For amateurs, and it would seem pros, boundary play theory as exemplified by Robert is massive overkill, and has more to do with mathematics than go.
The very fact that it is difficult to trace thinking about boundary plays in the Oriental literature tells us a lot about the proper priorities. We know that de-iri counting was known before 1844, because that was when Genan published what he claimed was the first book on counting boundary plays. He did not use the term de-iri, which may not have been used until the early 20th century, and he did not claim to have invented the concept (despite being brash enough to publicise himself) and he did not feel the need to explain the counting method. All this strongly suggests the method was already well known then, and since the endgame play of Edo players is regularly praised even today we may safely assume it already had a long history. That is, there had already been ample time for refinements to emerge.
But when we look at Genan's work (which I have long made available in the New In Go package that comes with the GoGoD database), we see the rather simplistic model that we are familiar with even today. Genan essentially uses three criteria to evaluate a boundary play: de-iri points, sente/gote, and (as an optional third element) incidentals such as aji, ko threats.
Nowhere does he speak of double sente or reverse sente, but he clearly understood at least some of the niceties of the term sente that Bill regularly refers to. For example, at one point he says a move is "worth about 9 points and
can be regarded as sente" (my emphasis).
A few times he does occasionally mention half-point fractions, but feels no need to go into sixteenths and so on, and usually he is satisfied to say a play is worth "about" X points. This very simple model was clearly all that the Edo masters needed, and over the many years since then it has become apparent that it was sufficient for generations of later pros, who never (with a straight face) give obscure fractions in game commentaries - they will always say, like Genan, a move is worth "about" X points and, again like him, may occasionally mention half points.
But after Genan we had to wait 80 years before boundary plays were properly treated again in print. This is despite the minor explosion in go journalism in the late 19th century. Instead we had fusekis, josekis, tesujis - but no yose. Again, this almost certainly tells us something about what pros such as Honinbo Shuho thought about priorities.
In the late 1920s, the likes of Kato Shin and Kubomatsu Katsukiyo (who did not use ghost writers) started presenting series on boundary play counting in Kido. The articles were not as well structured as they would be nowadays by a professional go journalist, but essentially they were just presenting what Genan knew.
And by and large that has been the same situation ever since. Periodically, articles of the type Kato and Kubomatsu wrote have been recycled, and have occasionally made it into books. Sometimes you get a whole book on boundary plays, but when you examine it you see that it may be ten pages of "theory" and 230 pages of problems. The only real difference is that the more modern books are apparently written by amateurs. We can say this, not with complete certainty, admittedly, because we see characters and allusions in the text of a level that requires education at an elite university supposedly written by a pro who left school early to devote himself to go. And after the war, we see magazine articles about esoteric technical details of the endgame but these are
always by amateurs. For example, Sakaguchi Junei (a very strong amateur) introduced the concept of miai counting to Kido readers in 1955. We also see amateurs' articles about fractional followers and corridors and the like. There had been a minor tradition of amateurs writing about the mathematics of go even before the war, but in the main these were related to the rules of go.
Robert has claimed pros have kept their secrets hidden - the bounders! I think rather that this simply confirms the picture of boundary plays as something that pros did not fuss about very much. They were apparently content with the simplistic model used by Genan and we can infer that they thought we amateurs should be content with it, too. After all, there is a strongly practical reason for thinking that way. Boundary plays tend to come at the end of the game, when time is short. Even pros find it hard to calculate multiple positions using fractions in one-minute byoyomi (and byoyomi can be much shorter!), so they rely on "tricks" such as "that obviously cancels this out so skip the calculation" and they accept the theoretical but highly remote possibility that they may lose a game because they don't quite know how to finesse the fractions. This is exactly how many chess players treat the endgame, too. Quite a lot of pros don't know how to play the K+B+N vs K endgame and so have, on rare occasions, ended up with an embarrassing draw instead of a win. But they have correctly reasoned that they may never see such an endgame in their entire lives.
In fact, in many endeavours, a characteristic of the pro is that they are eminently practical and a characteristic of the amateur is to fuss about trivial details.
Then O Meien came along. He introduced to Japan a new way of treating boundary plays called absolute counting. He said he was surprised Japanese players did not use this superior technique, but it was normal in China (I think he meant what we would call Taiwan). Apart from the novel method, his book was remarkable in that he clearly wrote it himself. Another important feature is that his method, once understood, is very easy to apply. In actual play O Meien does not fuss about sixteenths, corridors or other esoterica any more than his Japanese colleagues do. It's just that his method (or perhaps better: approach) cuts through the confusion caused by de-iri and miai counting (caused by overzealous amateurs?) and also overcomes the delusions that Bill mentions often, such as sente being worth something.
Now Robert says this about O Meien's book:
You only understand this good book if you read Japanese or already know all the theory explained in this book. It teaches the very basics of modern endgame theory for local endgames and the most basic global decisions. Endgame 2 - Values teaches more, more details and also the microendgame, scoring and school mathematics but avoids global decisions before the microendgame because they will be the topic of Volume 5. O Meien's book is well worth reading but non-essential if you read Endgame 2 - Values. On the other hand, modern endgame theory has been neglected in the other literature so reading both books can further improve one's understanding.
I have quite a few problems with this. Despite the implication, Robert does not read Japanese. I do read the Japanese, and actually have translated the book. I did this to practise my shorthand (getting very rusty in retirement) and as a way of forcing myself to think about the book. I normally just speed read go books.
So, if Robert cannot read the book, how does he know he already knows all the theory explained in it? His further comments suggest to me he doesn't. For example, the book does not teach much about the very basics of modern endgame theory, because the whole point of the book is to stand up against modern theory as exemplified by what Japanese players use. He barely touches on the very basics such as how to count de-iri style because the basic tenet of his book is about changing attitudes. On top of all that, the climax of the book is the use of a formula for making not "basic global decisions" but rather advanced global decisions to do with winning the game. I see no sign in Robert's self-review that much of what O says appears in "Values", so to say "Values" makes O's book non-essential is, well valueless.
My own view of O's book can be inferred from the fact that I bothered to translate it (nb for myself - don't ask). But I'll be explicit. It can be a slightly irritating book because O is not a born technical writer, but it is not just "good" but excellent. It is the best yose book for amateurs available (by a long way) because the method works. It retains the simplicity and sensible prioritising of previous Japanese books (amateur ghost writers' pages on sixteenths excepted), so it is easy to learn. It is sufficient in itself. OK, it is in Japanese, so you may have to make do with inferior English works, much of the content of which is made redundant by O's insights. But that's hardly O's fault.
)
