Getting started with shape

PeterPeter · Post by **PeterPeter** » Sun Mar 24, 2013 2:56 pm

EdLee wrote:The correct atari is the other direction, after which B has miai of (d) and (e)

Is this outcome good for black? I don't understand why.

Click Here To Show Diagram Code: [go]$$W $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . O . . X . O . X . . . | $$ | . . O , X . . . . , . . O . . X . . . | $$ | . . . . . . . . O . X . . . . . . . . | $$ | . . . O . . . . . O X . O . . . . . . | $$ | . . . . . . . . X O X X 2 3 . . X . . | $$ | . . . . . . . . . X O X O . . . . . . | $$ | . . . . . . . . . . O 1 . . . . . . . | $$ | . . X , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

billywoods · Post by **billywoods** » Sun Mar 24, 2013 3:29 pm

PeterPeter wrote:Is this outcome good for black? I don't understand why.

White's shape is littered with cutting points. Here's one possible continuation:

Click Here To Show Diagram Code: [go]$$B $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . O . . X . W . X . . . | $$ | . . O , X . . . . , . . W . . X . . . | $$ | . . . . . . . . O . X . a . . . . . . | $$ | . . . O . . . . . O X . W 3 5 . . . . | $$ | . . . . . . . . X O X X X O 4 . X . . | $$ | . . . . . . . . . X O X O 1 b . . . . | $$ | . . . . . . . . . . O O 2 . . . . . . | $$ | . . X , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Black has miai of a and b to connect out; white's marked stones are dead, provided black's stone at the top doesn't die.

Of course, white could just play

elsewhere and give the stone up:

Click Here To Show Diagram Code: [go]$$B $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . O . . X . O . X . . . | $$ | . . O , X . . . . , . . O . . X . . . | $$ | . . . . . . . . O . X . . . . . . . . | $$ | . . . O . . . . . O X . O . 4 . . . . | $$ | . . . . . . . . X O X X X O . . X . . | $$ | . . . . . . . . . X O X C 1 2 . . . . | $$ | . . . . . . . . . a O O 3 . 5 . . . . | $$ | . . X , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

...but then what was the point? White's group in the upper right is looking endangered, and a is a nice move for black to start chasing the central white stones later, perhaps.

PeterPeter · Post by **PeterPeter** » Sun Mar 24, 2013 5:28 pm

Bill Spight wrote:Here is another ancient game, a castle game, in which the table shape arises early in the play. But you can't say that the player aimed at making table shapes. I have included some other comments on shape.

Thank you for these comments on shape during the course of a game.

It seems that in the first half of the game, shape is more to do with the density of stones, rather than their exact arrangement? For example,

.

There are 2 examples of white making kosumis on the second line towards the corner (moves

and

), which were described as good shape and approaching the corner, but which were easily blocked by black. The resulting arrangement I understood to be generally better for black:

Click Here To Show Diagram Code: [go]$$W $$ . . . . $$ . . O . $$ . 1 . . $$ . 2 X . $$ . . . .[/go]

Is this good shape for white in these cases due to its height on the board, or its function in helping to secure white's base, or its strength in a contested area, or some other reason?

EdLee · Post by **EdLee** » Sun Mar 24, 2013 5:37 pm

PeterPeter wrote:Is this outcome good for black? I don't understand why.

But wait! There's more! B still has the wedge (f) -- this has been sitting here for this and other variations. You play with it.

Click Here To Show Diagram Code: [go]$$W $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . O . . X . O . X . . . | $$ | . . O , X . . . f , . . O . . X . . . | $$ | . . . . . . . . O . X . . . . . . . . | $$ | . . . O . . . . . O X . O . . . . . . | $$ | . . . . . . . . X O X X 2 3 . . X . . | $$ | . . . . . . . . . X O X O . . . . . . | $$ | . . . . . . . . . . O 1 . . . . . . . | $$ | . . X , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

amnal · Post by **amnal** » Sun Mar 24, 2013 5:49 pm

I think you're thinking about it all wrong. Just forget the entire concept of shape and look for good moves. Shape is an emergent set of behaviours in that certain layouts of stones turn out to be good moves more often than others. In principle, we don't play shapes because they're shapes, we play good moves and happen to notice the common shapes afterwards.

Of course, in real games shape turns out to be an important topic because the statistics turn out to be an excellent heuristic by which to adjust our reading. A strong player's eyes in a given situation will probably first be drawn to shape points, which are shape points specifically because he's played hundreds of games and recognises that they were often important.

It seems that in the first half of the game, shape is more to do with the density of stones, rather than their exact arrangement? For example, .

There are 2 examples of white making kosumis on the second line towards the corner (moves and ), which were described as good shape and approaching the corner, but which were easily blocked by black. The resulting arrangement I understood to be generally better for black:

Of course shape is important, but I think you're grasping too rigidly define something that isn't generally viewed that way and (I'd even perhaps say) isn't useful to view that way.

As above, lets ignore the entire idea of shape when analysing these kosumis and analyse the board purely in the context of good moves. In both cases, black threatens to cut off one of white's stones and white has decided he wants to prevent that by connecting out somehow. There are several ways to connect:

Click Here To Show Diagram Code: [go]$$W $$ . . O . . $$ . d e . . $$ . c f X . $$ . b O . . $$ . a . X . $$ . . X . . $$ . . . . .[/go]

Any of a,b,c,d,e,f will connect white's stones. Each have advantages and disadvantages. f, for instance, doesn't impact the corner but does damage black's stone slightly. e has more effect higher up the board but again doesn't impact the corner.

Now, in the context of the actual board position white decides that he's sufficiently strong on the left that he doesn't risk leaving himself thin and so doesn't need a move up there like e, d or f. On the other hand, he doesn't feel strong enough to push and cut so he'll just try to connect. He could play b, but why hold back when a also connects but also threatens a jump into the corner? Therefore, he plays a. This gives the maximum payoff per connectedness; if black tenukis, white can jump into the corner.

Now if we play lots of games and observe lots of results, then we can begin to make some observations about shape. We can see that in situations like this (where defining 'like' is an exercise in itself) white's best shape is a because it simply has the best followups - either being sente or allowing the best jump into the corner. It's a little weaker further up the left, if the board was different then maybe white would have a different best shape.

Your attempts to categorise shapes seem...well, too naive. Even with the strongest shape proclamations we can make, like empty triangles being bad, it's absolutely trivial to find many many situations in everyday real games where the shape is actually good. That's because shape is something far deeper than being just good or bad, and when making judgements based on shape you have to think about the constituent factors. For instance, in the example above white's kosumi is good shape because it gives the best payoff in this situation. More, situations like this are common enough for the outside kosumi to be a recognisable good shape for trying to get a good payoff. However, there are also lots of situations you might think are similar but are different enough for a different move to be best, and understanding shape must involve understanding this.

This is all very long winded, but it all still comes down in the end to playing shape being the same as playing the best move. Frankly, at your level I think by far the best exercise is to play lots of games (plus reviews) and let your understanding of shape develop by actually getting the statistical experience you need to make the judgements. As you read many different situations, good shapes naturally reveal themselves and become a part of your reading patterns. If you try to force a false understanding, you'll just end up playing bad moves by misunderstanding what a shape means. Of course studying shape may be important, but I think the real meaning of that will be clearer with more experience.

Bill Spight · Post by **Bill Spight** » Sun Mar 24, 2013 6:16 pm

PeterPeter wrote: Thank you for these comments on shape during the course of a game.

It seems that in the first half of the game, shape is more to do with the density of stones, rather than their exact arrangement? For example, .

Density is a factor, at least in bad shape.

You do not want to be overconcentrated. But exact placement matters. Compare, for instance, instead of

, plays on each of the four adjacent points.

There are 2 examples of white making kosumis on the second line towards the corner (moves and ), which were described as good shape and approaching the corner, but which were easily blocked by black. The resulting arrangement I understood to be generally better for black:
$$W
$$ . . . .
$$ . . O .
$$ . 1 . .
$$ . 2 X .
$$ . . . .
Click Here To Show Diagram Code
[go]$$W $$ . . . . $$ . . O . $$ . 1 . . $$ . 2 X . $$ . . . .[/go]
Is this good shape for white in these cases due to its height on the board, or its function in helping to secure white's base, or its strength in a contested area, or some other reason?

Both stones make connections as well as threaten to enter their respective corners. Remember, shape is about local efficiency.

That is why it is senseless to show configurations of stones of one color and call them good shape.

Bill Spight · Post by **Bill Spight** » Sun Mar 24, 2013 6:32 pm

@ amnal

While it may be true that the shape play is often the best play, I don't think that you can learn shape by learning the best play. It certainly did not work that way for me.

Learning about efficiency and shape really helped my game. More than just about anything else, other than learning to play the whole board.

Everybody is different. There are high dan players whose shape, well, leaves something to be desired. But they have developed other skills.

PeterPeter has his own path. He certainly starts interesting discussions. More power to him!

Uberdude · Post by **Uberdude** » Sun Mar 24, 2013 6:34 pm

PeterPeter wrote:
There are 2 examples of white making kosumis on the second line towards the corner (moves and ), which were described as good shape and approaching the corner, but which were easily blocked by black. The resulting arrangement I understood to be generally better for black:
$$W
$$ . . . .
$$ . . O .
$$ . 1 . .
$$ . 2 X .
$$ . . . .
Click Here To Show Diagram Code
[go]$$W $$ . . . . $$ . . O . $$ . 1 . . $$ . 2 X . $$ . . . .[/go]
Is this good shape for white in these cases due to its height on the board, or its function in helping to secure white's base, or its strength in a contested area, or some other reason?

I wouldn't call the move at b5 a good shape. I would call it a good move. As amnal said good moves are more important than good shapes. As you said that kosumi next to a 2 stone wall is, in the abstract, a bad shape for white. The important point it is connects c6 to c9 in sente (assuming black answers in the corner) which means white has sente to play a big move elsewhere. This way of connecting is stronger and maybe we can say it is better shape but it is gote which is a big minus:

Click Here To Show Diagram Code: [go]$$Wc Big picture $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X O . . . O . . O . . O . . . . | $$ | . . X , . O . . . , . . . . . , X . . | $$ | . . . X . . . . . . . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . . O O . . . . . 2 . X . . . . X . . | $$ | . . . . . . . . . . . . . X . X . . . | $$ | . . . , . . . . . O . . O . O X O . . | $$ | . . O . . . . . . . . O . . O X X . . | $$ | . . . . . . . . . . . . . . O O . . . | $$ | . . 1 X . . . . . X . . X . X O O . . | $$ | . . O . . . . . . . . X . . X . X . . | $$ | . . . . X . . . . . O O X . . . X . . | $$ | . . X , . . . . . , . . O X X , . . . | $$ | . . . . X . . . . . . . O . O O O . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

PeterPeter wrote:Around 100-200 games, I guess.

Add a 0 or two

Bill Spight · Post by **Bill Spight** » Sun Mar 24, 2013 6:48 pm

Uberdude wrote:PeterPeter, how many games have you played? It's all well good theorising about things, but nothing beats actually playing to build experience. After 1000 games you'll have a better feeling of shape (and after 1000000 even better!).

PeterPeter wrote:Around 100-200 games, I guess.

Uberdude wrote:Add a 0 or two

Geez, I only played 200 plus games in my first year.

Uberdude · Post by **Uberdude** » Sun Mar 24, 2013 6:57 pm

Bill Spight wrote: Geez, I only played 200 plus games in my first year.

Now we haz Internetz!

Life In 19x19

Getting started with shape

Re:

Re: Re:

Re: Getting started with shape

Re: Getting started with shape

Re: Getting started with shape

Re: Getting started with shape

Re: Getting started with shape

Re: Getting started with shape

Re: Getting started with shape