Tewari analysis.

[Toge]

Tewari means analyzing the position either by reducing the superfluous stones and deciding which side made more effective moves, or by changing the order of moves played to determine better shape.

I've had fleeting glimpses of positions that could easily be understood with tewari. Correct answer to odd move in invasion for instance respects the usual joseki sequence. Life and death problems are epitome of tewari. Poor move fails the objective while tesuji accomplishes it. "Almost correct" lines of play are often just played in wrong order.

Is shape study meant to give foundation to tewari analysis?

[Kirby]

It seems to me that tewari is about finding "good moves". It might be the case that many good shapes contain good moves.

[Bill Spight]

Is shape study meant to give foundation to tewari analysis?

It is pretty much the other way around. For instance:

$$ Bad Shape
$$ ------------------------
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . X O O . . . O . .
$$ | . . . X . X O . . , . .
$$ | . . . . . X . . . . . .
$$ | . . . X . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .

Click Here To Show Diagram Code

[go]$$ Bad Shape
$$ ------------------------
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . X O O . . . O . .
$$ | . . . X . X O . . , . .
$$ | . . . . . X . . . . . .
$$ | . . . X . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .[/go]

Tewari reveals Black's shape to be inefficient, although it looks like good shape.

Edit:

$$ Permutation
$$ ------------------------
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . 9 8 2 . . . 4 . .
$$ | . . . 1 . 5 6 . . , . .
$$ | . . . . . 7 . . . . . .
$$ | . . . 3 . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .

Click Here To Show Diagram Code

[go]$$ Permutation
$$ ------------------------
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . 9 8 2 . . . 4 . .
$$ | . . . 1 . 5 6 . . , . .
$$ | . . . . . 7 . . . . . .
$$ | . . . 3 . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .[/go]

In this sequence, - is plainly not good for Black.

(Sorry, I thought that this was well known.)

[daniel_the_smith]

Trying to understand Bill's position...

Of course this would be the normal way it would happen:

$$ Bad Shape
$$ ------------------------
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . 7 6 2 . . . 8 . .
$$ | . . . 1 . 3 4 . . , . .
$$ | . . . . . 5 . . . . . .
$$ | . . . 9 . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .

Click Here To Show Diagram Code

[go]$$ Bad Shape
$$ ------------------------
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . 7 6 2 . . . 8 . .
$$ | . . . 1 . 3 4 . . , . .
$$ | . . . . . 5 . . . . . .
$$ | . . . 9 . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .[/go]

If we reorder it a bit, 5, 6 & 8 are questionable, but 9 is clearly terrible.

$$ Bad Shape
$$ ------------------------
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . 1 6 8 . . . 4 . .
$$ | . . . 7 . 9 2 . . , . .
$$ | . . . . . 5 . . . . . .
$$ | . . . 3 . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .

Click Here To Show Diagram Code

[go]$$ Bad Shape
$$ ------------------------
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . 1 6 8 . . . 4 . .
$$ | . . . 7 . 9 2 . . , . .
$$ | . . . . . 5 . . . . . .
$$ | . . . 3 . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .[/go]

Am I even remotely close?

[aurik]

Also looking forward to "the answer".

Here's my guess:

$$ The jump attachment is small and overconcentrated
$$ ------------------------
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . 5 4 2 . . . 6 . .
$$ | . . . 1 . 7 8 . . , . .
$$ | . . . . . 9 . . . . . .
$$ | . . . 3 . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .

Click Here To Show Diagram Code

[go]$$ The jump attachment is small and overconcentrated
$$ ------------------------
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . 5 4 2 . . . 6 . .
$$ | . . . 1 . 7 8 . . , . .
$$ | . . . . . 9 . . . . . .
$$ | . . . 3 . . . . . . . .
$$ | . . . . . . . . . . . .
$$ | . . . . . . . . . . . .[/go]

attach-hane-extend from here is a silly idea. B gains very little, and W gets to reinforce his slightly-overextended position.

[Shaddy]

I find that if both players in a tewari analysis make strange moves, it's best to be wary. It only really comes out right for me if only one player makes strange moves.

[RobertJasiek]

Tewari means analyzing the position either by reducing the superfluous stones and deciding which side made more effective moves, or by changing the order of moves played to determine better shape.

Chapter 4.6 of Joseki Vol. 2 Strategy informs about four further kinds of tewari:

- assessing the difference of numbers of played stones
- forming pairs of stones (or plays) with equal meanings
- move reversion (comparing to another known move order)
- comparing a sequence or shape to other known (similar) sequences or shapes

[Joaz Banbeck]

... Life and death problems are epitome of tewari...

Not really, IMHO, just one of the essential ingredients.

The goal of tewari is to establish whether or not a stone is good, regardless of sequence. ( I think Bill is correct in saying that tewari contributes to the foundation of shape. )

L&D requires proper sequence.
Tewari aims to be independent of sequence.

Indeed, one might say that shape is what remains after tewari has condensed the wisdom of L&D by removing sequence.

[Kirby]

Tewari is kind of interesting in that respect... It makes me wonder: In go, is order important or isn't it?

I'm inclined to think that order is always very important - but then, how can we measure the benefit of tewari?

[daniel_the_smith]

Tewari, if anything, means order is even more important-- it's not enough to merely play stones in the right order. You have to play in such a way that there is, in retrospect, no order which makes those stones look bad*!

[*] Or, at least, worse than the other guy's moves...

[Li Kao]

AFAIK what tewari is based on is:
If you find a reordering of the moves so that each exchange of a black and white stone is either even or bad for the same side, and at least one of them is not bad for that side then the whole sequence is bad for the side.

This implies what Shaddy remarked: If both sides make inefficient moves in a reordering, that reordering isn't fit well for tewari.

It's main use is to find out if certain moves are efficient. It's not useful in life and death, since there the ordering of moves matters a lot, and there is a clear cut goal.

[Kirby]

Tewari, if anything, means order is even more important-- it's not enough to merely play stones in the right order. You have to play in such a way that there is, in retrospect, no order which makes those stones look bad*!

[*] Or, at least, worse than the other guy's moves...

I guess this is kind of where I am confused: If you play in such a way that no order makes the stones look bad, it seems to say that the order in which you play the moves is not important.

As I write this, though, I suppose there is the possibility that order in which you play moves could be important in that you force your opponent to commit to particular moves. The moves that you do to achieve this must, themselves be good in any order. I guess this makes me a little bit less confused.

[Joaz Banbeck]

... You have to play in such a way that there is, in retrospect, no order which makes those stones look bad...

Sounds like another definition of shape.

[John Fairbairn]

Tewari has absolutely nothing to do with life and death.

It has virtually nothing to do with order of moves.

It is a post hoc method of analysing a sequence, almost always a joseki, and is most often used by pros to compare one putative joseki with another, not to analyse good shape.

It means dissecting the moves (te wo waru, or te wo waridasu), that is breaking them up into small units. There are two main techniques. In one, you try to put the moves back together in another order (this is the only area where move order comes in, but it is purely hypothetical). This often reveals a nonsensical move (e.g. you end up having peeped at a bamboo point, although when you played this "peep" it wasn't then a peep and there was no bamboo joint). The other technique is to remove stones in black-white pairs (remembering to include prisoners). This often reveals glaring examples of overconcentration or redundant stones.

It's a bit like pulling legs off a spider to see if it can still walk.

It supposedly (with no proof) goes back to Dosaku.

[Kirby]

Is there a method to evaluate a tewari analysis that you have performed? That is, it seems possible to do it the wrong way. I like Shaddy's suggestion regarding playing "normal" moves for one side. Are there any other ways to ensure that your tewari analysis is correct?