Life In 19x19

Ever since I learned to play go I've always had at least one strategic concept I knew I didn't even apply; usually, more than one.

Is there a point where you already know everything about go and just have to learn to do the same but better? Or are there subtler and subtler concepts to learn and apply, all the way to the top.

All the "concepts" are just heuristics that abstract something that may or may not be meaningful about any particular game. I guarantee that no matter how many times you decide you have "learned it all", the very next day some obnoxious SOB that hasn't studied anything but the bottom of a beer glass in years is going to clean your clock! There is no end to the complexity of Go; there is only overconfidence.

Alguien wrote:Ever since I learned to play go I've always had at least one strategic concept I knew I didn't even apply; usually, more than one.

Is there a point where you already know everything about go and just have to learn to do the same but better? Or are there subtler and subtler concepts to learn and apply, all the way to the top.

Until go is solved ( ) there will always be new strategic concepts to learn.

ez4u wrote:All the "concepts" are just heuristics

No, they are better than that! They come with principles and methods for advice of decision making.

Bill Spight wrote:Until go is solved ( :mrgreen: ) there will always be new strategic concepts to learn.

Bill, how long do we still need to solve it, eh?:)

Practically speaking, a few weeks ago, I have rediscovered a very important strategic concept I have never seen mentioned in literature, never heard of and completely overlooked in my games, but I am sure that must exist verbally among professionals and inseis. Don't worry, it is such a nice concept that I will write at least one book about it (for that reason, I do not reveal it now). The type of thing you'd wish would have reached the Western knowledge pool decades earlier.

RobertJasiek wrote:

ez4u wrote:All the "concepts" are just heuristics

No, they are better than that! They come with principles and methods for advice of decision making.

Well yes, that is what heuristics are, right?

You can never run out of strategic concepts, because you can always come up with more and more complicated, less and less applicable heuristics, and you can always give names to more and more tenuous abstractions that arise on the board.

But you definitely run out of useful concepts rather quickly.

Bill Spight wrote:Until go is solved ( ) there will always be new strategic concepts to learn.

Even solving games doesn't automatically give us a human-friendly strategy. Checkers has been solved; does that mean a human can go look up the relevant proof and quickly reach checkers 9d ? What about an easier game like connect 4?

palapiku wrote:more and more complicated, less and less applicable heuristics,

This sounds so negative, but actually current strategic concepts are mostly basic or close to basic. (Connection is basic. Thickness relies on connection etc., so is just a level 2 concept.) Future go theory will see higher levels, but that does not mean "less applicable". It just means that players wishing to become strong need to learn a bit more, e.g., will need level 3 concepts.

But you definitely run out of useful concepts rather quickly.

No, not as long as people like me are searching for mightier theory or an understanding of existing but still verbal theory.

What one is hitting is something different: the individual researcher will have to slow down his enlisting of concepts after a very short time. The few dozens of known concepts are written down quickly:)

emeraldemon wrote:Checkers has been solved

IIRC, the variant played as International Checkers is not solved yet. Right?

RobertJasiek wrote:

emeraldemon wrote:Checkers has been solved

IIRC, the variant played as International Checkers is not solved yet. Right?

True, only 8x8 checkers has been solved, not 10x10. I think my point still holds: you can go to the research website and "play the proof", showing each move as win, lose, or draw down the tree. But I don't think that would help you beat a clever player who doesn't follow a line you memorized.

link: http://chinook.cs.ualberta.ca/users/chinook/index.html

So the proof is only "weak" (without explanation of strategy).

RobertJasiek wrote:So the proof is only "weak" (without explanation of strategy).

That seems like a pretty strong proof to me... Strategy is not meaningful when talking about solving checkers.

palapiku wrote:You can never run out of strategic concepts, because you can always come up with more and more complicated, less and less applicable heuristics, and you can always give names to more and more tenuous abstractions that arise on the board.

But you definitely run out of useful concepts rather quickly.

Oh, ye of little faith!

RobertJasiek wrote:Future go theory will see higher levels, but that does not mean "less applicable". It just means that players wishing to become strong need to learn a bit more, e.g., will need level 3 concepts.

Consider the following demanding definition of usefulness: A truly useful theoretic concept is one that, if learned by a strong professional, would further improve his play.

If a concept helps weak players but has no effect on professionals, it means that, while useful to these weak players, the concept is replaceable by other kinds of go knowledge or skill, which professionals already have from other sources. In a sense, such a concept is redundant. Such concepts may help weaker players get stronger faster than with traditional instruction, but they don't go past what can be achieved with traditional instruction.

Under this definition, I suspect very strongly that none your theory is useful. What's more, it may actually be harmful. To quote Kageyama:

When the ladder becomes slightly difficult like this, there is a widespread tendency to give up, and wonder if there is not something like a triangle theorem, some mechanism one can apply and get the answer instantly. If you want to create such a thing it is not much trouble to do so, but having it will only prove destructive to your game.
[...]
Occasinally some periodical proudly announces that it has discovered a shortcut to reading ladders - some worthless white elephant with four or five dotted diagonals and heavy black lines. Even if you could understand it, it would not do your game the least good. Such things are ridiculous.

I suspect that your ideas such as the formula to estimate the influence in joseki are exactly this kind of white elephant, and may actually be harmful in the long run.

I will be delighted if I'm proven wrong by seeing a wave of Jasiek-school professionals demolish the current greats.