Quotation reference:
viewtopic.php?p=149062#p149062



There are players (example: I) who learn more than 10,000 times faster from the combination of go theory, playing games etc. than from playing games only. This presumes that go theory is available for such players' needs. For dan players, explicit go theory is still scarce.

We have also been told that other players improve from playing games only and that they would not need any go theory. This, of course, is utter nonsense, because all such go players understand a lot of go theory, when they speak about games or participate in discussions. It is easier to buy that such players do not learn or do not prefer to learn go theory from explicit source and with much conscious effort. This is what they should say: they prefer to learn from implicit sources and subconsciously. IOW, it is not go theory per se that would be useless for them, but it is explicit representation of go theory that they dislike.

Go theory books can be a great shortcut for those wishing and being able to learn from them.

Go theory shortcuts are not magic, but should be available in terms of reasoning and explanations. Unfortunately, not every go theory is presented in that manner. This, however, does not make it magic, but this creates a duty of finding and describing reasoning and explanations.

If one wants to become strong only by examples, one needs hundreds of thousands of them: not only games and not only stones, but moves, positional contexts for moves, sequences and trees of sequences. Suppose you need to study 10,000+ games etc. It means to study hundreds of thousands of moves and their contexts. Even if done subconsciously, it won't be less.

Magic of go theory? Many players, especially kyu players, greatly underestimate the amount of necessary go theory for becoming a strong player. (For this assuming one would be learning from explicit go theory.) It is in the order of thousands (or maybe even a few dozens of thousands) of principles or other general knowledge pieces. Most of them are specialised, but nevertheless necessary. The number is one or two orders of magnitude below the number of examples needed to learn only by examples. However, the number is not small. There are already dozens (or a few hundred?) topics of go theory, before one has even started to study actual theory for a topic. There are at least hundreds of principles etc. per major topic.

There is no magic also because the number is not small. There is also no evidence for go theory being useless just because somebody gives up after the first few dozen principles etc. When somebody gives up so early, then of course he will not reach a state of appreciating the power of go theory. Somebody reading "Avoid premature endgame." but not applying this piece of go theory regularly in his games will of course stay weak and so fail to appreciate even the basic theory, and much more the more advanced theory.

Making money of weak players? Books teaching only by examples make a lot of money: they teach too little go theory per book, so that the weak players need many such books. Go theory books save a lot of money, because fewer books are needed to reach the same understanding. Relatively fewer does not mean "only a few" though, although a few players manage to reach dan level with only a few books. And the major reason for the few needing only a few is that they get a lot of also necessary other go theory (such as strategic planning, which is still hopelessly underrepresented in go books) from other sources. For the knowledge of an amateur high dan, however, quite a few books are needed (but mostly not available yet), if go theory shall be learnt mainly from books. (Don't forget: alternatively one needs hundreds of thousands of examples.)

It is not go theory books per se that make a lot (BTW, "a lot" is actually "very little") of money, but there are very different go theory books. Part of them teach dozens of times as much contents as others. Even without extrema, a factor three between books teaching much and others teaching little go theory can be assessed easily. Why would some readers complain about having to buy too many go theory books? Because they prefer to read those with little go theory contents, so they need a few times as many such books as they would need if they read carefully only the books with much and dense contents. Nowadays, at least for a few topics and at kyu level, readers have the choice between more books with less contents per book and fewer books with more contents per book.

One does not become stronger from go theory books just "somehow". One becomes stronger by understanding the contents and applying it! What are all those readers doing? Where are the countless discussions on topics of go theory books not understood well enough to apply them in their games? A missing energy to learn from books is hardly ever the books' fault. Eager players learn even from the dullest dictionaries. Of course, there are didactically worse versus better books, but without much time spent to play more games for applying read theory, there is little point in blaming the books instead of oneself as a player.

If you depreciate go theory as gimmick, you are not trying hard enough to understand and apply the theory, well, of course unless it is a bad example of faulty theory. But how would you ever find out if you simply blame pros and authors, instead of discussing the theory itself? You do not learn go theory from meta-discussion, but from discussion about and reflection of the theory itself.

For once, I don't disagree with much of what you say except for here. However, I find most kyu players greatly overestimate the amount of go theory necessary. Too many people (me included) focus on go theory books and not enough on improving reading. I find lots of kyu players who read a lot of the fundamentals books but can't read. They don't realize that why they are losing is not spending the time more efficiently on problems.

Of course - it all hinges on what we define as 'go theory' - and on the question if we all enter this discussion with the same definition in mind. So I think this what should start all this - a solid working definition of 'go theory', just to avoid misunderstandings.

Having said the above, I think you make an excellent point here. Go players, as they get stronger, understand more and more of what I think of as 'theory'. The stronger the player, the more of that he/she seems to know. The source is not only playing, but also discussions with other players, game analysis, and books - many of which present some aspects of said 'theory' in one form or another, often without intending it - for the user to assimilate.

So personally, I would find it much more magical if a player got strong *without* any input of go theory, only on examples, problems, and games.

I think what RJ means here is books which are specifically written to present and explain 'go theory', rather than presenting it piece-by-piece and often between-the-lines like many other books do. In this case - it is a shortcut indeed, but a legitimate one. Who wants to read 10 books instead of 1 for the same effect? Such 'shortcuts' should be supported, I think.

Still... to repeat myself - it all hinges of what we define as 'go theory.'

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

I think it's possible that a series of examples and problems could be understood to introduce theory as a second order function, much in the way that quite a few beginners don't get explicitly told about 2 eyes, or seki, but discover it themselves.

Defining what we understand as the phrase "go theory" is interesting, but, Bantari, please forgive me that at the moment I lack time for that. Surely we will discuss this later:)

Possible? Yes
Desirable? Not sure, possibly, it depends
Efficient? Not sure, possibly, maybe not so much - it sort-of connects to EdLee's and mine discussion about teaching approach.

I think you might be right that it depends on the level the student and what exact aspect of 'go theory' we are talking about.

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

There is a lot of truth to this... This is why I like the level up books. Its a great mixture of very basic theory and a very systematic approach to improving reading, beyond life and death problems. The simplest and most obvious example of this is the "can you cut" problem variety. Doing these sorts of the problems over and over with slightly increasing difficulty has done wonders for my ability not to lose as much as I do and all it is really doing is making sure one reads the moves for both players out properly.

There are several serious flaws in RJ's reasoning, and some tendentious wording, but just to pick up on a couple of things apart from oren's good point about over-reliance on text books as a way of putting off the work needed.

1. Top experts in a profession such as go do not seek to "learn" or "understand" go theory, but rather to acquire it. As in various other fields, the experts are then able to exercise superior skill without explaining how they do it. Furthermore, they apparently soon learn not to even attempt to explain it because that appears to interfere with the automaticity of their responses. We've all experiences this - you type away unthinkingly at high speed, but as soon as you stop to think where the gefurtel key is or how you spell grok, you become all fingers and thumbs.

2. No-one, as far as I know, has yet demonstrated extreme skill in go, or anything close to it, by having learned theory from books. Until someone does, it is fatuous to argue that learning theory is a superior method to doing grunt work if you want to become a real expert.

3. There is an underlying assumption by RJ that everyone studying go wants to become superstrong. This is nonsense. Some (most adults?) simply want to learn to appreciate the game better as fans, and for that they need ways of verbalising theory not to get stronger but to talk to or understand other fans.



4. This is not go theory, nor is it accurate. Even within its own little scope the advice should be "avoid premature boundary plays" because many of the plays referenced occur in the middle game, but the main point is that you have to avoid premature anything - premature forcing plays, premature side extensions, premature reductions, premature probes, etc etc. - because that's what "premature" tells us. This is just plain common sense. You might just as well pretend "avoid bad plays" is go theory. The fact that some people are too dozy, lazy or stupid to apply common sense does not mean that they need a dose of go "theory". They probably just need a dose of caffeine or a kick up the backside.

Hmm... a few comments:

To 1,2,3, and in general :

I did not see Robert's post as addressing the 'top experts' or the 'super strong' - just the opposite, he mentions kyu players quite a few times, and so I understood that this is the demographics that his comments consider. And the books which are the subject here are also most likely addressed at kyu players.

I fully agree with your comparison about typing, it is very good point. However - it is my contention that, as with typing, we first have to learn the skill consciously before it can become 'second nature'. Judging by myself, I do play many moves subconsciously, but this does not mean I don't know the underlying theory or reasons - its just that I don't need to bring them up to a forefront of my mind when making the decision. But I still need to know them. I am not sure if the same applies to the 'super strong', possibly now, but since I will most likely never attain that level, I have to do what I can to do what I do. Same for the rest of us.

The 'super strong' and the 'top experts' usually have at least one huge advantage over the players the books in question are targeting - they can actually afford to devote a big chunk of their life to Go and take their sweet time 'inhaling' the necessary theory while doing problems or looking at examples, or whatnot. They also have another huge advantage - constant/frequent access to other top players to discuss ideas with and top-level competition to test idea against.

And so for them, as the end result - I agree with you that what they do in the long run is a much more efficient method of becoming stronger (this is probably where my disagreement with RJ lies.) But most of us do not have this luxury, and learning stuff more explicitly can be a very pleasant alternative.

To 3:

As far as I know, nobody demonstrated any great skill with absolutely no knowledge of 'go theory', so I would assume it *is* important and learned - either explicitly or implicitly, and the knowledge is internalized on either conscious or subconscious level. Most likely - it is a combination of both.

This all reminds me of the 'learning' thread from a few days back...

To 4:

This is why I said misunderstanding can be avoided when we define exactly what we mean by 'go theory'. I have a feeling we are not there yet.

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

This reminds me of what Maeda Ryo was saying at US Go Congress. Children who want to be strong are not given access to joseki or other books, so that they are forced to solve problems on their own while their brain is developing. For adults, he says that's not quite as reasonable. We need shortcuts, and that's fine.

It's similar with language, start talking a four year old through the grammar of a language will get you nowhere, exposing them to the language constantly and repeating correct sentences to them will do the trick though. Trying that with adults or older children will only get you mixed results, many people will get nowhere without word lists, grammar manuals and books on idiom.

You don't even have to correct them. As a kid I said, "instreting", something that my parent thought was cute. They also liked, "clapsed".

(The latter for "collapsed", not "clasped". )

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

Visualize whirled peas.

Everything with love. Stay safe.

Easy enough. A theory is a collection of statements of facts along with concepts that permit concise statements of those facts. A theory also allows statements that have exceptions.

Thus, an empty triangle -- pardon me for not saying what that is in less concise language -- is bad shape. Except when it isn't, of course.

----

One thing about go theory is that much of it is vague. That means that it takes judgement and experience to understand many concepts. Sakata noted that in postmortems with other pros they sometimes would disagree about whether a move was kikashi or not. Sometimes it seems like advanced go concepts are like swing in music. What is swing? If you gotta ask, you'll never know.

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

Visualize whirled peas.

Everything with love. Stay safe.

Good definitions, I can certainly live with that.
Is that the same 'go theory' the OP was talking about? I dunno...

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

In professions such as go? This is important here: go is activity, where, in principle, reasoning and explanation are available to those wishing to think in those terms. This is so, because go is a complete information game (in its definition, not necessarily in its playing). "Top experts" in go are a) top players, b) top researchers or c) top computer programs. For (c), explanations are dry in our eyes, but they are available (at least to the interested programmer). For (b), explanations are essentially mandatory. For (a), all top players use explanations by reading and endgame calculation and decision making due to these aspects. For other aspects of go theory, quite a few appear to rely on a mixture of explicit go theory and subconscious thinking, which they cannot explain (well enough) to themselves and others. This does not prevent part of the top players to write books that also contain explanations for aspects of go theory.



There is no such automaticity, or they would not consume any noteworthy thinking time.



1) Are you sure that none of the top players has read at least some theory of books, such as tesuji types, life and death techniques implied by LD problem books etc.? ;)

2) If you said "learnt most of his go theory knowledge from books", then, of course, you would be right, because by far the most go theory needed for amateur and pro dan players is not available in books yet.



1) For kyu players, your argument does not hold, because go theory books for them are available, and quite a few players have become significantly stronger also by reading, learning from and applying them.

2) As I have said many times, go theory books are not the opposite of other means for becoming stronger, but playing, self-review, others' reviewing, learning reading + problem solving, studying games and taking lessons are / can be also important for improving.



Not exactly. My underlying assumption rather is that many want to become an amateur dan, possibly amateur high dan, possibly stronger, but anyway want to understand what dans do even of they do not reach dan playing level themselves.

Those aiming at 5 kyu maximum need only relatively little go theory. Everybody else, IMO, can as well learn much of an amateur high dan's go theory quickly, instead of delaying most of such go theory. I think so, because most go theory for amateur high dan is as easy or difficult as go theory for SDKs. The real difference is a) greater completeness of an amateur high dan's knowledge and b) his knowledge of a lot of more details of go theory. A kyu player does not need all the details yet, but can decide when to learn how many of them. However, apart from sheer details, a kyu player should rather learn an amateur high dan's kind of knowledge than partly bad knowledge the kyu player needs to unlearn later. This is so, because an amateur high dan's knowledge and a kyu player's knowledge both consist of simple and complicated knowledge pieces. The amateur high dan's simple knowledge pieces can simplify some of the kyu's unnecessarily complicated knowledge pieces, and the kyu should better abandon those of his simple knowledge pieces that are wrong because of being too simplifying. Instead of sticking to bad, simple knowledge, it is better for the kyu to use already the good, simple knowledge of an amateur high dan. Only WRT possibly too detailed knowledge, the kyu should decide how much of it he already can and wants to learn versus postpone for later.



The same theory can be used by them, except that they would want to skip more of the detailed knowledge.



Whether it is "go theory" surely depends on how one defines what is "go theory":) It is not scientific go theory, but it is applicable-principles-based go theory.

Of course, this particular principle is not accurate. Would you use this as an excuse for a DDK not to become stronger up to a few ranks, because you would advise him not to apply anything that is only 99+% correct in practice? It is incorrect only when one misjudges greatly whether a move is only-endgame and whether that is premature, i.e., the game has not reached the endgame stage yet. (Note that a middle game move's secondary meaning of also being good endgame is not prohibited by the principle.)



"Avoid premature boundary plays" has a meaning different from "Avoid premature endgame", because the former principle also advises to avoid aji keshi in boundary play that are not pure endgames. The latter principle has the advantage that it can be applied also if one does not want to complicate matters by aji keshi considerations. It is worth discussing whether the former principle is a good generalisation of the latter principle. IMO, this is unclear so far, simply because it has not been studied enough.

"Avoid premature endgame" does not have an only "little scope". The principle is applicable throughout the game and relevant outside the endgame phase(s). Application versus non-application of the principle improves the score by dozens of points. I call such a "great scope".



It is also worth discussing whether a yet broader generalisation can be your suggested "Avoid premature anything". OTOH, a principle with such a broad generalisation needs to be broken down again into its parts. Otherwise we end up with principles as powerful as "Win or else tie the game".



If it were, I would not have needed many years to acquire this particular piece of go theory, and countless of kyu players would not be violating "Avoid premature endgame" too frequently.

It is common sense only when one has already absorbed and accepted the principle. Before, it is without sense, as can be seen from the many violations of the principle out of unintentional ignorance of its existence and contents.



There is nothing to pretend about. It is go theory, and it is good go theory. (Of course, it is insufficient go theory, because "bad" demands explanation in detail.)



All players start with the absence of "Avoid premature endgame" in their games, when they are still weak DDKs. All, because every weak DDK makes such mistakes. This shows that the principle is NOT common sense, but its contents must be learnt, before it can be applied regularly.



Wrong, see above.



Wrong, because common sense does not tell weak DDKs to apply the contents of the principle, regardless of how much caffeine of kicks they get.



In conclusion, where is there any flaw in my reasoning? All you have identified is a possible (worth discussing) embedding of "Avoid premature endgame" in a broader principle, i.e., a possible advance of go theory, which KEEPS the principle as a special case of another, broader principle.

It comes close to kinds of go theory I prefer, but there are also weaker forms of go theory.

I don't mean correct them, I mean just speak correctly around them. Correcting them is usually an utter waste of time.

I say "I love you" to my daughter (3), she says "I like you too Dada" to me, in Irish romantic love, love of your country etc is grá, but the construction to say you love a sibling, child, friend or whatever is the same as "I really like/love ice cream," "Is brea liom tu, a Dhada" or whatever, whereas a lesser form of "I like ice cream" would be "Is maith liom uachtar reoite." One just figures out by context what a person means. She gets mixed up with the phrases since she's 3 and she doesn't really bother with merely liking things she either really likes it or doesn't have an opinion. Correcting her is pointless, responding to her correctly "I love you too Siún" does eventually do the trick though.

I have some trouble with this (I'm not saying you're wrong, but this may be why people misinterpret you in this respect). When someone says theory to me I really don't think of applicable principles. Applicable principles are a how-to/pedagogic thing, quite often distilled from either experience or theory, but not in themselves theory since they instruct rather than explain or prove. Theory, for me, would be proving why in most or all cases premature boundary plays are sub-optimal play, better would be to prove the majority of cases and then the classes of possible exceptions to the maxim. Advising someone not to play premature boundary plays is advice rather than theory. Neither is more or less valuable than the other, merely different things.

For a mathematician, theory is what is or can be defined or proven formally. For a physicist, theory is what does model or is expected to model reality, as it is or shall be observed. For an inventor of go principles (etc.), theory is those principles modelling reality in a sufficient (e.g., specified as "at least 95%") fraction of cases, as they are or shall be observed. For an informally talking or teaching player (when he does not rely on aforementioned principles), theory is any piece of knowledge believed to be relevant, but often without evidence or specified threshold of applicability, except that specialised evidence (such as the reading of related variations) for particular examples might be available.



Probably, this is so for many traditional principles / proverbs. It is different for most of the principles written down by me, because I rely them on careful study (such as empirical study), implications / refinements from earlier principles or in some cases mathematical methods. IMO, it is also different for a few other authors or researchers, who do not just copy tradition, but do their own careful studies or least careful checking of empirical evidence collected in their own memories.

I have a naive definition of what a go theory book is. If more space is spent on words than diagrams, it is theory.

In other words, if I have to read a lot of text to get anything out of a book---theory.