Life In 19x19

RobertJasiek wrote:Your suggestion to gather information at every single node is bad and very inefficient. It suffices to gather information at the terminal positions, branching positions and the start position!

Where do you know from that a node is NOT a branching position ?

Cassandra wrote:Where do you know from that a node is NOT a branching position ?

Such a (non-terminal) node has exactly one interesting next move and all the other moves are, without doubt, obvious failures or obviously inferior.

RobertJasiek wrote:It does not help us much that a) variations become the easier the more they approach terminal positions (because we still need to read on and compare with other variations) and b) an already completely read variations tree allows easy interpretation (because, except for the trivial problems, we can never construct such a complete tree.

I would like to assume that we start studying "reading" with simple / trivial problems. Therefore, I doubt your conclusion to a), just because we will stop "reading", when we have reached an already known position.

In his postscript to Igo Hatsuyôron, Inoue Dôsetsu Inseki declares that the number of move sequences in NOT infinite. So, “there will in the end be nothing you cannot see”. Therefore -- in combination with a) -- I also doubt your conclusion to b).


In principle, I would also like to assume that -- in your concept -- you largely underestimate the importance of "shape" (i.e. "position"). It was Inoue Dôsetsu Inseki again, who stated that the more you understand about "sequences", the more it will be absolute necessary to further study "shape".


And Dôsetsu finished his postscript by noting that if you were unable to find a solution, you simple have not grasped the right approach.

RobertJasiek wrote:

Cassandra wrote:Where do you know from that a node is NOT a branching position ?

Such a (non-terminal) node has exactly one interesting next move and all the other moves are, without doubt, obvious failures or obviously inferior.

You really want to claim that your theory is based on such unsteady terms like "interesting", and "obvious" ?

Quite apart from this, with our earlier steps on the path of "reading" (here: in the sense of "solving Tsume-Go") we will have accumulated some knowledge, which enables us to evaluate these unsteady terms. So, in the very beginning of our path, every node had been a decision point.


By the way: If something can be stated to be "without doubt", it will lose its property of being "obvious" only.

RobertJasiek wrote:

Kirby wrote:what Inseong refers to as "1-2-3 reading"

Is this a) reading the most interesing move first, then the 2nd, then the 3rd most interesting moves etc., (a good reading principle) b) reading move 3 before move 1 as a tesuji technique (not anything like a general reading theory but very specialised and often wrong) or c) something else?

Well, it works like I described, so you can interpret what you'd like from it. Maybe my description wasn't clear.

But my feeling is that this is the process:
1. Read the "normal" move first.
2. If the opponent has a good response, it's possible that it's a 'key move'.
3. Try playing the 'key move' first, and it might have better result.

Applied to life and death problems, the "normal" move is to reduce eyespace. The opponent may have a good defense, though. It's possible that their response to the first move should be played first.

Doesn't always work, but sometimes it helps in finding the answer.

Catalin Taranu's version:

Think about playing "3" first.

Might be that the "1"-"2"-exchange is not really necessary beforehand, or even hurts.

Learning by problems only should start with simple problems and proceed to more difficult problems. (My) general theory is applicable to every problem - simple or difficult, so you need not have doubts about (a) or reaching already known positions.

Finite or inifinite calculation does not depend on Dosetsus (in this respect limited) insight but on the rules. (Not that it matters in practice;) )

In order to get rid of your doubt about my conclusion (b), take a problem of intermediate difficulty, write down the complete tree, and after a hundred years you find that you do not have enough time for doing so despite finity;)

I do not underestimate shape but I also do not overestimate it:

1) It cannot be said often enough that VERY similarly looking shapes can have VERY different behaviours. The literature is full of examples.

2) The more I have studied problems carefully the more I realise that shape considerations tend to be unimportant. Of course, one must see all the shape tesujis; but do you also know where this ends? In advanced problems, every intersection provides a shape tesuji; there is no point in calling them tesujis - simply consider each interesting next move! There is one application of shapes in local problems that is really useful (AFA it can be applied): as a database of terminal local positions with known outcomes; if you know 100,000, you can end deeper reading a bit earlier on average than if you know only 10,000. Achieving this extra step of knowledge by heart is useful for strong dan players. For weaker players, basic reading theory does enough of the job if one's knowledge of terminal shapes includes some hundreds of nakades and a few standard shapes (such as arising in josekis or simple shapes occurring frequently in practice). If you want to see more in shapes and rely on Dosetsu, also show us his justification and explanation beyond what I write in this paragraph. Further shape study beyond that is not "absolutely necessary" but can also be achieved by slightly deeper reading.

To not find a solution can mean "not grasping the right approach" (Dosetsu), giving up prematurely while using a valid approach, or being unable to read broadly and deeply.

I do state that my theory is also based on the ambiguous terms "interesting" and "obvious", as I do state that my theory is also based on unambiguous components. While I condemn intuition as an excuse for not explaining things, the ambiguity I use in my theory has the flexibility to fall back to unambiguity whenever needed: "In case of doubt, consider EACH interesting next move." (And if "interesting" is ambiguous, consider, if necessary due to inapplicable simplifications, each local next move if the problem is local.) Reading by humans does not need the perfection of formal mathematical proofs or the algorithmic accuracy of expert system programs. One does not need to consider long lists of detailed principles (such as "Do not fill your own territory elsewhere on the board if the problem is local and a ko cannot occur.") but an "obvious failure" is good enough to exclude, e.g., such moves. Reading has lots of things that require meticulous thinking, so it is also important to save time whenever possible. For obvious failure moves, time can be saved easily. Of course, one must not repeat the frequent mistake of the past to pretend to oneself that a highly unusual move (such as an early 1st line move) would be an obvious failure just because of being highly unusual. One must not forget the other, aforementioned principle.

In fact, unpruned complexity of local problems is so extraordinarily great that ambiguous, fast pruning of obvious failures and the like is essential at every imagined position and the problem's current position. - If there is the slightest doubt, do not prune a move yet.

Yes, a go beginner is faced with the greater reading problem that each moment lets each legal move appear interesting. The beginner has a good sense of having doubts when skipping moves.

Kirby, that 1-2-3 tool is one of the few things professionals do like to teach verbally. Unfortunately, it is not so useful, except maybe like a proverb for a 5 kyu giving his thinking a bit greater flexibility to think in a way he has not considered before.

RobertJasiek wrote:To not find a solution can mean "not grasping the right approach" (Dosetsu), giving up prematurely while using a valid approach, or being unable to read broadly and deeply.

Sorry, but

"giving up prematurely while using a valid approach, or being unable to read broadly and deeply" are examples of not having the right approach.


Concerning your apparently deep aversion to "shapes" / "positions" (at least you seem to regard these as worth less than "sequences"):
Dôsetsu writes in his postscript to Igo Hatsuyôron that it will become extremely difficult to reach perfection without harmony between "shapes" / "positions", and "sequences". And I think that he is right with his statement. Relying too much on "sequences" means that one has not yet sufficient understood "shapes" / "positions".

RobertJasiek wrote:Of course, one must not repeat the frequent mistake of the past to pretend to oneself that a highly unusual move (such as an early 1st line move) would be an obvious failure just because of being highly unusual.

We can be very, very sure that Dôsetsu was very aware of your concern. Igo Hatsuyôron #120, Inoue Dôsetsu Inseki's life-time masterpiece, has several decisive elements of this type, and as such very well hidden before professional eyes.

RobertJasiek wrote:
Kirby, that 1-2-3 tool is one of the few things professionals do like to teach verbally. Unfortunately, it is not so useful, except maybe like a proverb for a 5 kyu giving his thinking a bit greater flexibility to think in a way he has not considered before.

I am not 5k, but I still find it useful. I agree that it doesn't always work, but it gives me a map in solving problems. When faced with a problem for which I do not know where to start, it gives a systematic ordering that I can use for searching.

I find this useful, because otherwise, I would have no systematic way of searching for the next move (my choice would be arbitrary).

It doesn't always work, but I do think that it has improved my KGS 1k/1d mind's flexibility.

Catalin said (with regard to his version, which has a similar effect) that this kind of "tool" may serve you well to make you thinking about (at least) an alternate move.

RobertJasiek wrote: 1) It cannot be said often enough that VERY similarly looking shapes can have VERY different behaviours. The literature is full of examples.

Where is the problem with it ?

Click Here To Show Diagram Code

[go]$$ Igo Hatsuyôron #43
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . O . . X . |
$$ | . . . . . . . . . . . . . X O . O . . |
$$ | . . . . . . . X . X . X . . X O O O . |
$$ | . . . , . . . . . X . . X . X , . X . |
$$ | . . . . . . . . . . . . . . . . X . . |
$$ | . . . . . . . . . . . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |[/go]

Cheng Xiaoliu mentioned that White could not be killed unconditionally any longer, if any of the Black stones was absent.

So, it's a matter of shape, isn't it ?

Cassandra wrote:Concerning your apparently deep aversion to "shapes" / "positions" (at least you seem to regard these as worth less than "sequences"):

My aversion is about the myth that shape knowledge (other than that of terminal positions) could significantly increase reading skill by something like providing "missing links" between different phases of reading into a problem or tesuji shapes guiding towards faster reading. I am beyond that illusion. John's advertisement for viewing shapes as something dynamic comes closer to reality because shapes are constantly changing during the sequences. But then we speak of techniques rather than only shapes.

Dôsetsu writes in his postscript to Igo Hatsuyôron that it will become extremely difficult to reach perfection without harmony between "shapes" / "positions", and "sequences".

Uh, is he saying anything else than that sequences consist of alternating moves and positions? I.e., both are equally important so to say.

Relying too much on "sequences" means that one has not yet sufficient understood "shapes" / "positions".

Sequences consist of moves and positions, even if one writes them down only as moves. Each move leaves and reaches a position. So what meaningful are you saying?

Igo Hatsuyôron #120

Not only #120 :)

Where is the problem with it ?

The problem with relying on shapes although similar ones can have very different behaviours is that one must not draw (naive) conclusions from one known shape to a very similar unknown shape, although it may be very tempting to do so.

Kirby wrote:otherwise, I would have no systematic way of searching for the next move

In such a case, if necessary, consider each interesting move. Preferably, in an order of perceived decreasing likelihood of success, else random.

Of course, AFTER having discarded the obvious failures and obviously inferior moves.

RobertJasiek wrote: In such a case, if necessary, consider each interesting move. Preferably, in an order of perceived decreasing likelihood of success, else random.

Of course, AFTER having discarded the obvious failures and obviously inferior moves.

Sure. But the question is, which moves have good likelihood of success? As far as I'm concerned, selecting the opponent's refutation as is done in 1-2-3 reading has as much likelihood of success as any alternative method that I know.

I mean, the method works on several problems, even if it doesn't work on all problems.

Perceived likelihood of success can be guessed from the moves' functions or initial purposes. E.g., moves with some functions are more likely successful than such without any apparent function. A single threat is less promising than a) a direct move or b) a multiple threat. Etc.

Kirby wrote:Sure. But the question is, which moves have good likelihood of success? As far as I'm concerned, selecting the opponent's refutation as is done in 1-2-3 reading has as much likelihood of success as any alternative method that I know.

I mean, the method works on several problems, even if it doesn't work on all problems.

I think, this is similar to answer the question where your opponent would play if it was their turn. And it will be even better than choosing candidate moves by chance.

A simple example is capturing a single stone for an eye (your opponent's turn) vs. stretching your one-stone-group into a two-stone-group (your turn; probably causing some kind of shortage of liberties).

This is just a strong hint, where a vital point of the shape / the problem might be, nothing more. But we can assume that your opponent will choose their moves on purpose, so it is worth considering to try destroying their initial idea (as a matter of course this is true for your opponents move as turn #2). However, there is no guarantee that this kind of hint will always lead to the "real" vital spot of the problem. Just because there are problems of a more complex kind.