My experience bears out what Bill says, but I'll add a few words from a different standpoint. Often, the more sidelights, the better we can see.
I think there are two aspects worth highlighting most. They are related.
As to the first, I have an English friend who is a taiji expert. He studied in Beijing Sports University and has won gold medals in international competitions in Asia. He is also a high-level classical guitarist. He teaches both disciplines for a living. They are quite different, mentally and physically, yet he has achieved an extremely high level in both. His own explanation can be summed up in an observation he often makes: people often say (to themselves or others) that you should practise, practise something until you get it right. That is so wrong, he says. What you must do is practise until you never get it wrong.
This difference in attitude is subtle in words but huge in terms of commitment and time. However, there is a trap for the unwary. If you are a beginner at something, by definition almost you are likely to be doing it wrong. If you keep practising, practising, you run the risk that you are reinforcing what you are doing wrong. That's one reason why a teacher is always recommended, but in practical terms most of us have to do without. This is where the second aspect comes in. But before going on to that, let me mention a go example of the "never get it wrong" attitude.
I don't recall Go Seigen making a misread - his few mistakes were usually strategic. And this shows in the significant number of very hard classical problems where for years 9-dan pros of the "practise till I get it right" variety have published books of tsumego problems with the wrong answer, only for "never get it wrong" Go to come along and expose their errors.
My second aspect is visualisation. My taiji/guitar friend does not use this term, though does follow the principles. So as to retain that term let me switch to Walter Gieseking. He was a famed piano virtuoso, and part of his fame rested on the fact that he could learn a long and complex classical piece on a plane trip and then play it from memory, faultlessly. He co-wrote a famous book called "Piano Technique." He uses the term "visualisation" a lot. To me, his usage doesn't seem to have anything to do with 3D images or rotations in space. I would say it is more or less "pattern recognition," a term that was perhaps not current when he wrote, but with the proviso that it includes dynamic as well as static patterns - sequences as well as shapes.
For example, one thing he says is: "The first thing to be done is to visualise the notes."
He takes us through several examples of how to do this. They are frightening at first. The page looks black, with forests of demi-semi-quavers. But (with no reference to fingering or other elements of physical technique) he analyses each piece, showing the hidden patterns. He is not writing for beginners, so the analysis is high level, but to satisfy curiosity here is one example:
Quote:
The accompaniment in this case is a broken C major triad in the left hand. In the first measure a quarter note C is followed by a quarter rest; and in the second, third and fourth measures the quarter notes on E, G and C are followed by corresponding quarter rests. Measures five to eight are the same for the left hand as the first four measures. With the help of visualisation, these eight measures can therefore be played easily, without music; that is, after careful reading without notes.
By analysing this paragraph in a way similar to his own analysis of the piece of music, we too can garner useful information. For example, his reference to E, G and C as a group is not explained but is something that all beginners would learn, and so he is building more complex patterns on a base of simpler patterns. Also, although again I imagine the term "chunk" is not one he would have been familiar with, we can see that he chunking information into groups such as EGC, so that instead of memorising/visualising many discrete items we only have to memorise/visualise a few chunks.
It should be noted that Gieseking demands intense work on this analysis:
Quote:
Having thus carefully visualised several measures, we practise and play them. There are problems to be solved in every measure, which, if they are to be satisfactorily mastered, require careful and concentrated analysing and practice. It is therefore advisable to tackle one measure at a time, and to continue practising it until all the difficulties have been overcome.
in other words, never get it wrong.
Quote:
By this method of visualisation, this careful thinking through of the piece of music in question, the pupil will be capable of writing down the whole exercise from memory. After intense concentration, most of my pupils have been, to their great astonishment, able in a few minutes of time to play the entire exercise from memory. Visualised reading at the same time affords the pupil the best insight into the form of the composition under study. ... This at first somewhat mechanical process will quickly enable him to grasp the import of a composition.
Again to try to render this in nowadays more familiar terms, Gieseking is recommending us to use our conscious brain (analysis) so that our unconscious brain (practice) can never get it wrong.
Applying this to go, I can mention a couple of examples. One is the Meijin Inseki's "Today We Have a Splendid Feast" in which he picks out a specific tsumego technique (analysis) and then gives a large number of problems focusing repeatedly on that one technique (practice).
Another example is the shape pattern recognition implied in the proverb "the L shape is dead." You still have to know specifically how to kill it, of course (the sequence), but this chunking has already done most of the work for you, and with "intense concentration" you can (and should) even chunk the killing sequences. But even this is not mastery of the L shape. You need to know what happens if there are extra legs, hanes, liberties, of course, but did you know that if you are a Go Seigen kind of never-get-it-wrong pro you will also know how many liberties the L shape has when it comes to killing it? There is always something more to analyse! That might be an esoteric example, but actually the principle applies to almost every problem done by almost every amateur - we find
a solution and then move on to the next problem, instead of finding
all solutions to the first one.
From the above, what I believe follows for the go player is:
1. Those who recommend solving a problem by just sitting there calculating until the solution pops out are misguided (unless they are masochists, of course). They risk reinforcing bad habits and are probably not using chunking enough. It may be impressive to see an amateur struggle and struggle over a problem and eventually come up with a (possibly tentative and probably partial) solution. But I promise you it's a LOT more impressive when you see a pro, as I have done with Kato Masao in a press room, walk past and casually pick out the right answer in front of the strong amateur in a split second. And the explanation is not having seen the problem before. I can remember being overawed by Mrs Sugiuchi who was analysing an amateur game that obviously she had never seen before or had had time to study. She mentioned that the reason for some move being good or bad was the ladder. What ladder? I asked, seeing a totally ladderless board. She then looked at me with the same amazement that I was looking at her - for quite different reasons, of course - and then rattled off a 30-40 move sequence that created a ladder in another part of the board. I've seen that too often to question it now.
2. Those who recommend peeking at the solution of a problem either at once or, more usually, after some effort to solve it ourselves, are on the right lines. But they are usually guilty of poor pedagogic technique because they never explain how to analyse the solution, or to stress that "intense concentration" to do this is still required - just a different sort from pure calculation. In particular, they are usually unlike the Meijin Inseki and never give a swathe of similar problems to reinforce the techniques.
One reason, though, that the advice of the latter group seems basically sound is that seeing the solution is a form of visualisation. You are seeing the goal that you have to work towards. That seems to stress efficiency, and humans need efficiency (in the form of heuristics, admittedly) to compensate for lack of brute calculating power. This also accords with a tip I got from someone about how to memorise a piece of music. It's obviously not Gieseking's method and I can't say it worked for me, but the tip was to start with the final passage of the music (so that you know it's supposed to resolve, and thus how it may evolve).
Another underused tip is to wipe off the stones after looking at a problem (which of course implies using a board not just a book - just as we use a piano and not just sheet music) and then, as Gieseking implies, put down the entire position from memory. We probably do this by laying out certain obvious shapes first, so we are in a way analysing the structure of the position (in the "EGC is a chord group" kind of way, but we can build on that later).
We can be as eclectic as we like about tips. As Bill say, everybody's brain is different. But I'm sure there are enough shared elements to make it worth at least heeding the example of pros like Kato and Sugiuchi and the advice of masters like Gieseking and the Meijin Inseki.