Re: Yose = endgame? No way, Jose!
Posted: Fri Feb 08, 2019 7:44 am
It can be an uphill struggle to get a discussion going here and so I often tend to be provocative. But usually there is a serious underlying intent. Those who wish to be VERY serious about yose may care to look at my article on kechi (the forerunner of yose) in The Go Companion. There I note how even the famous translators have made a mess of the go passages in Tale of Genji and other Heian classics. Their misunderstanding of kechi is one such case.
But continuing in serious mode here for a change, I will make some other observations about go terms in general.
As I have already mentioned, I am close to completing a book on the games between Genjo and Chitoku. There are almost 90. As usual I have collected a vast number of pro commentaries and synthesised the results. But on this occasion I have gone one step further. I have added an appendix, which I call "Go Wisdom." Here I discuss a very large number of go terms, quite a few of which will be new to an English-speaking audience. I do not so much define them (I am already bumping up against the maximum page limit) as discuss them (e.g. list how and in what contexts they are used) and I also list proverbs associated with them. The idea is that as the readers read through the games, if thoughts occur to them about certain references in the commentaries (e.g. to momentum) or about moves in the game that are not even covered by the commentary, they can turn to this appendix and be supplied with a set of memory joggers or suggestions for further thoughts. I combine this with a new format for the commentaries themselves in the hope that it will stimulate self-study and also embrace all levels of go strength.
As a refinement to the appendix I have added an eclectic set of cross references to each term in the commentaries. The idea here is that instead of an individual game you can study a concept by referring to the wealth of examples in many games. You may find, say, definitions of thickness hard to understand but once you've looked at over 120 examples, plus explanations of their significance in the commentaries, your neural network rather than short-term memory will take over - and we all know how powerful that is in go!
Some spin-off results of this exercise were interesting to me.
I have long noticed that the way westerners talk about the game is rather different from the way it appears in Japanese books. I even ran a thread called (I think) the Big Game on rec.games.go which explored this phenomenon.
There are two aspects to this. One covers written texts (I.e. translations). For example, the Japanese seem to make use of many more terms. This is probably largely to do with the effects of generations of translators which have all used different English terms for the single Japanese term. This dilutes the term so much that it can make it look as if there is no concept there at all (e.g. choshi, ijime). But there are also cases where the Japanese simple does have inherently more nuances, which are lost either by using the raw Japanese term in English (e.g. yose, aji) or by using Emglish terms that carry their own range of very different nuances (e.g. influence, trade, forcing moves).
The other aspect relates to speech or other informal discussion. Here the obvious difference is that the Japanese texts are provided by pros. English original texts are mostly provided by amateurs. We should expect a stark difference, and one way in which this shows is that the pros make much more use of certain terms than amateurs do.
For example, the two most common terms in G-J, i.e. by pros, were thickness and forcing moves. (Just in passing, sente kikashi as opposed to just kikashi is - or was - a term beloved of British amateurs but does mot occur even once here; they also use sente much more, instead of forcing moves).
Another very common term was "settling", which was as common as tesuji. I suspect the proportion in amateur usage would be more like 1:100 rather than 1:1.
A further high scorer was "order of moves." I suspect this scores highly because of the pro obsession with efficiency of moves, and I'm sure that is also an explanation of the very high score forcing moves (timing = efficiency).
Erasing/reduction scores high, which seems interesting in the light of AI games.
A couple of other examples, less frequent but perhaps eye-opening for amateurs: heavy" - as early as move 9; kamae (construction) where amateurs might say moyo. But there was no example of jimoyo, which seems to be used only of modern games.
A further noticeable pro-am difference is nuances. I referred earlier to nexus go theory. That is my own term. When I was working on shogi computers (our British team led by David Levy produced the first programme to beat a pro, on four pieces), I developed a nexus shogi theory and in the course of visiting Japanese researchers in Tsukuba, I learned about a somewhat similar approach to go by a Japanese academic. The main difference was that his groups of concepts had to have a clear hierarchy. That proved too difficult to program, I gather. It was easier to handle a nexus in shogi because at an early point depth of search and material gain made up for hierarchical imprecisions.
Pros seem to think in nexus terms. They may not realise that, of course! So where we talk of running battles, or at most leaning attacks and twisting attacks, the pros talk about aori, seriai, karami, motare, semedori, torikake, etc. Or forcing moves, momentum, ersure, probe. Or aji, semeaji, aya, fukumi, te ga aru. Or settling, bases, kamae. The difference is not just that between a vintage claret suffused with overtones and a bottle of plonk. It is also that a term within one nexus can also apply within a different nexus, and so you can get a nexus of nexuses, each informing the other. Like a good red wine enhancing the taste of the beef. This rich network of associations (rather than the actual terms) surely leads to greater range of vision and creativity.
Does this richness of terminology help produce stronger amateurs? Does a lack of richness handicap us? I suspect it does make a big difference of amateur level, and of course even pros have to start at amateur level. But maybe, leaving strength aside, the biggest difference is to amateurs' understanding enjoyment of the game as fans. Think petrolheads vs F! racers.
Of course I can only speak in impressions, though they do cover over 50 years of reading about go. But this little index has hardened up my main impression: terms matter, but not necessarily as ways to get stronger - unless we are also willing to spend the vast amount of time needed to form neural networks of the associations they represent.
But continuing in serious mode here for a change, I will make some other observations about go terms in general.
As I have already mentioned, I am close to completing a book on the games between Genjo and Chitoku. There are almost 90. As usual I have collected a vast number of pro commentaries and synthesised the results. But on this occasion I have gone one step further. I have added an appendix, which I call "Go Wisdom." Here I discuss a very large number of go terms, quite a few of which will be new to an English-speaking audience. I do not so much define them (I am already bumping up against the maximum page limit) as discuss them (e.g. list how and in what contexts they are used) and I also list proverbs associated with them. The idea is that as the readers read through the games, if thoughts occur to them about certain references in the commentaries (e.g. to momentum) or about moves in the game that are not even covered by the commentary, they can turn to this appendix and be supplied with a set of memory joggers or suggestions for further thoughts. I combine this with a new format for the commentaries themselves in the hope that it will stimulate self-study and also embrace all levels of go strength.
As a refinement to the appendix I have added an eclectic set of cross references to each term in the commentaries. The idea here is that instead of an individual game you can study a concept by referring to the wealth of examples in many games. You may find, say, definitions of thickness hard to understand but once you've looked at over 120 examples, plus explanations of their significance in the commentaries, your neural network rather than short-term memory will take over - and we all know how powerful that is in go!
Some spin-off results of this exercise were interesting to me.
I have long noticed that the way westerners talk about the game is rather different from the way it appears in Japanese books. I even ran a thread called (I think) the Big Game on rec.games.go which explored this phenomenon.
There are two aspects to this. One covers written texts (I.e. translations). For example, the Japanese seem to make use of many more terms. This is probably largely to do with the effects of generations of translators which have all used different English terms for the single Japanese term. This dilutes the term so much that it can make it look as if there is no concept there at all (e.g. choshi, ijime). But there are also cases where the Japanese simple does have inherently more nuances, which are lost either by using the raw Japanese term in English (e.g. yose, aji) or by using Emglish terms that carry their own range of very different nuances (e.g. influence, trade, forcing moves).
The other aspect relates to speech or other informal discussion. Here the obvious difference is that the Japanese texts are provided by pros. English original texts are mostly provided by amateurs. We should expect a stark difference, and one way in which this shows is that the pros make much more use of certain terms than amateurs do.
For example, the two most common terms in G-J, i.e. by pros, were thickness and forcing moves. (Just in passing, sente kikashi as opposed to just kikashi is - or was - a term beloved of British amateurs but does mot occur even once here; they also use sente much more, instead of forcing moves).
Another very common term was "settling", which was as common as tesuji. I suspect the proportion in amateur usage would be more like 1:100 rather than 1:1.
A further high scorer was "order of moves." I suspect this scores highly because of the pro obsession with efficiency of moves, and I'm sure that is also an explanation of the very high score forcing moves (timing = efficiency).
Erasing/reduction scores high, which seems interesting in the light of AI games.
A couple of other examples, less frequent but perhaps eye-opening for amateurs: heavy" - as early as move 9; kamae (construction) where amateurs might say moyo. But there was no example of jimoyo, which seems to be used only of modern games.
A further noticeable pro-am difference is nuances. I referred earlier to nexus go theory. That is my own term. When I was working on shogi computers (our British team led by David Levy produced the first programme to beat a pro, on four pieces), I developed a nexus shogi theory and in the course of visiting Japanese researchers in Tsukuba, I learned about a somewhat similar approach to go by a Japanese academic. The main difference was that his groups of concepts had to have a clear hierarchy. That proved too difficult to program, I gather. It was easier to handle a nexus in shogi because at an early point depth of search and material gain made up for hierarchical imprecisions.
Pros seem to think in nexus terms. They may not realise that, of course! So where we talk of running battles, or at most leaning attacks and twisting attacks, the pros talk about aori, seriai, karami, motare, semedori, torikake, etc. Or forcing moves, momentum, ersure, probe. Or aji, semeaji, aya, fukumi, te ga aru. Or settling, bases, kamae. The difference is not just that between a vintage claret suffused with overtones and a bottle of plonk. It is also that a term within one nexus can also apply within a different nexus, and so you can get a nexus of nexuses, each informing the other. Like a good red wine enhancing the taste of the beef. This rich network of associations (rather than the actual terms) surely leads to greater range of vision and creativity.
Does this richness of terminology help produce stronger amateurs? Does a lack of richness handicap us? I suspect it does make a big difference of amateur level, and of course even pros have to start at amateur level. But maybe, leaving strength aside, the biggest difference is to amateurs' understanding enjoyment of the game as fans. Think petrolheads vs F! racers.
Of course I can only speak in impressions, though they do cover over 50 years of reading about go. But this little index has hardened up my main impression: terms matter, but not necessarily as ways to get stronger - unless we are also willing to spend the vast amount of time needed to form neural networks of the associations they represent.
