Life In 19x19

At the moment, I am losing too many of my fights, and I am not sure why. I play moves that look natural and sensible, then my opponent plays a reply that I was not expecting, and one or two moves later, I find out that I have come out the worst.

I think this is because I do not have much of an understanding of good shape. I have done a bit of reading around on SL and elsewhere, and most information is in the form of a series of examples of 'good shapes' on empty areas of the board. So, I am perfectly happy with what a table and a bamboo joint looks like now. The issue is applying this in a game, where one tends not to spend ones time making isolated shapes in the middle of the board.

I am looking for some context. When a few of your stones come into contact with a few of your opponent's stones, do you think: "How can I make a table shape here?", and then your stones will look after themselves? Or is it more a case of looking for moves that achieve some basic objectives (connecting or cutting) and then checking that they form a 'good shape'?

What is so special about the table shape? Which way round is it supposed to go, and what is its function when pointing in a certain direction?

I know that an empty triangle is usually bad, but in the middle of a fight, that idea gets overshadowed by the more pressing concerns of keeping my stones connected and not getting boxed in.

Is there a book that explains this stuff at a ddk level? I had a browse through Shape Up, but most of it is a bit beyond me.

Peter, very good questions, as usual.

PeterPeter wrote:I think this is because I do not have much of an understanding of good shape.

I'm 99.999% certain you're correct.

PeterPeter wrote:When a few of your stones come into contact with a few of your opponent's stones, do you think: "How can I make a table shape here?"

No, not necessarily. I start with an objective, locally and/or globally, then I look for the move to achieve it.
This is where tesuji (including life-and-death) problems are important:
to cut, connect, escape, live, kill, reduce, get sente, probe,
create ko threats, create good follow-ups (miai points),
make good shape (sorry, self-referential), make bad shape for the opponent, etc. etc.

Bruce Wilcox's interactive lessons, "Contact Fights," helped me a lot. I bought it from him.
It used to be shipped on a CD, but these days maybe he has an option to download?

In general (dangerous, I know), I've noticed three groups of people:
- street fighters learn by fighting, fighting, fighting; mostly "street smart" experience (Chinese Go servers, Tygem, etc.);
- theorists try to derive many things from first principles, and/or from reading Go books;
- osmosisers [sic] -- sounds too funny? -- take lessons from a good teacher (say, a good pro), and absorb good shapes over time,
from pro moves, from joseki study, from reviews, etc.

Probably, there's some combination of all three in most of us?
Your questions are obviously part of the eternal thread discussion, "How to study shape."

Greetings,
if you study at the Guo Juan Internet school, the lectures, group lessons, private lessons or whatever talk about shape all the time.
Best wishes,
Buri

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!).

Around 100-200 games, I guess. Shape seems like such a subtle concept; you don't realise that the shape you are making is worse than your opponent until the end result slaps you in the face. I could do with a few pointers. I like to learn an idea first, then practice it in my games.

EdLee wrote:[..]

Bruce Wilcox's interactive lessons, "Contact Fights," helped me a lot.

Ah, nice to know …

I bought it from him.

… because I did, too, recently. Sadly … so little time …

It used to be shipped on a CD, but these days maybe he has an option to download?

[..]

I got ’em via email.


Greetings, Tom

Shape basics:

The following are generally good locally.

Click Here To Show Diagram Code

[go]$$B Tap
$$ . . . . . .
$$ . . . . . .
$$ . . X 1 . .
$$ . . . O . .
$$ . . . . . .
$$ . . . . . .[/go]

Click Here To Show Diagram Code

[go]$$B Hane
$$ . . . . . .
$$ . . . . . .
$$ . . X . . .
$$ . . O 1 . .
$$ . . . . . .
$$ . . . . . .[/go]

OC, you can also extend, which gives you the tap pattern.

Click Here To Show Diagram Code

[go]$$B Hane at the head of two stones
$$ . . . . . . .
$$ . . . . . . .
$$ . . X X . . .
$$ . . O O 1 . .
$$ . . . . . . .
$$ . . . . . . .[/go]

Go proverb: Don't look, hane at the head of two stones.

OC, the "Don't look" is an exaggeration. Sometimes it is right to extend.

Click Here To Show Diagram Code

[go]$$B Extension
$$ . . . . . . .
$$ . . . . . . .
$$ . . X O . . .
$$ . . O X 1 . .
$$ . . . . . . .
$$ . . . . . . .[/go]

Go proverb: "Extend from a crosscut."

That avoids "atari, atari!"

Many exceptions, though.

Click Here To Show Diagram Code

[go]$$B Extension
$$ . . . . . . .
$$ . . . . . . .
$$ . . a X 1 . .
$$ . . O O b . .
$$ . . . . . . .
$$ . . . . . . .[/go]

Note that a play at either "a" or "b" would allow White to play at 1, making a good shape.

Click Here To Show Diagram Code

[go]$$B Counter hane
$$ . . . . . . .
$$ . . . . . . .
$$ . . O X a . .
$$ . . b O 1 . .
$$ . . . . . . .
$$ . . . . . . .[/go]

Many players are unaware of this play, but it follows from the above. If Black plays at "a" or "b", White can play at 1, making good shape.

Click Here To Show Diagram Code

[go]$$B Solid connection
$$ . . . . . . .
$$ . . . . . . .
$$ . . O X 1 . .
$$ . . . O X . .
$$ . . . . . . .
$$ . . . . . . .[/go]

Avoids atari, atari.

Click Here To Show Diagram Code

[go]$$B Turn
$$ . . . . . . .
$$ . . . . . . .
$$ . . . X X . .
$$ . . O O 1 . .
$$ . . . . . . .
$$ . . . . . . .[/go]

Surprisingly big.

That should get you started.

You might also take a look at these pages on SL. http://senseis.xmp.net/?BillSpight%2FBasicClosePatterns

PeterPeter wrote:What is so special about the table shape?

Most shapes you can make with four stones are somewhere on a spectrum between loose and solid. The table shape is a nice middle ground. It's very difficult to cut or squeeze or prod at, but doesn't involve any horrible empty triangles, and also contains some eyeshape. Perhaps if you laid out some other shapes you can make with four stones on a board you'd see why the table shape was usually better than them.

(You might like to start simpler: try making a few shapes with two stones, and ask yourself in what cases they're good shapes and in what cases they're bad shapes. You'll soon see that it's very situation-depedent: when is a knight's move bad, and when is it good? When is a nobi bad, and when is it good? How are these affected by distance from the edge, or presence/absence of opponent's stones? These are pretty hard questions. Try to get a feel for them, then maybe move on to three- or four-stone shapes.)

PeterPeter wrote:When a few of your stones come into contact with a few of your opponent's stones, do you think: "How can I make a table shape here?", and then your stones will look after themselves? Or is it more a case of looking for moves that achieve some basic objectives (connecting or cutting) and then checking that they form a 'good shape'?

Neither. I look for moves that achieve some basic objectives (connecting or cutting or similar), and then play them. They very often form good shape just naturally, because "good shape" is close to synonymous with "local efficiency". Sometimes they also don't. That's fine too.

PeterPeter wrote:I know that an empty triangle is usually bad, but in the middle of a fight, that idea gets overshadowed by the more pressing concerns of keeping my stones connected and not getting boxed in.

If an empty triangle is the only move that works, or the move that works best, play it. The reason that an empty triangle is usually bad is that there is usually a better move - but not always, so if you've done your reading, you don't need to worry about following or not following this proverb.

Post games you've lost for reviews. Then we can tell you exactly what you're doing wrong

PeterPeter wrote: What is so special about the table shape?

Here is a game with two --count'em-- two table shapes early in the game.

PeterPeter wrote:At the moment, I am losing too many of my fights, and I am not sure why. I play moves that look natural and sensible, then my opponent plays a reply that I was not expecting, and one or two moves later, I find out that I have come out the worst.

There are a lot of points to cover in these statements.
First is "fights" and "losing them": I had a game where I "lost" every fight and won by 20 points, the reason was simply that my opponent loved to rush with frenzy in all my positions and make a very small two-eye-group there. All his attacks were unreasonable, I could have played a game of whole-board fights with him (most likely he had hoped for it) or I could just let him live overall and win solid on points.
To know whether or not to fight in a given position or how much you can give away to stay ahead is in most cases a matter of experience and counting (positional judgement). It will do you no good to focus on that in the DDK-range.

Second: "natural and sensible looking moves". A very complex topic and it would be easier when you could post your game and comment on the moves which you thought were natural and sensible. In a nutshell we as (not high-dan) amateurs don't play the "sensible" move too often. It has a lot to do with lack of positional judgmeent (see above) and lack of reading skill (Tesuji and Life-and-Death problems). The simplest method to work on playing more natural and sensible moves in the DDK-range is to solve Tesuji and Life-and-Death problems. And to solve them regulary. And repeat them.

Third comes the "not expected reply" and the "worst" outcome. Actually this is the same as above. More Tesuji and Life-and-Death problems will surely help you overcome this. Good shape will not necessarily. The reason is that a lot of formations fall into "good shape" but in a fight only a few (maybe just one) will be the "correct" shape (not necessarily the "good shape"). See the game Bill Spight posted, although the bamboo joint is a "good shape" it would not be the "correct" shape in this one situation.
In my opinion good shape or the correct shape in a local context has to be resilient against attacks, respectively offer enough flexibility to settle your group (bonus points when it also threatens your opponent stones ; ) ). For me the best way to learn such shapes is - again, I begin to sound like a broken record - to solve Tesuji and Life-and-Death problems (especially the former). You will experience that a lot of problems are solved through shape weaknesses like a liberty shortage or a move that threatens to cut at two places. By repeatedly solving such problems you will start to become aware of weak shapes and thus play stronger shapes more naturally. You may not know how they are called then but, seriously, who cares? =D

Bill Spight wrote:Here is a game with two --count'em-- two table shapes early in the game.

This position, immediately before black makes a table at A, is instructive.

Click Here To Show Diagram Code

[go]$$Wc
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . O . . X . O . X . . . |
$$ | . . O , X . . . . , . . O . . X . . . |
$$ | . . . . . . . . O a Z . . . . . . . . |
$$ | . . . O . . O . . . . . O . . . . . . |
$$ | . . . . . . . . X . X . . . . . X . . |
$$ | . . . . . . . . . . . . O . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . X . . . . . , . . . . . b . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

Firstly, how black purposely leaves the very big point around B to instead secure his group, which looks to have mostly escaped to the centre anyway.

Second, how effective the table shape is in that position.

Which still leaves me wonder about black's thinking process.

Was his idea to 'make good shape', and this was the best shape he could find?

Or did he have specific tactical reasons for playing at A (strengthen the marked stone?), which happily coincided with good shape anyway?

Perhaps I am splitting hairs, or over-thinking this.

Bill Spight wrote:Shape basics:

The following are generally good locally.

$$B Tap
$$ . . . . . .
$$ . . . . . .
$$ . . X 1 . .
$$ . . . O . .
$$ . . . . . .
$$ . . . . . .

Click Here To Show Diagram Code

[go]$$B Tap
$$ . . . . . .
$$ . . . . . .
$$ . . X 1 . .
$$ . . . O . .
$$ . . . . . .
$$ . . . . . .[/go]

$$B Hane
$$ . . . . . .
$$ . . . . . .
$$ . . X . . .
$$ . . O 1 . .
$$ . . . . . .
$$ . . . . . .

Click Here To Show Diagram Code

[go]$$B Hane
$$ . . . . . .
$$ . . . . . .
$$ . . X . . .
$$ . . O 1 . .
$$ . . . . . .
$$ . . . . . .[/go]

OC, you can also extend, which gives you the tap pattern.

Thank you. I think it is these attachments that stump me the most, as there are usually 4 default responses to consider (stretch or hane, on each side), and with so much open space around, it is not easy to see where the fight is going to go.

PeterPeter wrote:Firstly, how black purposely leaves the very big point around B to instead secure his group, which looks to have mostly escaped to the centre anyway.

If black doesn't play the capping move at , black will still get out to the centre and probably live, but much more uncomfortably - white will be able to continue to chase black while building up his moyo in the upper left, and the white group in the upper right will be able to chase in sente and make more shape for itself. The cap by black makes black's group stronger, stops the expansion of white's moyo, threatens a big invasion / reduction in the upper left very soon (notice defends), and - as black's group is now stronger, and in sente to boot - white's group in the upper right is weaker, and will get chased.

PeterPeter wrote:Second, how effective the table shape is in that position.

The table shape connects most of black's stones (notice that, if black tenukis, white can think about cutting) in sente, and makes eyespace. Look at the sequence through to 30, thinking about the following things:

• What happens if white plays at K14? Why is black's cap at good shape here? Where does white actually play, and why?
• Why does black play at K15? "To make a table shape" is not the right answer! takes a liberty off a white stone, so is sente; it also connects some of the black stones together. then takes another liberty off a white stone, so is sente, and connects all of black's stones. Compare this to if black had played and at, say, L14 and L16.
• Compare the strength of black's groups after and . Compare white's upper-left moyo after and . Notice that white got something out of black's sente moves too.
• White threatens to cut right through the table shape. Why doesn't black care?

As others have indicated one doesn't make shape just for the sake of making shape. Shape is an indicator of local efficiency. If you see a play that makes good shape, that is a point in its favor. If you see a play that makes bad shape, or is such that the opponent can force bad shape, that tells against it. A sense of shape can help you to evaluate a position or variation. It can suggest plays for you to make.

Shapes with a number of stones, like the table shape, rarely arise because the player aimed at making them. If he did, that would be somewhat questionable. The table shape is nothing special. It arises in response to the opponent's play. To take things to the extreme, to make the table shape while allowing the opponent to take three big plays elsewhere on the board would be ridiculous.

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.