseki ! Now I will win the game." There is also another danger of learning joseki: Joseki as a source of bad habits.
BobMcGuigan: Joseki are sequences of the locally best moves but often different moves are required because of circumstances in neighboring corners or elsewhere on the board. It has often been said that joseki should not be memorized but understood move by move. Many sequences in Ishida's dictionary are no longer considered correct as a result of research and experience in pro games. If you really understand why each move is made you will be able to take advantage of mistakes and make up your own "joseki" if you don't like the standard moves. It's also worth noting that joseki "mistakes" that are considered fatal by pros might only result in a loss of one or two points, hardly likely to be fatal for most of us.
Malweth: The advantage of learning joseki is to more fully understand the flow, shape, strengths, and weaknesses of the stones in the begin to mid-game. Joseki study (IMHO, of course) should concentrate on ideas and meanings behind the "best move." Commentary on Joseki is essential to this and Ishida is one of the few available (and is excellent). It won't bestow on you instant good play in the begin-game, but it will give you the ideas you need to learn on your own (through playing). A certain level of reading is required for Joseki study and this (starting) level hovers around 3-7 kyu. Some Joseki are accessable to the "younger" (8-20k) such as the 3-3 invasion, but serious Joseki study does require good reading skills.
Joseki memorization, which is often confused for "studying" or "learning" joseki, is harmful and useless. It does no good to play out a memorized sequence if you can't give reasonable responses to new variations.
After some true Joseki study you will not have all the best moves, but you will be more adequately prepared to deal with bad moves (overplays) or new variations. A stronger teacher can help you when you see variations that you cannot understand.
ilan: Actually, there are 2 kinds of joseki: The first are standard sequences for invasion or connection, e.g., the 3-3 invasion under the 4-4 stones, and the second are what people most commonly consider joseki, that is, standard corner exchanges. I believe that the first type of josekis are probably a good thing to learn, and the 3-3 invasion should be known to beginners starting at 20K or so. I'm not so sure about the value of opening josekis for sub dan players, except to know some of the traps and pitfalls. There is definitely no value in knowing them if you don't know the reason for every single move. Basically, when you play joseki, you are playing 9p moves, with those concerns in mind, so the subtleties of the "best" moves have no bearing on all but the best amateur games. For example, take the standard 3-3 joseki that almost all players know:
Now, apparently, the move is rather subtle, and the usual pushing battle would be to play one line higher. However, it appears that this produces weaknesses in Black's position, which might prevent a later extension from
. I found this out by reading it in books, e.g., Lessons in the Fundamentals of Go, and I conjecture that most people playing either side of this joseki don't know the reason for the low posture of
. Furthermore, the reason for play
on the fourth line instead of between
and
doesn't seem clear to me at all. In fact, this move has been played against me numerous times, and I found out that it has also been tried in professional play. Therefore, the difference between it and joseki cannot be very serious. Maybe it's the possibility of a push and cut, but anyway, it is not an error, maybe simply a sub 9p move.
On the other hand, here is a situation where knowing joseki is very useful.
Many amateur players might play , but I don't know how many would find
without having seen it in a joseki book. In any case, White is in trouble.
Interestingly, this seems to have an application to this non joseki move that has been played against me many times.
Since is in no book that I have found, it was up to me to figure out what is wrong with it, and what to do about it.
After some deep thought and knowledge of the above tactic,
I concluded that maybe the best response is to play
here.
zinger: Surely black must cut with at a instead?
The reason is that if White plays what seems to be the most natural move at , then Black gets a favourable transposition into the previous variation.
Therefore my conclusion is that knowing the traps and pitfalls helps you deal with non standard plays.
malweth: I agree with most of your meaning, though I don't concur with some of the terminology. There are many different types of joseki, but all look for a settling of the stones. An invasion, a sacrifice, a peaceful or fighting joseki are possible... in all cases the goal is to make your stones strong and/or alive.
Furthermore, simply playing joseki isn't the same as playing 9p quality moves. Joseki is also about board position and fuseki. Playing one joseki is often a bad move based on the rest of the board. Even taking this into account, this only gets you to maybe 3-5p. Making new moves that give you a different and better board position is where the top pro moves are found.
Joseki requires a different style of reading and intuition from other parts of the game (mid-game is most similar). Studying joseki gives a good idea of what moves are possible (vital points). Once these vital points are recognized, real reading skill must come into play.
Chris Hayashida: You have to be careful analyzing that way, though. For example, even though the diagram "Unfavourable transposition for White" was a bad result for White, it doesn't necessarily mean that in that diagram is the best response to
. As zinger pointed out,
might be better. This result would also need to be compared to other variations. In this result, a and b are miai, capturing one of the two white stones.
ilan: I believe that the words "that way" can be omitted from your text, that is: "You have to be careful analyzing." In any case, all these responses are typical of Go player discussions, which are endless debates about the minutest technical details, and which end up obscuring general conceptual points. Of course, this is simply a reflection of the type of thinking the game requires.
---
phenomene : Well, I think a sequence linked from the page Joseki without further explanation should refer to something locally leading to an even result. Of course there are special circumstances where this sequence is good (when one has strength on the side as in your examples), but it is usually bad to play Kosumi-Tsuke when the opponent has room to extend. It may be called Joseki because it matches the translation of the Japanese "settled stones", but in usual go language, one uses the word Joseki for equal results only, and it is common to say "this is not joseki" for sequences better for one player. In short, it is dangerous to suggest a sequence is a rule when it is an exception.
Bill: I think it is what joseki books call a play that depend on the circumstances Baai no Te.
Andre Engels: I don't agree with phenomene. There are joseki in the joseki books for which they say, "Locally this is good for Black/White, but it may be different regarding on the overall position." Ishida for example, regards the sequence to the left a 'basic joseki', but also notes that "just taking this area in account, the exchange [...] is clearly favourable for Black."
BobMcGuigan: The famous statement that "for a meijin there are no joseki" suggests that very strong players make things up as they go along, or maybe that "joseki" are determined by social convention among go players. There are also various sequences that were given the "joseki" stamp of approval in, say, Ishida's dictionary and widely played in professional games but some years later are no longer considered "joseki". Maybe all joseki dictionaries should have a warning label that indicates that the moves in this book give results that are generally considered equal at the time of writing but are not guaranteed to be so in the future. Nakayama Noriyuki, writing on this topic, said that erroneous joseki sequences seem to persist in dictionaries for 30 to 50 years. In other words, it takes a long time for go writers (and even pros) to become aware of some of these changes of judgment.
Charles At least. If one works closely with them, one becomes aware of the uneven nature of their content. For example, repeating old analysis not based on high-level play has been common up until recent times (when, to be fair, there are many thousand more games played per year by pros). You can find minor mistakes in suji, too; which is probably a sign that an insei compiled a section, rather than a strong pro. The Igo Daijiten must quite largely consist of research rather than joseki from real games. One reason that Korean go has had such an impact is that a more sceptical attitude has been applied to 'things everyone knows'.
Warp: AFAIK it's wrong to think that a joseki is a pattern which (always) gives equal result to both players, and this is a common misconception. A joseki is an established pattern which gives the best possible outcome for both players given a certain initial position of stones. A joseki doesn't have to always give equal result to both as long as it gives the best possible result to both starting from a certain arrangement of stones. It may perfectly be that eg. in a corner situation one of the players is already much stronger than the other (because he has for example more stones there) and there's absolutely no way for the other player to get an equal result anymore, no matter what he does (assuming his opponent plays well). However, he may still have a good sequence which gives him the best possible outcome in that specific situation. In other words, he minimizes his loss. For example, if there's a stone at 4-4 and the opponent plays at 3-3, that 3-3 play is not a joseki, it's simply an invasion. If this happens too early in the game it's usually a bad invasion because the invader is in an inferior position. Thus invading at 3-3 too early in the game is usually a mistake. However, that is irrelevant with respect to the defition of joseki. Now, the classical sequence which usually follows from this arrangement of stones is a joseki because it gives the best possible outcome from this position forward for both players. It was a mistake to play the 3-3 stone in the first place, yes, but the following moves are not mistakes anymore, they simply minimize the damage. Thus those following moves are joseki. Curiously, even some pros misunderstand this. (By the way, I think that the definition of joseki in the joseki page could include this notion.)
Arieh: The distinction between old and new Joseki is irrelevant to all but professional and near-professional strength players. If a joseki was considered the best option for decades by top pros in the past, the fact that today's top pros have found moves that give a slightly better position has no relevance for most of us. If an old Joseki leaves black better by a quarter of a point, it doesn't mean much for us as white when we are making mistakes that cost us 2+ points.
Lynx: Sure, but the goat in me says that while fixing that 2 point mistake would be great, if you win by just half-a-point, getting that quarter of a point was worth it. Knowledge is its own reward, to a degree, and playing Go for any non-competative reason is calculated firstly to maximize pleasure and only secondly for winning.
Arieh: I agree entirely. Memorizing Joseki seems like the epitome of knowledge for the sake of winning. Studying Joseki can be the opposite - appreciation of shape and general principles of territory vs influence. However, the distinction between old and new Joseki is by definition subtle enough that, except for very advanced players, it has to be memorization; if even advanced players could really understand why the new Joseki is better than the old, than it wouldn't have taken the best pros decades to find. We may remember the explanation that a new Joseki is better for black because it yields more center influence, but to really understand the distinction we'd have to be at a level approaching that of at least the pros who missed the better move for decades. Harping on an exceptionally subtle distinction between Josekis is like an amateur skiier spending lots of time waxing his skiis. Yes, the wax will make you ski faster, but it has much less impact than other, more useful, and more fun training.
Bill: I do not believe that a 1/4 point difference is enough in itself to scuttle a joseki. Surrounding conditions will be right for it to be chosen often enough to remain joseki. I think the difference has to be at least 1 point for the size of the play to be decisive. (Style also matters, as does the type of fuseki.)
Say that a certain old joseki play loses one point versus the modern joseki play. To be sure, in amateur games a difference of one point hardly ever decides the game. However, a difference of one stone in strength comes to a difference of about 15 points in the final score. Since a player makes around 120 plays in a game that is counted, the average difference per play is only about 1/8 point.
If you simply memorize the new joseki, you gain 1 point every time it comes up. But if you understand why the move gains 1 point, you can use that understanding to play better in general, and that understanding can be worth one or more stones in strength.
Besides, the differences between old and new joseki are not always all that subtle. I can think of at least one old joseki that you never see anymore. Why? The answer is really quite simple. The player who is pincered makes a contact play to secure his group. It was later realized that that was unnecessary, and it strengthened the opponent's stones as well. So people stopped playing it. The principle is not hard to understand, and, in fact, it applies to a number of old joseki. To be sure, recognizing when you do not need a contact play to make your group safe is another question, but the principle is worth learning, and it is not particularly subtle.
Floris?: I think the current pages on joseki are very flat and not very much of a help. They just give ya a few sequences and a bit of information that really I (and I asume there are others) don't care about such as "This is a traditional joseki that is still played". It really gives you the idea of: "Here is a sequence, play it or lose". I think SL should add extremely detailed commentary explaing when _it is appropriate to use the joseki, when not to use it, what the joseki focuses on (influence, corner, left side, etc.), and i guess also some history and many heavy commented examples (from pro games) of global situations of when and when not to apply this joseki sequence. I realize this seems like an impossible task, which it is. But still I feel it would add a lot of value to SL. I can contribute a lot too, but I feel some high dan player should take the initiative.
P.S. Should i add discussion items to the bottom of the page or the top? I've seen both done.
Hu: Discussion is usually added at the bottom, unless one is addressing a specific point in context, as I am here. In a few cases, discussion is added at the top, most notably Messages to People Currently Present in the Library and similar pages, which are voluminous.
Velobici: Gobase.org provides one element of what Floris? has suggested, the many examples from professional games of when a particular joseki was used and when not. The game records are not commented. It is beyond the scope of SL to provide all that is suggested. However, general recommendations of when a joseki is appropriate and when it is not would be most welcome. One question is at what level does this become suitable. My son (now 1D AGA, 2k KGS, regularly plays winning a fair share of games against a 5D AGA/3D IGS at three/four stones) has been taking lessons weekly from a professional residing in the United States for a year. The teacher has indicated which direction of play is larger/better and what general type of joseki is desired, but never said that a particular joseki is to be used in a particular situation. This is interesting. The idea of "use this joseki when these conditions arise" seems to be too formulaic to support good go play. Perhaps, when my son is stronger this will change. Perhaps, the direction of thought is not the right direction of play.
Charles My explanation.
First, it may be that SL isn't the ideal joseki dictionary. That wouldn't be surprising: no joseki dictionary that exists is the ideal joseki dictionary.
Secondly, when I post a joseki page here, it is never copied from a book. It is always researched from a database. So the information is actually more reliable in certain ways than many joseki books. We point out here some ways in which the Ishida Joseki Dictionary is misleading, for example.
Thirdly, I think the most important achievement so far is to get a structure on SL to which joseki information can be added, and then found. Please note that even this much took a number of people many hours of work.
Fourthly, it is basically a lazy attitude that someone else should work hard, so that others should have free instant access to research. My experience is that giving people hundreds of free web pages doesn't really cause gratitude. Well, human nature as usual. But it makes no sense on a Wiki to say 'someone else should do this for me'. A wiki is a community first: if it is treated like a service, eventually those doing the real work may lose interest.
Fifthly, the importance of joseki study is vastly overrated. I do think we need a reference section here on joseki. The reason is that links out of other discussions can happen. And so that backlinks can handle 'how do you use this joseki' questions. Floris, have you thought of using backlinks? In time, they should build up naturally.
Sixth, for myself I would certainly give fuseki patterns, and in particular side patterns, priority over joseki.
Seventh, there is a fair amount posted here already that isn't easily available in English. If it is due to me, then that's because I think it has some value. I have posted comments before, putting words in the mouth of Sensei (our mythical founder). I don't believe that Sensei approves of trying to tell the whole story on a joseki. It's too narrow a view of the game.
Finally, I have just checked and there are 411 pages here with the Joseki keyword. That's a number of books' worth, already. And that's without some sort of Editor figure.
Hu: As usual, Charles has expressed a thoughtful point of view, cogently, which I endorse. To address Floris' concerns, yes, it would be nice. Just do it. Where it might be wrong it will soon be corrected. What will be most right about it will be that it will have been started. Coincidentally, just a day before, I finally got working on a Joseki Heuristics page, despite my lack of knowledge, thanks to Bob kindly sparking me into life. This may be a place to help guide some selections between joseki, or perhaps a jumping off point to some more specific guidelines for selecting between joseki.
Floris?: Ok, my points:
Charles The point about navigation is fair; all I can say is that it is constantly being worked on here.
For the rest:
Floris, why don't you post some particular positions here for discussion? You know, there are other ways, besides asking for a 'perfect' set of joseki pages - as if someone could make them appear, by magic.
Back on the main topic, we know that Michael Redmond, starting as an insei at amateur 5 dan level, had to forget all his joseki knowledge learned up until then. He has told us that. What I don't know is how he then replaced it by joseki knowledge to take him to pro 9 dan. But I suspect it was painstaking work, based on researching in pro games. Sound at all familiar?
jfc: What is best for one person is not necessarily best for another. While it is safe to say that nearly all players want to become the best that they can, different players have different resources at their disposal. as an insei Michael Redmond had the following at his disposal:
- a great natural ability.
- fellow insei's to compete and study with on a daily basis
- regular professional instruction
- Lots of time to devote to Go. As an insei (and later as a pro) studying Go was his job.
For someone with these assets, forgetting all joseki may be the correct approach in the quest for acheiving the highest level of play. For others with more meager resources, learning joseki and then focusing more time and energy on other aspects of the game (e.g. tsumego) may give better results.
There is a very good chance there will be very sensible comments, discussion within 48 hours (my experience: 5 minutes, by Charles and unkx80, but that may be exceptional).
[1] AvatarDJFlux: I believe it is exactly the other way round, there is no Joseki without Fuseki...
Since when Tactics come before Strategy??!?
I think Pro's choose a Fuseki strategy and then seek the proper Joseki (if any, otherwise they invent it) to reach their goal...
This discussion made me write A Zen Way to Joseki, something I had in mind since long...
(Sebastian:) The diagram here starts in the upper right corner, while the diagrams in the individual pages start upper left. The advantage of starting upper left is that diagrams are better comparable - they always start at the same point, regardless of how much of the board they show. My question is: Would it make sense to flip the diagram on this page?
I think this should be romanised as jungseok. Charles
Alex Weldon: One of the problems with the Korean language is that there are three different vowels that are sometimes romanized as "u." The only one which is "correct" to romanize is the one that looks like a fat T, and which is pronounced like "u" in "flute". The others are -| (pronounced like the "u" in "hung"), which should be romanized as "eo" and -- (pronounced sort of like the "eu" in the French word "peu," closest English equivalent maybe "u" in "put"), which should be romanized as "eu."
So, if Charles' romanization is correct (probably is), the word would be pronounced like "joong sock."
Perhaps jungseok should be added as an alias for this page.
Charles Just checked - it's the -| vowel in both cases, so jeongseok or even cheongseok.
Bill: I believe that the closest English vowel sounds like jung-suck.
maruseru: I'm neither an English native speaker nor a Korean native speaker, but the Korean character ㅓ is, I believe, not pronounced like the "u" in "hung", but more like the "ou" in "soul" (hence the capital Seoul, which is the romanized version of 서울, is pronounced close to "soul"), or maybe the "aw" in "saw". I guess the closest English vowel sounds for 정석 would therefore be "joungsouk".
Bob Myers: This is just wrong. Why would you write something you know nothing about? Seoul is pronounced "soul" in English but certainly not in Korean; "suh-ul" is probably the closest we can come, with the "uh" part being what's in question here.
maruseru: Oh, I'm so very sorry to offend your ego. Have I disturbed a discussion among the local demi-gods? So sorry. Henceforth I shall ask your permission to write something I might not know everything about. Would that be alright? Would you, as the authoritative voice here, lower yourself to explain how "suck", which is pronounced, as far as I dare to believe, like the german "sack", would be a close vowel sound for 석? Or how "uh", which I, in my limited understanding, associate mostly with "uh-uh", pronounced like the german "ah-ah", would be a vowel sound for ㅓ? Kindly enlighten this unworthy soul and forgive my impertinence on this subject.
Bill: English sock is closer to German sack. English suck is closer to German söck.
maruseru: Maybe the confusion is between American English and British English. British English "sock" is like German "sock"; American English "sock" does sound more like German "sack". NB: please no reprimand from Mr Myers; I'm from Austria and I've lived in London for five years. My knowledge of Korean is indeed limited; I've only taken the first semester course of Korean at the department of Koreanology at the University of Vienna, but I do hope that I've learned the pronounciation of the Korean characters.
Bob Myers: Sorry for getting your undies in a twist. If you know the pronunciation of the Korean characters, why would you say that the native pronunciation for Seoul is something like the English "soul"?
xela: The vowels in "sock" and "suck" can vary a lot depending on which part of England (or elsewhere) the speaker is from. One day I will learn how the international phonetic alphabet works, then I will be able to discuss these things without causing mass confusion.
Phelan: I would like to use this time to remind everyone to keep calm, and discuss rationally. Emotion and intention don't translate well in a written discussion, and could be very misinterpreted.
maruseru: (To BobMyers) Like I said, ㅓ to me sounds closer to how an English person would pronounce "soul" than to how an English person would pronounce "uh". But I see now that we're probably talking about different English dialects. Maybe we can agree that it sounds like the "aw" as in "thaw" - that, to me, sounds rather similar in both American and British English. But xela is right; the international phonetic alphabet would work better. To speakers of German, it's even more muddled: The sound is somewhere between "a" and "au", like an "au" with more rounded (pursed?) lips.
Bill: I am certainly no expert in Korean or Korean pronunciation. However, I attended the 4th International Conference on Baduk in Jeonju this past fall. The first spelling I saw for Jeonju was Junjoo, which is how it sounds to me. Incheon also sounds like Inchun. They do not sound anything like Jawnjoo and Inchawn.
maruseru: Hm, I just looked it up in Elementary Korean and they say the character they use for the phonetic transcription of ㅓ sounds like "aw" as in "thaw" or "uh" as in "uh-oh". Interesting. I'd have pronounced the latter differently, but then I'm not an English native speaker either. But we have "thaw" on good authority now. :) About the different pronounciations - I guess there are different accents and dialects in Korea as well, so the vowel could be pronounced differently in different parts of Korea.
Anyway, they use ɔ as the phonetic transcription for ㅓ; there is a Wikipedia article on the International Phonetic Alphabet that has a vowel chart that shows how the tongue should be positioned for this vowel to form.
Bill: Thanks, maruseru. :-) That's very helpful. As for the aw/uh difference, it is not uncommon for the same vowel (phoneme) to be pronounced somewhat differently in different contexts. It may be that before n or ng the uh variant predominates, and before k or another vowel the aw variant predominates. Maybe the closest English is something like jung-sock.
(Later.) Wikipedia gives ʌ for eo; i. e., open-mid back unrounded, not rounded. It also says that for most speakers, the long form is əː instead of ʌ: . That explains the uh.