Yose = endgame? No way, Jose!

[John Fairbairn]

Many people here regard yose as meaning the endgame. I have long tried to correct that misapprehension, pointing out that yose moves can occur in the opening. There has been a grudging acceptance among a few strong players that, yes, well, there are a few josekis where you can play yose moves, but most people are surprisingly resistant. Which is one reason I like to harp on about it

At any rate, I happen to believe Honinbo Shusai knew more about go than most people, and in the position below it is Black to play move 37. He calls the move played, by Yasui Chitoku, the start of the yose. And interestingly, LeelaZero seems, implicitly, to agree...

Where did Black play?

Chitoku played A. More fully, Shusai said this move an be seen as signalling the start of the boundary-play stage and since Black has managed to play first in this stage without having lost any of the effect of first move, he can be optimistic about the outcome. It is useful further to note that Black is adopting a one-weak-group strategy in this game with the three-stone wall down below, and he wants to be cautious about any spill-over from the attack on that leaking into his moyo at the top. Chitoku's move can also be called a kakoi (surrounding move).



LZ gives an interesting insight into this. It seems confused by the position in that it offered no less than 19 moves. Its best move was B, but very late in the search it came up with C as second best. However, all the moves presented were very, very close together in terms of win rate. It seems that a multitude of moves for a bot may represent quiescence in the position.

LZ's array of moves did not include Chitoku's A, but it did include D, E, F, G and H, and so it clearly likewise gave a high priority to this area. Furthermore, I would characterise all its 19 choices as yose moves - 5 of them were on the second line. The reason for not including Chitoku's A can quite likely be explained by the fact that there was no komi, i.e. Chitoku could afford to play conservatively.

Yose is a verbal form from a causative verb of various meanings, but the root sense can probably be taken as gathering together. I prefer the image of pulling tight the draw strings of a bag. Keeping this sense in mind also explains the relationship with yose-ko, which likewise has nothing directly to do with the endgame. I prefer the term "boundary plays", which usefully also contains some information about what you are trying to achieve. Of course, nearly all boundary plays occur in the last third of the game, so "endgame" usually passes muster. But not here!.

BTW if you like Japanese terms (and advanced go theory) you should note that after Black 37 White has to make a choice between a torikake strategy and a yoritsuki strategy. Because Black has chosen a shinogi strategy, a karami strategy is not available to White.

[Uberdude]

From http://rusgo.org/an-interview-with-ivan-detkov/, I highlighted part of it

Interviewer: Yose was the strongest part of your Go. Please give advices to our readers, who would like to improve yose. Most of them think that studying the endgame book by Tomoko Ogawe is the only method available.

Detkov: Many amateurs do not like yose. There is no fun to calculate territorial benefits of each move and pick the best, especially if you are already tired. Fuseki abstractions and middle game fight are more interesting for most players. This is true but things are different if you really care about the result. I like yose because I like to win.You study yose when you study go. The more time you spend the better. I did all kind of things including book reading, working on tests, and more but here is my main “secret”. It will work for everybody with a good visual memory. You should have a good visual memory to be a really strong player anyway. I am looking through as many different professional games as I can. I rarely try to analyze it deeply. It is more like series of beautiful pictures going through my vision and imprinting in memory. It builds up intuition because the brain is collecting a library that helps to chose candidate moves later. I guarantee you good results after ten thousand games.Be patient and look through all games to the last move. It is all about yose, right? You can do it quickly if nothing interesting is happening. While you are looking, try to feel a critical moment when the game shifts and predict a move. You can do it as often as you like. Try to check who is ahead regularly if it does not kill fun to watch. This process must be a pleasure or it will not work. If you see an unexpected interesting move, put a diagram in a notebook. I have collected hundreds of yose tesuji over the years. These tesuji are beautiful and they are so different from killer or cut placements!Knowledge of a good set of technical tricks and general understanding of miai and values in sente and gote make the job done if you add consistency and patience. If you want to improve further, you should know who is ahead all the time in the game in general and especially in yose. It helps to make a right decision when it comes to a level of acceptable risk. I confess that I was always too lazy to do it consistently enough. At some moment I realized that the whole game except an early fuseki is in fact preparation for yose. It seriously improved my skills. You should start to build your plans for yose very early been constantly aware about possible losses and gains on territorial borders when they are just emerging. How else you can correctly estimate territories? Notion of miai is an important key. Any pair of moves with the same or about the same properties could be thrown away, which helps to keep the process of analysis reasonably quick. The worst error is to realize that the yose already started when your partner is going full steam ahead. I always felt bad when I could not play this sweet first kosumi on the second line in sente.It would be interesting to ask Alexey Lazarev the same question. It was normally not a problem for me to turn -5 or even -10 points into a victory but not with him. If I was losing 5 points before yose, it was the game result. This is the main reason why he was the top player in Soviet and later Russian go for so many years.

[Kirby]

Maybe it’s more of a Japanese misunderstanding than a go misunderstanding? Using the “boundary play” term could be helpful.

[gowan]

From http://rusgo.org/an-interview-with-ivan-detkov/, I highlighted part of it

Interviewer: Yose was the strongest part of your Go. Please give advices to our readers, who would like to improve yose. Most of them think that studying the endgame book by Tomoko Ogawe is the only method available.

Detkov: Many amateurs do not like yose. There is no fun to calculate territorial benefits of each move and pick the best, especially if you are already tired. Fuseki abstractions and middle game fight are more interesting for most players. This is true but things are different if you really care about the result. I like yose because I like to win.You study yose when you study go. The more time you spend the better. I did all kind of things including book reading, working on tests, and more but here is my main “secret”. It will work for everybody with a good visual memory. You should have a good visual memory to be a really strong player anyway. I am looking through as many different professional games as I can. I rarely try to analyze it deeply. It is more like series of beautiful pictures going through my vision and imprinting in memory. It builds up intuition because the brain is collecting a library that helps to chose candidate moves later. I guarantee you good results after ten thousand games.Be patient and look through all games to the last move. It is all about yose, right? You can do it quickly if nothing interesting is happening. While you are looking, try to feel a critical moment when the game shifts and predict a move. You can do it as often as you like. Try to check who is ahead regularly if it does not kill fun to watch. This process must be a pleasure or it will not work. If you see an unexpected interesting move, put a diagram in a notebook. I have collected hundreds of yose tesuji over the years. These tesuji are beautiful and they are so different from killer or cut placements!Knowledge of a good set of technical tricks and general understanding of miai and values in sente and gote make the job done if you add consistency and patience. If you want to improve further, you should know who is ahead all the time in the game in general and especially in yose. It helps to make a right decision when it comes to a level of acceptable risk. I confess that I was always too lazy to do it consistently enough. At some moment I realized that the whole game except an early fuseki is in fact preparation for yose. It seriously improved my skills. You should start to build your plans for yose very early been constantly aware about possible losses and gains on territorial borders when they are just emerging. How else you can correctly estimate territories? Notion of miai is an important key. Any pair of moves with the same or about the same properties could be thrown away, which helps to keep the process of analysis reasonably quick. The worst error is to realize that the yose already started when your partner is going full steam ahead. I always felt bad when I could not play this sweet first kosumi on the second line in sente.It would be interesting to ask Alexey Lazarev the same question. It was normally not a problem for me to turn -5 or even -10 points into a victory but not with him. If I was losing 5 points before yose, it was the game result. This is the main reason why he was the top player in Soviet and later Russian go for so many years.

Haven't various pro players told us the same thing, to the effect that every move even in the fuseki, is aiming toward the yose. And, just as the Japanese word yose is misunderstood by many Western amateurs, so is the Japanese word fuseki.

[John Fairbairn]

Isn't all yose also by definition kakoi (not kakkoi, though)?

No. In go pro talk kakou is reserved for a high-level concept that hasn't really made it through to English yet. It's part of a nexus of prophylactic moves that includes e.g. mamoru and honte. It refers mainly to surrounding (loosely) an area that otherwise would be prone to invasion or erasure. Because it is prophylactic it is also gote and so it is the timing of it that makes it high level. Fujisawa Hideyuki is especially good on this term, and of course Sekiyama Riichi's famous gote no sente underlies this nexus, too.

Come to think it, nexus theory in go, which is more than a couple of decades old in the academic Japanese computer literature, may be the key to progress in understanding AI play. The bottleneck originally, as I recall, was sorting out the hierarchies within each nexus. I imagine that could now be done by running AI simulations on carefully controlled positions, perhaps a little like the one above. Here (leaving aside Shusai's use of "yose"), for example, kakoi seems to have a higher level than any mamoru or honte move.

[Elom]

Having just happened to be confusing over game stages, thank you for this analysis! But also have a question of sorts. . .

The start of game appears to me to be equated to the fuseki, so much so the start of the game in the west is named as such. But is this also true for Japanese pros? Because the meaning of the word (Iwamoto Sensei's seed-scattering go) comes to mind); it makes sense that such plays are common early on in a game, however is it not possible that in the late stages of a kyu match where the boundaries seem decided, one player may plant a seed and have it live. Perhaps this was a fuseki play, an opening play, that opened up new possibilities!

With that in mind, it is easier to conceive that a move which shuts down invasion possibilities is a closing move, closing territory up as you would a bag . It becomes obvious that you could play a yose move near the start of the game as you could play a fuseki move near the end of a game.

But following this line of reasoning, it seems to me that fuseki ends up closer in nature to the word aji, with the same being true for yose and aji-keshi; in both cases, the former could be said to cause the latter.

But that is odd to my eyes, so would you say this is closer to the nature of yose (and fuseki)?

[jlt]

If someone could give an accurate definition of fuseki, I would be interested.

I guarantee you good results after ten thousand games.

Good thing to know that I will finally get results by year 2049.

[Mike Novack]

I think the problem is that the term is used in two senses, "stage of game" and "type of move".

Taking the latter first, means a move that locally affects the final score but has little to no effect on the overall board.

Then we have "stage of game" meaning just moves of this sort left.

Thus a really big yose play might be the best move even though well before the yose stage. Or it might be premature with plays affecting the board as a whole still more important.

[RobertJasiek]

Endgame can be a game phase, characterisation of local position (local endgame, often with boundary settling) or characterisation of those aspects of a play good for local endgames or boundary settling. Long ago, I have overcome a strict phase splitting and prefer to view opening, middle game and endgame to run simultaneously from move 1 on. Computer programs (expert systems and probably neural nets) often do likewise. I would expect similar flexible thinking from pro players.

[John Fairbairn]

The start of game appears to me to be equated to the fuseki, so much so the start of the game in the west is named as such. But is this also true for Japanese pros? Because the meaning of the word (Iwamoto Sensei's seed-scattering go) comes to mind);

The fuseki is the stage of spreading out stones (or groups - same word) in open areas. So in a game that begins with a long variation of the nadare there may be no fuseki. Similarly, a game that starts off with fighting can become quiescent and so fuseki moves can be played in the middle game.

The old term was ishidate - establishing groups. It was the term used, for example, by Honinbo Sansa. Another, rarer one, was ishikubari, which was "spreading groups out." I suspect that fuseki (where seki = ishi and fu = spreading) came to be preferred to both partly because it was the posh Chinese reading of one but partly because it was ever so slightly more in tune with more modern play then the other - you didn't quite "establish" groups by playing on the fourth line, for example, but you did spread your stones and your influence about.

It's interesting that Sinified words for yose exist but are rarely used (e.g. shuusoku, which is apt to come up only in the Autumn when rice is being harvested - for obvious reasons).

The opening in terms of time is joban, and the endgame in terms of time is shuukyoku. Both are Sinified forms and not specially common, but are certainly in constant use, especially joban.

And, as a reminder, a goban is a go board, not a goban, and a jubango is a ten-game match, not a jubango. And don't get me started on kifu and moku...

Iwamoto's "seed-scattering" style should really be "bean-scattering". It's a sly joke on his time as a coffee farmer in Brazil.

[Knotwilg]

If "yose" does not mean "endgame" but "boundary play", then what did Honinbo Shusai mean with a move that is "the start of the yose"?

I like "endgame" and "boundary play". They are clear terms. And early boundary plays may be a sign of an early start of the endgame.

I'm willing to like "yose" if I know what it means, less so if I'm told what it's not.

BTW how different is "a kifu" from "a porte-manteau", "a nacho" or "a zugzwang"?

[John Fairbairn]

If "yose" does not mean "endgame" but "boundary play", then what did Honinbo Shusai mean with a move that is "the start of the yose"?

Japanese does not normally distinguish singular and plural, so yose is "boundary play" or "boundary plays", and as it is a verbal noun it also covers "playing the boundary plays".

I like "endgame" and "boundary play". They are clear terms. And early boundary plays may be a sign of an early start of the endgame.

Or not.

BTW how different is "a kifu" from "a porte-manteau", "a nacho" or "a zugzwang"?

Two ways to start off with. (1) There's a clear (and MUCH clearer) English equivalent - game record. (2) In Japanese go it actually has two meanings - "game record" and "go manual" (and if it comes to that it also means "shogi manual)." Why bring in confusion? If I tell you some kifu with new games by people like Dosaku have recently been found in the papers of Honinbo Shusai (who was also an avid shogi player), which do you think I mean?

Furthermore, the English is portmanteau and it doesn't mean what porte-manteau means in French, which is essentially the point of this thread.

[Knotwilg]

If "yose" does not mean "endgame" but "boundary play", then what did Honinbo Shusai mean with a move that is "the start of the yose"?

Japanese does not normally distinguish singular and plural, so yose is "boundary play" or "boundary plays", and as it is a verbal noun it also covers "playing the boundary plays".

I like "endgame" and "boundary play". They are clear terms. And early boundary plays may be a sign of an early start of the endgame.

Or not.

BTW how different is "a kifu" from "a porte-manteau", "a nacho" or "a zugzwang"?

Two ways to start off with. (1) There's a clear (and MUCH clearer) English equivalent - game record. (2) In Japanese go it actually has two meanings - "game record" and "go manual" (and if it comes to that it also means "shogi manual)." Why bring in confusion? If I tell you some kifu with new games by people like Dosaku have recently been found in the papers of Honinbo Shusai (who was also an avid shogi player), which do you think I mean?

Furthermore, the English is portmanteau and it doesn't mean what porte-manteau means in French, which is essentially the point of this thread.

Well, the argument is weakened by the fact that I'm not a native speaker, but the point stays: people do borrow words from other languages if they think that is fashionable, effective, enriching, being inclusive. And these words then tend to drift from their original meaning. I do acknowledge that this can be irritating to the knowledgeable spectator.

As it happens "a game record" may, to a non-native speaker, suggest that the Guiness Book will soon be involved. Both kifu and record have multiple meanings in their language. How confusing that is, depends on the education of the user.

But again, point taken.

(Totally unrelated, only this week I came to realize that Mick Jagger didn't abuse the English grammar when he sang he was waiting on a friend.)

[Kirby]

Maybe it’s more of a Japanese misunderstanding than a go misunderstanding? Using the “boundary play” term could be helpful.

As I implied earlier in the thread, the whole topic here is pretty much a language/classification issue: some people use the word "yose" to mean something different than what is often meant in Japanese. Given this, to avoid confusion, I agree that a new term makes things easier and less ambiguous.

That being said, borrowed words in foreign languages naturally take on a life of their own. Korean has a class "konglish" words, which are basically borrowed English words, which have (sometimes) taken on their own unique meaning. 핫도그, for example, is borrowed from "hot dog". But if you order a 핫도그 in Korea, you'll more often than not be getting what would more aptly be called a "corn dog" in English. In the interest of facilitating communication between Korean and English speakers, one might argue that Koreans use 콘도그 (sounds like "corn dog") rather than something that sounds like "hot dog". After all, this would bring better parity between the English and Korean words.

But on the other hand, there's nothing necessarily *wrong* with using a word that could cause confusion to English speakers, when speaking in Korean. It's a different language, after all, and they can use borrowed words however they'd like. In the same way, there's nothing necessarily *wrong* with the English word for "yose" to carry a different meaning than the Japanese ヨセ (yose). After all, English and Korean are different languages.

But it's a little less confusing for everybody if we agree on common terminology, given the chance - the main point is effective communication, and if we can do something to improve that, why not?

[EdLee]

Can someone explain the etymology of ヨセ ? Thanks.