Having returned to the active Go scene in 2011, I started documenting experiences again at Sensei's Library, using the position search function there and the one at Gobase to study patterns and techniques. I'm also looking into acquiring GoGod and using it with Kombilo and/or Drago.

When I first learnt about gobase, back in 2003 I think, I was intrigued by the fact that a go research database existed in the Western go world and apparently not in the Eastern one. I'm sure digital pro games databases must exist over there too, but apparently they're not open. There is a faculty of Baduk in Seoul but I know of no way to see their research results (a general issue with universities I feel - their investigations remain reserved for internal use).

Back to 2011, I'm rather disappointed with the degree of activity of professionals and amateurs in go research. Wikipedia has succeeded in popularizing human knowledge and keeping it fresh (in my opinion) but one cannot deny that Sensei's Library, "the" wikipedia for Go, has stalled. The community is dispersed over a couple of forums, but even then it's rather absurd that opening patterns are conclusively discussed by amateur 2d like me, while there are probably hundreds of thousands in China, Japan & Korea who are stronger.

Is Chess research similarly stalling or is it better and if so is that due to it being a Western activity, while Go research is more reclusive being an Eastern activity?

In short, why are we not seeing intense go research on the internet by a few dozen 6d-7d?

Because playing a game is more interesting than typing up what you already know into a computer.

i am sure research is going on, at least in study groups of pros and strong amas in Asia. but it seems that their discussions and conclusions are not recorded and/or online

part of the research also may be done privately by single pros and they don't publish their results for obvious reasons - to use them to get an edge in few important games until the innovations get adopted by other competitors. i suppose this is only minor part of research, but i like the idea, like when Akaboshi Intetsu played the secret taisha variation of the Innoue house in his game against Honinbo Jowa

another reason might be that strong players can teach their knowledge for a fee, so they don't want to make it free online (reminds me of a 5d who jokingly said that he was also adding mistakes into his lectures, so that he had always something more to teach)

and when Cho Hye-yeon's blog was active, she had written a lot her thoughts about the game, various openings and counter-strategies. pretty much what you are talking about

_________________
Spilling gasoline feels good.

I might be wrong, but probably not.

Presumably you are referring to Myongji University. I can't remember for certain, but I think the baduk faculty is not in Seoul but in Yongin. However, it has nothing to do with the sort of research you seem to mean. Its goals are:

1) To promote [amateur] students' Baduk strength above the 5 dan level

2) To acquire a deeper knowledge of life from Baduk culture

3) To develop and educate the leaders in the field of Baduk

4) To teach necessary foreign languages to those who can introduce Baduk culture to the world

The research results are essentially limited to those who do postgrad work, but the theses are usually available (usually in an Oriental language but some of the work gets disseminated in English through the ICOB seminars).

Today, the best example of the sort of research I think you mean is probably the Ch'ung-am Brigade in Seoul, for alumni of a posh school there, but there have been very many such groups throughout the decades and the GoGoD CD gives details of all the most important ones. Pros normally keep their results to themselves, but there have been several cases of amateur groups publishing their research.

Read everything that has been published in 2011 and you can find a lot of new research results. Example:
http://senseis.xmp.net/?JosekiVolume2Strategy
Among a dozen major inventions, it defines influence and thickness precisely in general! There is no reason for you to be disappointed.

I did a quick search on http://www.infoseek.co.jp for go research ( 碁　研究 ), and got a lot of hits. There are quite a number of research groups, many in universities. There is even a research group for Hikaru no Go.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

Here is some go research:

http://ci.nii.ac.jp/naid/110006407445

Move Distribution and Pattern Frequencies in Tsume-Go [in Japanese]

Abstract:

In this paper, we have investigated some of the statistical aspects in Tsume-Go based on problems taken from a major Tsume-Go database. In particular, we have focused on the distribution of moves and the frequencies of patterns. First, we present the distribution of moves related to each point on the Go board. Then, we have applied these results onto the Tsume-Go problem and examined to what extent the problem can be solved using this information. Second, we have examined the frequencies of certain patterns in the Tsume-Go database. Since each pattern also contains a recommended best move, we have tested to solve the Tsume-Go problem by applying these moves. At the time of writing, these tests are in progress and final results will be presented during the seminar.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

There was a time when I tried to keep up to date with university type go research - the result of visiting places like the Tsukuba campus and company laboratories to talk about shogi programming rather than go, but there was some overlap.

However, titles like "Estimating the possible omission number for groups in go by the number of n-th dame" seemed like a perversion of the game to me, and I soon lost interest. Still, there was some more accessible work. The foundation of the article cited by Bill above is much-cited earlier work also by Yoshikawa and Kojima in which they tried to set up a model for pattern acquisition using lots of game records. The result, as I recall, was a progam that had about 10,000 patterns based on lozenge-shaped areas of the board, and there was a set of statistics associated with each pattern that gave the likeliest moves for each side in that pattern. A program called (I think) Katsuya was developed in the 1990s based on this idea - positions where no matching pattern was found used the most similar one, I think. I have no idea what happened to Katsuya.

But Yoshikawa and Kojima wrote another paper which was even more accessible - it was on the relationship between skill and the use of go terms. They found there was a definite relationship but it seemed unclear which way the cause and effect went. It was of special interest to me because the method of the experiment was first to listen in to comments made by individual players (they called this a go protocol: you may recall Bruce Wilcox doing something similar), but then they used consultation go (amateurs) in the same way, and finally, to get a pro perspective, they used the two books on the famous consultation games involving Go and Kitani. This still is of special interest to me, you'll understand, having since produced a book on one of these games (Consultation Go, by Slate & Shell, now in exciting techno e-book form !!!!). I don't think Y & K discovered anything startling, but one conclusion I recall was that as players become stronger they are more likely to use terms in an elliptical way, i.e. they assume a shared understanding or empathy with other players. (I must say, wryly, the brouhaha and lack of empathy over "joseki" recently here comes to mind).

A further paper I recall strongly from that era (1990s) was one which categorised go terms in a hierarchy. This struck me then - and still does - as a major insight that ought to be developed, both for programming computers and for teaching humans. I'm afraid I've forgotten the author and the paper is somewhere up in the loft, and I can't be fashed to hunt for it. But I'd be interested if anyone else knows what happened to this line of research.

Incidentally, the abstract for the paper cited above has a slightly misleading (?mis)translation: "problems taken from a major Tsume-Go database". The paper is in Japanese and actually refers to "a large database", which is actually about 10,000 or 12,000 problems taken from almost 100 very standard problem books. As far as I know the collection is not available as a publicly accessible database. Having all or most of the books used, I can say that there are many duplicate problems (which may have affected the results unless there was some unmentioned procedure to screen them out), so the number of distinct available problems from the standard literature (if I've remembered 10-12,000 correctly) may be about 8,000.

All I'm really saying here is that I can confirm Bill's point about there having been lots of academic research done, and at least some of that was done in Japan using a database of go games over a decade ago. The GoGoD database has also been used for the same purpose by some western researchers.

As perverse as saying "Black is ahead by 3.5 points? More numbers are the future! It is very meaningful to say "Can play elsewhere 7 times.", in general n times.



I do not think it is that exciting. Having a possibility to play through variations distracts from what a book's author wants to emphasise and tempts the reader into avoiding the exercise of reading ahead.



Well known terms can be abbreviated when confusion is unlikely. Otherwise my experience is a more detailed vocabulary with increasing strength. Greater strength requires a greater variety of represented contents.



Why. My vocabulary (in books or research papers) enjoys being structured hierarchically.

Thanks Bill and John to relate to the question. So we may assume there is indeed going on some Go research, both at the typical academic level of specialty that is of interest to none but the authors and one respected member of this community, or at the more mundane and holistic level.

When I took lessons with Minue 6d, I was struck with the volume of knowledge that he apparently drew from the Korean Go scene, knowledge that is not available in any textbook I know and that I can also not derive from pro games because it's about standard patterns in the middle game that punish bad technique or deal with an attack in a certain way. Much of these leaning, sabaki or other techniques are yet undocumented. But it went deeper than that: Minue seemed to have a theoretical foundation for these techniques, rooted in as basic proverbs as "straight is better than diagonal", which I have tried to get to in my SL series on Go Theory, obviously failing miserably due to my very modest level.

John, rest assured that your contributions about joseki or atsusa are resonating with me. I'm having an occasional revelation about the dynamic nature of the game. I think it is precisely this dynamic nature which is at the root of the issues with translation go concepts into the Western language. But now I've ranted off topic.

Not only have you seen the situation clearly, I think you may have given me a blinding insight - not so much in the quote above, but in your word "holistic".

I wouldn't want to make too much of it at this stage, as I may come up with counterexamples or other angles myself, but I toss some early impressions out for discussion. The reason I went off on the track I am about to describe is that as I read "holistic" I was sitting facing a bookshelf on the other side of the room where the book nearest to me is a "Practical Encyclopaedia of Go" which is edited by a Lu Guoliang, but the main technical contributor appears to be the inspired and inspiring Zhao Zhiyun. He was a go historian and journalist and also a 6-dan pro. He died regrettably early in 1993.

This encyclopaedia is the epitome of holism. It has over 700 pages and covers every aspect of go in great detail, all under the same cover. What dawned on me was that the Oriental countries all have this "holistic" kind of book, but also their magazines are equally holistic, and you could even say that their playing environment is equally holistic - a wide range of players of different grades available to play most of the time.

Whilst that may not seem very remarkable, my first impression is that it is a stark contrast with the western go world. Obviously we haven't got the encyclopaedias. Although we now have a lot of books in English, the vast majority, maybe even all, cannot be described as holistic, as they deal with separate parts of the game. Even a very long running and excellent magazine like Go World cannot be called holistic. It is almost exclusively devoted to commentaries. Even if you had a large library of English books, you will be hard pushed to say that the combination of all them - taking out all the repetitions and mistakes - even approximates to what is available in the above encyclopaedia. Until very recently, as well, it was very hard for any western players to get anything like a holistic playing environment, with teachers and players of different grades always on tap. For most of us of a certain age, it was damned hard just to get a game for most of the year.

I think this has also become something of a vicious circle. At least, even here I see people talk about deciding to studying joseki, or the endgame, or life and death. All fine things in themselves, but not holistic. It may seem that doing a bit of A, then B then C amounts to a mixture of ABC, but I'm not so sure. I'm beginning to believe that there is a undervalued synergy in doing A and B and C together in a way that is not really possible here. It's possibly not just convenient but hugely more beneficial to have a book, as above, where you can read about josekis and turn immediately to a section on life and death problems that result from josekis, or to a section on counting the resulting endgame plays, whenever you have a query. Being able to play every day surrounded by stronger players who laugh at you because you didn't count liberties is chastening but valuable. This ("five alive") is something that's so basic that it rarely gets mentioned in books, but it is part of the "street" knowledge in the Far East (though not here, in my experience). We don't even have a street, let alone street knowledge.

I think the gappy western approach also leads to distortions. For example, it is my judgement that, as a percentage of their utterances about go, western players say "joseki" far more often than Orientals. That may not necessarily be a bad thing, but I suspect it militates against a holistic approach.

Over the years I have spent a lot of time trying to reveal to the western world that there are go concepts that they haven't considered yet (either at all or properly). Though I still believe this was necessary and, I hope, useful, I'm beginning to wonder if even this is more atomistic than holistic. It's just providing more ingredients for the mix, but the finished cake still needs a cook's expert hands.

Even if the above analysis proves to be a reasonable summary of the situation, it doesn't directly offer a solution, of course. But (so long as minds are kept open and those tempted to try to apply killer logic can rein themselves in) L19 itself does at least provide the glimmerings of a holistic way of talking about go.

Apparently you do not have any idea of what "great detail" actually is. I am writing a first volume on semeais and my current estimate is: About 600 - 700 pages are needed for a great detail, that is reasonably complete treatment of only basic semeais between exactly two essential strings, without kos and without non-basic shape defects. Therefore an encyclopaedia with only 700 pages cannot be both holistic and have great detail. It can only be an overview with some details.



I will contribute to filling more gaps in Western literature. During that process, it will become easier to write overview-style encyclopedias.



The what? Not the approach is gappy but time has not proceeded long enough to fill all major gaps of Western / world-wide literature. Be patient!

Robert, I leave in the middle whether your approach and standpoints are correct, but you seem to have missed the point of John's statement. If you are going to provide a detailed or even exhaustive treatment of semeai and need 700 pages for it, that's fine. It is still a "western" approach of splitting knowledge into separate components, treating these components "in detail" (or not) and somehow relying on these treatments to lead to a unified improved knowledge.

John, in his enlightenment about holism, challenges that assumption.

As far as I know, this particular dichotomy in teaching has always existed in any field, particularly in teaching "skills". As much as analogies may distract, you can write tons of books on how the different parts of a bicycle function and need to be operated without any reader being necessarily capable of riding the bike afterwards. I won't go too far into analogies, but I can assure you that in the field of guitar playing there is just as great a battle between the "holistic" teachers who think the technique and the music go together and the purist technicians who believe the technique can be developed perfectly dissociated of any actual music. And recently we have even seen a revival of holism in the fields of mathematics, where algebra without geometry or geometry without algebra are regarded by scholars as different ways of being blind. (see Brett Victor's essay "Kill math")

What John seems to advocate as a beneficial characteristic of oriental knowledge transfer of go, is to resist decomposition and always treat a particular aspect in the larger view. There is no such thing as joseki without L&D-implications (sure) but even not without endgame implications. There is no move that exists without the global context. There is no opening pattern without an already inherent implication on the endgame. There is no capturing race that doesn't include the flavour of furikawari or ko. Etcetera.

Interestingly, John expands the idea of being "rich in context" to players as well. Certainly a player in a thriving club atmosphere is not the same player as he would be without. I didn't think of that kind of holism but the point is amusingly true.

Now Robert, it may be that Westerners are who they are and NEED to teach and learn in a deconstruct-reconstruct process. I personally don't think it's genetic but surely the idea of analysis is culturally rooted in the European renaissance and has provided so many successes in the scientific field that it has become second nature to us. And maybe it is, as you say, a matter of time and ardent effort of the Jasieks before analytic methods prove their value in the field of go theory too.

No, but he has overstated it and understated existing Western go books with non-specialising approach.



Avoiding decomposition leads to overlooking many detailed facts. Both is useful: the detailed study and the connecting overviews.



Trying to understand all semeais by starting with already including exchanges and kos is too difficult! Many have tried and failed. There are so many types and cases that even I, who enjoys generalisations, do not dare to make them without having studied the details carefully. A solid and safe approach is to study the details first and then make the generalisations and connections to other topics and yet broader contexts.



Countless proofs are already there. Maybe you mean something different though: analytic methods have not taken the lead in practical Go decision making yet.

Here is an extreme example of what I think John has in mind. (I critiqued the paper at the time.)

Click Here To Show Diagram Code

[go]$$ What is White's possible omission number?
$$ --------------
$$ | . . O . X . .
$$ | . . O . X . .
$$ | . . O . X . .
$$ | O O O , X X .
$$ | . . . X . X .
$$ | X X X . X X .
$$ | . . X X X . .
$$ | . . . . . . .[/go]

How many dame can Black fill before the next one is sente?

This is a trivial example for most go players. And we just say that White is alive. However, it is possible to construct a position where the PON of this group matters.

OC, a group's PON is more likely to matter when the position is more open. And when more than one group is involved. If you can make a play that reduces the PON of two groups to 0, you have a good chance of winning.

One thing I noted when I first started studying pro games was how often, in the opening, they would extend or bolster a live group. One point that some commentaries mentioned was that the opponent had a sente play against the group. IOW, the group had a PON of 1. Obviously, small PONs are important.

On the other side of the coin we have the proverb that says that large groups are alive. OC, that is not always so, but another way of putting it is that they have large PONs. Their very size means that typically they have many potential places to make eyes, and many dame in case of a semeai, thus a large PON.

The authors' estimation method may allow beginners (or computer programs) to recognize large PON without having to read variations. (The proverb does the same.) Yes, such estimates are fallible, but they make the game easier. I don't think that that perverts the game.

Several years ago, before the Monte Carlo advance in computer go, I played against a program that incorporated Thomas Wolf's tsumego module, which is extremely strong. To my surprise, the program was rather weak in actual life and death in the game. The reason, as the programmer explained to me, was that using Wolf's module to read out life and death was relatively slow, and so it was called only when the main program thought that it was worthwhile. I wonder if that program could have used the PON estimation technique to make better decisions of when to call the module.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

I don't think so. It is not just the non-specializing approach. It is the non-specializing approach combined with depth and (presumably) some systematicity. The latter two are where Western books do not match their counterparts. Books like Lessons in the Fundamentals of Go are non-specific, but are not systematic. Systematic books (which seem to be most often for beginners) are usually not in the depth you can get from an encyclopedia like John mentions.

_________________
Occupy Babel!

I am reminded of my first go research paper, delivered at a small conference, with maybe a couple of dozen in the audience. Afterwards I quipped that there were more people in the audience than readers of my paper. (Actually, it has gotten more citations than I thought it would. )

My research tends to be abstract, so that it may apply to various go situations. Other research, such as in tsumego or joseki, tends to apply to more specific situations. And then there is the research of Go Seigen. He has written five volumes in a series on his research on recent games.

http://www.amazon.co.jp/呉清源-最新打碁研究%E3%80%885〉-呉-清源/dp/4416507100/ref=pd_sim_b3

Vol. 5 on Amazon, Japan. I expect that his is the most holistic research of all.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

On the question of holism, at least in the context of learning go, my favorite method of study has always been pro games. They have everything. At least implicitly. Shusai Honinbo said that playing over 1,000 pro games was enough to advance to pro level. (I think you also need talent. )

I know that many strong players preach tsumego, tsumego, tsumego. Not to underestimate its importance, but there is more to the game, IMO. There are many specifics about the game that are worth studying. But I think that the best place for the holism that Knotwilg and John are talking about is in pro games.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

Bill, concerning PONs, they appear to be a special case of the more general n-alive, which allows also negative n for the dead cases. (Positive or zero: attacker first. Negative: defender first. *-alive = pass-alive. Similarly: n-connected.)

Shusai / pro after studying 1000 pro games: In those days, people spent much more for a single game than nowadays. Even Go Seigen spoke of just 20 minutes per game. With more available games, this has become an option:)

Re: hyperpape: Of course there are systematic books or other overviews with less depth than supposedly in the encyclopedia mentioned by John. This does not imply "great detail" though, maybe it implies relatively "greater detail" than books you describe, but this is a much weaker statement than an absolute "great detail".

Robert, I think that you have missed your vocation. You should have been a film critic. Of course, you wouldn't have to watch any of the films; just comment on what other people say about them.

Best wishes.

_________________
No aji, keshi, kifu or kikashi has been harmed in the compiling of this post.
http://www.gogod.co.uk