Unfortunately, books specifically teaching reading per se are essentially non-existing. There are quite a few somewhat related Asian books which a cute reader can also use to learn about reading; their example discussion studies variations sufficiently exhaustively so that the reader can follow the move decision-making for every particular example. Sorry, but ATM I lack time to dig out the books and their ISBNs, except for Report of Choong-am Research, for which you (if you are a dan player) find the ISBNs here: http://home.snafu.de/jasiek/isbn.html

In English, there are at least the following short parts of books about reading:

Quotation reference:
viewtopic.php?p=172162#p172162



It is useful but the introduction teaches mainly only one aspect of reading and by only one example: iteration. The other four basic reading principles are explained in First Fundamentals, chapter 10, which has 14 pages. Then there are 3 pages about the reading method Local Move Selection in Joseki 1 Fundamentals.

This is pretty much everything about reading in the English literature. (I know, there are many related topics, but they all do not really teach how to think for the sake of reading well.)

The next closest topic of perception in The Happening Theater of Baduk, ISBN 89-86555-73-5, Korean book contributes a bit to giving an idea about what to think.

The good news is that I plan to write a couple of books on reading, but the bad news is OC that this can need a few years because lots of topics are in my pipeline of books I wish to write.

Currently, reading belongs to the most important fundamentals and to the least covered topics in the literature. The contrast of demand and supply could hardly be any bigger. Presumably this is so because writing books about reading is as hard as reading.

Don't tell me that ordinary tesuji, tsumego, L+D or problem books would be about reading. They are not. They do not teach you how to read in general, but rather help you train reading if you already know how to read and can read well enough in principle for actually doing the reading needed for the given problems. Quite contrarily, such books can be misleading WRT "how to read" because usually they show so few variations that one cannot know from them whether one's reading was correct or just accidentally included one of the "correct solution" variations without really understanding why it is correct. The reader's major task for his problem book is to verify whether his reading was correct and complete, not whether his reading contained just the few diagram variants. The diagrams show only a small part of the necessary variants and often do not discuss most of the necessary decision-making.

When I was hard into learning and practice reading a few months ago I found a couple chess books that cover how to read and how to practice:

• Think like a grandmaster by Alexander Kotov (dry, but covers reading trees, what to read
• Pump up your rating by Axel Smith (very fun, covers reading/calculation, sorting, trees)

In go, the best I've seen is Davies.. Which is not to say much.

And agree with Robert: only reading gets you only "this far." Laying the stones on the board (even placing stones instead of reading variations), or memorising the problem and solving it completely (I'm doing this lately with the 5x5 Problems book before going to bed) get you out of a relative comfort zone, thus improving your abilities. In a similar vein, doing a lot of simple exercises is good, but it's also good to do some medium, some hard and probably some impossible as well as long as you give everything you can (for a certain amount of time.) Variety helps here. How much of which? That's the good question!

Ninja-edit: if I was to get back into chess, Silman's "The amateur mind" and Axel's book above would be my definite go-tos to read to improve my level of play (as well as a book on commented games by Paul Kérès I have)

_________________
Geek of all trades, master of none: the motto for my blog mostlymaths.net

I looked at a lot of chess books not too long ago for insights in this area, and there seemed to be lots of good ideas (though Kotov's work seems to be rubbished by today's writers, incidentally), but in the end I concluded that they were not specially relevant to go. I also think RJ's inventory approach also misses the special features of go.

In chess you have pieces with different moves, and you start every game from a set position which constrains the resulting shapes. For both reasons you have a texture that helps you feel your way through the analysis. In go, all the stones look alike and there is much freedom about how shapes develop. A prime task for the player, therefore, is to impose on this amorphous mass some sort of texture. Doing this by recognising oft repeating shapes might be called "chunking", a useful concept from western mind sciences. But it's actually more useful in texture-rich chess than go, because it doesn't solve the problem of how to recognise useful chunks in the first place.

Another major difference between go and chess is that there is typically more than one major fight. In chess, analysis can be usefully constrained because there is typically a killer move that signals overall success (checkmate, promotion of a pawn, win of a piece). In go, killer moves often have only local significance. You can let a group die in one area if you make a big territory in another area. In other words, the results of any analysis are just part of a subtle and difficult positional judgement. For this reason, compromises are often made and this has been a major characteristic of the Japanese style in particular. In the same way that you may cash in shares that have risen 15% even though they may rise much further later - but may also fall - Japanese pros will often accept the clear and known return now rather than speculate about later. Unsurprisingly, this concept of souba 相場 comes from the financial world.

This means that a Japanese pro will often not make the strongest looking move. In an interesting evolution in modern go, Korean and Chinese pros do try to make such moves. They have had much success, and have made the Japanese style look backward. But of course they have been aided in that by the much shortened time limits of international go where finding the rebuttal for power moves is much harder.

From the standpoint of understanding go, though, looking at the Japanese style is probably the best approach for amateurs. One major reason is that so much of our vocabulary and exposure to ideas have been based on Japanese books.

To get a clue about what is special about Japanese go we can start with Yoda Norimoto. In a book of over 200 pages of solid text devoted to how pros think, he makes the point that the perennial question from amateurs "How many moves can you see ahead?" is meaningless to a pro. They just don't think in the linear way of "if he goes there, I go here, and then if ..." He says that much more important than "thinking" 考える　or "reading" 読む　is "expanding images"　ｲﾒーｼﾞを広げる.

That may sound like recommending chunking (imaging) or giving advice that would apply to any national style of go, but it's really specifically Japanese because of the choice of word "reading", which is still part of the Yoda package. It would be easy to use "analysis" or "calculation" instead as a translation, and I and others have often done so on occasion. But there are good reasons to stick with "reading". It is the words that matter.

In chess it is the coordinates that matter. Players discussing a variation will say things like "bishop comes out to g5, knight to f6, pawn e4, Black castles, knight c3". In books this idea of coordinates is even more extreme: Bg5, Nf6, e4, Nc3. This does not work in go, but not just because we have a bigger board where our mental GPS system is always a bit fuzzy about the coordinates. A more important reason is that the shapes chess players talk about nearly always come up in one orientation - the coordinates are fixed. In go we see the same shapes come up in different orientations, including mirror images, and colours. GPS doesn't really work there. Even the iPad's ability to re-orient according to how you tilt the pad is quite as impressive as what the go brain does.

The way go players cope with describing variations is to use images (diagrams in this case) but - better and much more often - descriptive words. One of my own favourite anecdotes from a go commentary is (I forget the details but the gist is...) a group of pros discussing a variation and one says, "I thought about the hane" and the other murmur agreement, "Ah yes, interesting move". I look at the board, stupefied: "What hane? There isn't a hane available here or for miles around." Then it turns out they were referring to a hane deep in the line on, say, move 23. They had all followed the same mental route to that point and knew this was the first forking point.

But it would be wrong to assume that the result is a simple "Google directions" kind of route map. Buried in the words is also the intent of each move (often called the "meaning" of the move). Recognising that intent is actually a form of pruning the analysis tree. Knowing what you want or need to do means you can limit the tools in your toolbox to those that will do those jobs.

Because of the variations in English terminology used by various translators over the decades, often abetted by a fair degree of literalness, ignorance and stubbornness among the readership, too, it must be said, the meaning of a move as understood by a westerner is often not the same as the meaning understood by a Japanese person. On top of that, understanding among Japanese players can vary perceptibly according to grade.

A good example of that is provided by Ishii Kunio. Generally speaking there is a class of moves called "slide" 滑り in go, and their main characteristic is that they are a knight's move (large or small) from the third to the second line, or the second to the first. But the term hashiri ﾊｼﾘ is sometimes used instead, and sometimes just "knight's move" - why?

For a beginner, the slide he will see first or most often is probably the triangle move below.



In this case a Japanese beginner can easily visualise this as a slide, which will help him remember it. At this stage he is like the budding martial artist learning the form (the shapes) and has yet to get to grips with the applications (the intent).

But in the case of the following diagram, whilst a beginner would definitely use the term "slide" for the triangled move, and any other level of player could do so acceptably, a strong player would probably (says Ishii) call it hashiri.



The reason is that hashiru means "to run", and the idea behind the move - the intent, meaning or application - is more important than the shape or form. The idea behind "running" is to be in a hurry to get this move in before White plays A.

The distinction is not so much between weak and strong players, though. It is the "meaning of the move" that counts. Among both weak and strong players therefore, the triangled move below would probably not be called a "slide", because it is not the sliding that is important.



It would instead be referred to as "defending with a knight's move" ｹｲﾏに受ける or "sliding with a knight's move" ｹｲﾏにすべる, so as to indicate that the important shape to apply here (for defensive purposes) is the knight's move, as opposed to the push or the one-space jump.

In the beginning is the Word. From all this we can deduce that the advice often given in Japanese books, especially those on josekis, to grasp the "meaning" of the move has a sound basis, and is valid for western players so long as they take more trouble than usual to understand the meaning of a word (which tends to imply, I suggest, not blithely using the Japanese word in English).

Of course none of that obviates the need for hard work. To go back to Yoda, he likens studying go to "déjà vu". If you keep studying game records (in his case he says he went over the games of Honinbo Shusaku and Go Seigen over and over again") you will build up images of what you have seen repeatedly before, and this will give you clues as to which moves to try when looking ahead. Or to follow the great Yogi Berra: go is "like déjà vu all over again".

Is the book being referenced by Yoda Norimoto Vital Points and Skillful Finesse for Sabaki (http://www.britgo.org/node/3288) or Yoda Note (http://senseis.xmp.net/?YodaNote)?

What are what you call "the special features of go"? Only deja vu images?

More to the point, you cannot know well my approach to reading beyond the published basic principles because I have not published most of my thinking for reading (and need to research in reading). So it is too early to assess that I would be missing features:)

However, I guess that I emphasise strategic reading where you paint Japanese deja vu images. I'd rather agree to a made fresh positional judgement than to old games' patterns / game flow stories as part of the reading.

Whilst this is true for local shapes, centre fighting etc., whole-board fuseki does have an orientation for me: if my opponent plays the Chinese opening in any orientation other than down the right hand side of the board with the 4-4 at the top, I find it rather discombobulating .

Probably プロ棋士の思考術.

_________________
Dave Sigaty
"Short-lived are both the praiser and the praised, and rememberer and the remembered..."
- Marcus Aurelius; Meditations, VIII 21

I guess this is a bit OT for this thread but the quoted remark reminded me that in the Japanese pro fast play tournaments (TV tournaments) there is a lot more ransen (melee-type fighting) than in the slower big title games. The reason for this, I was told, is that in the slow games the players can read the fight through to a point where they can determine whether it is in their interest to engage in a complicated fight with more likelihood of both players making a mistake.

As for rebuttal of power moves, how often does it happen that in games involving these power move players, with short time limits, postmortem analysis shows that initiating these complicated fights was a mistake?

Not necessarily off-topic.

I wrote about reading quite some time ago now: http://gobase.org/studying/articles/matthews/dans/24/

That was for the MindZine, and our editor then was Jon Tisdall, author of Improve Your Chess Now, which went back to Kotov; maybe that was why JF was saying chess players are revisionist about Kotov.

Reading isn't really my thing. But there is an angle in that article which is not totally oblique to what you are saying. There are hints enough about what pros do find "unreadable", in game commentaries. Their reading is meant to be up to an order of magnitude faster than an amateur shodans. But not just magically more potent. Almost certainly much more accurate and systematic. For example, pros are supposed to be able to read ko fights through "to the end": not perhaps clear what that means, but streets ahead of the likes of us.

I'm pretty sure most amateurs would improve by getting into situations they can read out, avoiding those they can't by better self-knowledge. At the pro level, we have to assume that it depends on the pro. But the quality of "lucidity" is therefore more attractive to study (counter-intuitively) for amateurs who want to be better players. This perhaps doesn't work beyond ama 5 dan?

I have been working on my theory that "Sakata has a lot to answer for" over a number of years now. I'm probably about as likely to make it stick as to get to 5 dan. Very interesting that an ancient Go Review called his style "Apollonian".

O Meien has made the argument that there are two types of situations: finite and infinite. Finite situations are readable, and so you can determine how they will turn out with best play. Infinite ones are unreadable to their conclusion, at least to the players in the game, though you may make inferences about them to some degree e.g. white has more stones locally and so should be at an advantage.

Which way you aim to play depends at least to some degree on your standing in the game, and I'd argue that when players are behind and said to be looking for complications, it's really the infinite situation they are or should be looking for. He (O Meien) seemed to think that infinite situations were also more favourable to the poorer reader in the match, since both players were equally handicapped by the lack of clarity, rather than just the one who can't read as well. That's not to say, however, that one shouldn't work on one's reading.

Interesting stuff.
Some questions and comments here, things I have a problem with. Or something to think about...

• Cannot the "deja vu" principle be likened to "chunking" in chess? If not, why not? I don't quite see the difference. To me, in chess chunking is slightly better defined, like a certain "chunk" will come again and again. In Go some "chunks" are more amorphous - like "a wall" or something (although this might be more of a strateging issue, like a cotrol of a file in chess, for example.) But other "chunks" are also well defined, like joseki positions, or bent-four, stuff like that... like with chess "chunks" you learn these shapes, their properties, abilities, and potential - and this in turn lets you play better in positions involving these shapes. And as in chess, this allows you to read deeper. Replaying games of pros helps you enrich the database of such "chunks" you have in your head, as well as how to play with/around them - in both games. No?

• It is true that the orientation in chess is more invariable than in Go, but to me this only means that in Go there is slightly more complexity to Go "chunks" than to chess "chunks". Or maybe even not always that. Maybe when you talk about center position, this is the most pronounced - in Go a center formation projects or has weaknesses from potentially all 360 degrees around it. In chess, most likely, a center position only has to worry about what's up-front, and maybe a little what is to the sides... but even this could be debatable, depending on position. In Go, a joseki result is usually directed in 2 or 3 directions, while in chess an opening result is usually directed in 1 direction. So we have increase in possible permutation of shapes, but to me this is qualitative, not quantitative. It only means that I have to understand how a "chunk" behaves in 2 or 3 directions rather than in 1 or 2 directions, but understand it I still must, no way around that.

• Just like with a move which can be a slide or not depending on position or intent, in chess there are moves which can be, for example, either a fork or not, depending on position and intent. Or either attack or not... and so on. We can talk about a specific name for a move, but I find in chess there is also a lot to be said about a move having a meaning. Take a bishop moving to control a main diagonal. This is not just about Bg2 or whatever - you need to understand the meaning of this move before you make it or you will fail to use it properly and most likely lose the game. A move like a4 can be either to prevent/escape a mate, or to drive home a devastating attack, depending on context. And so on...

• Positional judgement and sacrifices in global sense - as you mention in Go you can lose a group in one part of the board while gaining large area in another, which can even out or even be to your advantage. In chess the same principle can apply, you can give up material on the queen side and so distract enemy pieces to drive an attack on king side. As a matter of fact, I often find such "trades" in chess more subtle and less obvious than in Go. And often harder to calculate. Be it as it may, you often give up one thing in one place to gain another somewhere else, or give up one kind of value to gain another kind (material for initiative in chess or points for influence in go, for example), in both games, even though Go has overall a larger playing field.

• Sudden death and "killer moves" - can be said to exist in both games, especially in pro play. Of course, Go does not have a "mate", but chess games seldom end im mate these dyas, especially at higher levels. I would say that in most final positions, a mate is not even a part of the calculation. In Go, like in chess, if you lose a bunch of "anything" without proper compensation, the game is practically over. Losing a knight in chess is no different than losing a group in Go if you get nothing in return. At top level, losing a tempo in chess or losing a sente in Go can be a game ender right there! In both games you resign, period. As a matter of fact, in Go this might be even more pronounced, because you often resign when you are behind and unable to catch up on clear and well-defined points which you can calculate - while in chess you often resign with material equality, and the reasons for resignation are qualitative, not quantitative, and sometimes much less well-defined and thus harder to calculate.

• And finally - it is always good to see how top pros regard the game and its aspects, in this case - reading. But isn't it true that us, mere mortals, still have to walk step-by-step? Reading move by move, he goes here, I go there, he goes here, I go there... and then do I like the result? It is good to know that pros might think differently, but how does this help me? I will never get to that level, nor would I want to, so will their method(s) really benefit me, or will I just chase my own tail trying to ape them mindlessly?

Having said the above, I find that from my experience and at my level, reading skills between chess and Go are in large part transferable.

To put it simplistically, I think that what we refer to as "reading" is basically composed of two parts:

• visualizing moves/sequences, the "he goes here, I go there" stuff - mostly concentration
• evaluating (final) positions, the "do I like the result?" stuff - mostly knowledge/experience

What's more, in majority of cases when we say "reading", we really mean only the first part, and this is what we train. Take for example tsumego... usually the end result (i.e. evaluation of position) is trivial - one side dies/lives. So when we say we can "practice reading by doing tsumego" we really mean we can practice our visualizing skills and our concentration. And even the pros advise to practice tsumego to improve reading, so I think I am on to something here.

I find this part to be transferable from chess, at least - practicing visualizing and concentration in chess also helps me visualize and concentrate in Go. As a matter of fact, I think that this process is actually easier in Go than in chess, since Go stones are mostly stationary and looking at the board helps visualizing much more than in chess, where the position is much more dynamic and it is not unlikely that by the end of the sequence you visualize all or most of the pieces will have changed positions. But then, reading in Go is usually deeper... but this is because it is easier, so you can go so deep.

Each time i take a break from Go to play chess for a while, and then come back to Go - I find this part of "reading" to be easier.

In general, almost anything that make you concentrate and keep this concentration up for prolonged period of time could be said to "train" this skill, and by extension - the reading skill in Go. One of my favorite examples is when I was preparing for a physics competition in my highschool. I studied for 14-16 hours a day for some two weeks, and afterwards, when I had time to play Go again, I noticed that I could concentrate much better and longer, and visualize much deeper than before - which increased my level immediately by a good chunk.

The other part of "reading" - the knowledge and experience, I am not sure you can really "train" in the classical sense of the word, it has to be learned and acquired over time. Maybe replaying pro games can help gain this "skill" without really feeling like "learning" - but its still learning, sort-of by osmosis. And I find this part slightly harder (but also more interesting and immediately rewarding) in Go than in chess.

This part is not really transferable, I think.
I include here also what John refers to as "chunking" in chess or "deja vu" in Go, as well as all the slight differences in the meaning of the moves and their japanese naming terminology and all that.

_________________
- Bantari
______________________________________________
WARNING: This post might contain Opinions!!

I have failed to notice a white kobayashi because of this.

While you both are far ahead of me (“Chinese”, “Kobayashi” ), I may note that many years ago (before the WWW ) I use to play Go regularly with a Real Life™ friend in Hamburg, Germany, and we often rotated the board every five or ten minutes by 90, 180 or 270 degrees … that was FUN! (We didn’t even know words like “Joseki” back then, and lacking better knowledge, we used to multiplicate “Komi” with the number of HC stones )

Cordially, Tom

_________________
“The only difference between me and a madman is that I’m not mad.” — Salvador Dali ★ Play a slooooow correspondence game with me on OGS?

Interesting viewpoint (although due my general ignorance, I needed to google Apollonian to understand it.)

I tend to rely heavily on reading. Sakata is one of my favourite professionals, which may not be unrelated! There is an anecdote in a book (possibly the Treasure Chest Enigma?) about a Sakata-Kitani game. There was a natural looking move, which Sakata didn't play. Kitani was the only pro to fully read out the result of the natural move, which was an approach move ko about 38 moves later. Kitani's reading assessed the number of available ko threats, and included the assessment that he stood better.

All the pros in the room were startled by the depth of this analysis.

There are two aspects of reading - depth and breadth. Kitani clearly excelled in deep analysis. To my amateur-ish view, Sakata excelled at breadth. In a number of games, he would play a move which looked incomprehnsible. After a few hours' analysis, I would conculde it contained a threat. After a few more, I would conclude that it also contains another threat and the opponent couldn't protect against both. I'll try to find an example of this later.

I posted a problem where breadth is far more important than depth:
http://www.lifein19x19.com/forum/viewtopic.php?f=15&t=7548