Quote: Well, I'd say keep reading until you think you can make a good judgement on if the resulting position is a success. If it's hard to judge, see if you can read more deeply to help.
Quote: I find it hard to believe that he counted 130 moves ahead and evaluated the score correctly even with 8 hours.
Generally speaking how far can pros read? Do you know some extreme cases?
We've had this discussion before, but I'll try to add something new. First, however, I think we need to recall the famous (and, I believe, unchallenged) finding of Adriaan de Groot for chess. He has been quoted countless times. Allow me here to select a quote from "Moonwalking with Einstein":
"What de Groot uncovered was [...] for the most part the chess experts didn't look more moves ahead [than top club players], at least not at first. They didn't even consider more possible moves. Rather, they behaved in a manner surprisingly similar to the [pattern-matching] chicken sexers: they tended to see the right moves, and they tended to see them right away. It was as if the chess experts weren't thinking so much as reacting. [... and] they described their thoughts in different language than the less experienced chess players."
Similar behaviour has been documented for go.
Now one important question for amateurs is: do I attempt to behave like the masters or do I accept my limitations and simply try to improve the efficiency of my amateur method?
I don't know the answer to that question, and I haven't seen a well argued case either way in either chess or go. But more often than not, writers appear to point their readers towards aping the masters, although in practice their specific advice seems to boil down to making the best of the bad job made by the amateur approach (add more proverbs, shake well, leave thoughts to simmer and hope your mental soufflé doesn't turn out too flat).
If the inference that it is desirable to ape the masters is indeed correct, and if the short-cut heuristic approach is indeed flawed, then one hypothesis is that we need to look at the second part of the summary of de Groot's work above: learn to talk the pro talk.
There are books and commentaries on chess that attempt to do this (and FWIW they seem to me to be on the right lines, but they point the amateur also towards the need for more work, which is perhaps not what a reader who has spent $20 on a book wants to hear). There are books in the Oriental languages that do the same for go. But what is lacking perhaps is material in English for go.
This lack may not be apparent when you consider the wealth of books and commentaries now available in English, but I believe this material is undermined in two serious ways which turn a potentially rich resource into a mound of nutty slack. First, there are still far too many people who insist on misusing or abusing Oriental words or concepts. It has taken a very long time, for example, to try to get westerners to think about thickness properly, and we are still only part of the way there (if you use 'influence' you are probably still part of the nutty slack school). Inventing new words in English is an extension of this problem, and takes us even further away from the words of the masters.
A more subtle problem is down to translators. Certain terms come up over and over again in, say, Japanese, and with repetition a term and its concept is inculcated in a Japanese reader. But different translators render the terms in a hundred different ways, and so the western reader not only gets no benefit of the repetition but often even gets no sense that the base word is either a technical term or an identifiable concept.
To illustrate in the context of the present thread, a pro does not read laboriously ahead down the branches of a tree, nor does he apply special pruning techniques. Instead he applies "chunking" (just like chess masters). Boiled down to its essence, 'suji' just means chunking. It is significant also that the Japanese word for looking ahead is 'reading': in the same way that we don't read words letter by letter, go words (suji) are not read move by move. In chess, too, weak players (and writers) tend to talk about 'calculation' (move by move) whereas top masters talk about 'seeing ahead' (in chunks). That is not to say that top chess and go players never calculate, but when they do it is a confirmatory phase to check something they have already found, and very, very often they are happy to trust their intuition and not calculate at all, instead devoting their time to evaluating the whole position.
What a go pro appears to be doing is not looking ahead, say, 21 moves. Rather he is looking ahead in 7 chunks (or sujis) averaging about 3 moves each. Like all humans he is constrained by the magic number 7 - the maximum number of items most of us can hold in short-term memory - but he can use his short-term memory better than amateurs because he can remember 7 big chunks instead of 7 individual moves, or mere chunklets.
Further (and this applies to the cases of people like Jowa apparently seeing over 100 moves ahead) he can offset one chunk against another (e.g. a hanetsugi for me here offsets a hanetsugi for him there; but pros often also use much bigger chunks) and these chunks are then eliminated entirely from the active processing, though of course their effects will accumulate in the far-off position being envisioned.
What happens when even the pro hits the limit of 7, even when he is chunking? At that point he has to make a judgement, and psychological factors may impinge as much as go moves. He may, for example, decide to play safe if he thinks he is ahead, he may follow his temperament (play riskily if he's a gambler), or he may try to play delaying moves in the hope of getting clarification later. In an extreme case he may try to maximise his chunks and go for a furikawari trade.
But does he always strive to reach his limit of 7 chunks? Apparently not, judging by the evidence of blitz games (and commentaries). Instead he will go only as far as necessary until a quiescent point is reached, at which point he will switch to a positional evaluation (there are techniques for doing this at very high speed, and counting is rarely needed - as Takemiya has said in a recent book, it is true go is a game of counting, but you only count at the end of the game, not during it).
What follows from this line of thinking is that it becomes important to know when quiescence has been reached. In chess it's rather easy - when there are no more captures to be made and nothing is left en prise. In go it's rather more subtle, but the quiescent points are nevertheless very commonly pointed out in commentaries. Perhaps the commonest word used is 'ichidanraku' (a pause). This is one major element of pro talk that I rarely hear amateurs use.
Exactly what constitutes a pause point I'll leave to the reader as an exercise, but apart from completion of captures or trades there are some interesting cases that don't seem to have an equivalent in chess, and the best example may be miai. When you reach a miai point you don't really need to analyse any further because you know you will always have a good continuation from there. It's almost as if the alpha-beta algorithm was discovered by go players long ago!
|
)its totaly wrong. 
