Shape for DDKs -- and thickness, too!

Bill Spight · #1

I have prepared some thoughts and material on the basics of shape and thickness. I by no means intend this thread simply to reflect my thinking and opinions. Please join in the fun.

Shape for DDKs

Discussions about shape often start out by classifying some shapes as bad, others as good. But IMO the basics of shape come before such classifications.

What is shape? A shape is a local configuration of stones of the same color. Shape as a topic is about the efficiency or effectiveness of shapes. The study of shape, then, includes how shapes relate to stones of the opposite color. Shapes can be overconcentrated, so that they use more stones than necessary to achieve their effect; they may even be heavy, so that not only do they invest a number of stones, they do not achieve their effect and are weak. On the other hand, shapes can use too few stones to achieve their intended effect, and can be thin and weak.

Suzuki Tamejiro, the teacher of Kitani Minoru, emphasized three basic two stone shapes for beginners, the nobi, a solid extension to an adjacent point, the kosumi, an extension to a diagonally adjacent point, and the tobi, a one space jump. The most frequent moves in go form one of these shapes. Depending on the circumstances they often go by different names. Of course, in themselves we cannot say whether these shapes are good or bad; we have to take into account their surroundings. I will also emphasize these shapes, as well as another one, the keima, or knight’s move, an extension two points in one direction and then one point in a perpendicular direction.

Click Here To Show Diagram Code: [go]$$B Nobi $$ . . . . . . $$ . . . . . . $$ . . X 1 . . $$ . . . . . . $$ . . . . . . $$ . . . . . .[/go]

Locally the nobi is strong. It cannot be cut. It is also slow.

The next two diagrams show the nobi with an opposing stone.

Click Here To Show Diagram Code: [go]$$B Nobi (ii) $$ . . . . . . $$ . . . . . . $$ . . X 1 . . $$ . . . O . . $$ . . . . . . $$ . . . . . .[/go]

Click Here To Show Diagram Code: [go]$$B Nobi (iii) $$ . . . . . . $$ . . . . . . $$ . . X 1 . . $$ . . O . . . $$ . . . . . . $$ . . . . . .[/go]

Click Here To Show Diagram Code: [go]$$B Kosumi $$ . . . . . . $$ . . . . . . $$ . . X . . . $$ . . . 1 . . $$ . . . . . . $$ . . . . . .[/go]

The kosumi develops more quickly, but it is not as strong as the nobi. It can be cut if Black ignores the threat to cut. The next diagrams show the kosumi with one or more opposing stones.

Click Here To Show Diagram Code: [go]$$B Kosumi (ii) $$ . . . . . . $$ . . . . . . $$ . . X . . . $$ . . . 1 O . $$ . . . . . . $$ . . . . . .[/go]

Click Here To Show Diagram Code: [go]$$B Kosumi (iii) $$ . . . . . . $$ . . . . . . $$ . . X . . . $$ . . O 1 . . $$ . . . . . . $$ . . . . . .[/go]

This kosumi, AKA hane or block, is an attacking play, but it can be cut.

Click Here To Show Diagram Code: [go]$$B Kosumi (iv) $$ . . . . . . . $$ . . . . . . . $$ . . X X . . . $$ . . O O 1 . . $$ . . . . . . . $$ . . . . . . .[/go]

This hane is normally a very good play. Go proverb: Don't look, hane at the head of two stones.

Click Here To Show Diagram Code: [go]$$B Kosumi (v) Counter hane $$ . . . . . . . $$ . . . . . . . $$ . . O X . . . $$ . . . O 1 . . $$ . . . . . . . $$ . . . . . . .[/go]

Many players are unaware of this play, but it is normally a good one.

Click Here To Show Diagram Code: [go]$$B Tobi $$ . . . . . . $$ . . . . . . $$ . . X . 1 . $$ . . . . . . $$ . . . . . . $$ . . . . . .[/go]

The tobi develops even more quickly, but is weaker than the nobi or kosumi. White to play can cut it — but usually does not.

Click Here To Show Diagram Code: [go]$$B Tobi (ii) $$ . . . . . . $$ . . . . . . $$ . . X . 1 . $$ . . . . O . $$ . . . . . . $$ . . . . . .[/go]

This jump attachment is often seen.

Click Here To Show Diagram Code: [go]$$B Keima $$ . . . . . . $$ . . . . . . $$ . . X . . . $$ . . . . 1 . $$ . . . . . . $$ . . . . . .[/go]

The keima, or knight’s move, is quicker than the tobi, but, curiously, harder to cut.

Click Here To Show Diagram Code: [go]$$B Keima (ii) $$ . . . . . . $$ . . . 5 . . $$ . . X 2 . . $$ . . 4 3 1 . $$ . . . . . . $$ . . . . . .[/go]

White may cut with :w2:

and

, but

may be able to capture :w2:

in a ladder.

Click Here To Show Diagram Code: [go]$$B Keima (iii) $$ . . . . . . $$ . . . . . . $$ . . X . O . $$ . . . . 1 . $$ . . . . . . $$ . . . . . .[/go]

is often seen.

Bill Spight · #2

On studying shape

Aside from excellent material in books and online, professional games provide a wealth of material on shape, because almost all pro moves are locally efficient, and therefore are good shape.

You just have to examine them from that standpoint.

The basics of efficiency

A number of years ago I studied professional plays in local contexts. At first I categorized them in terms of the number of Black stones and number of White stones. But I discovered that a better classification was in terms of the total number of stones and the difference between the number of Black stones and White stones. What I discovered about efficiency can be applied even by beginners.

The first principle of efficiency is that the efficiency of a shape or move decreases, in general, with the total number of stones of either color in a local area. The main exception is when a move captures a number of stones or saves them. This is obvious on the whole board, as moves become worth much less at the end of the game than they were at the beginning. But it also holds true, as a rule, in local situations. As they start to become settled, it is often time to play somewhere else.

The second principle of efficiency is that the efficiency of a shape or move decreases, in general, with the difference between stones of different colors. The most frequent plays are those that start from an equal number of stones by each player, so that one player gets one stone ahead, or those that end with an equal number of stones, so that one player catches up. The next most frequent plays are those where a player adds a stone from two stones behind, or where he gets two stones ahead. Getting three stones ahead in a local area is often inefficient. Not that it never happens, but in general it is to be avoided.

The second principle is related to the first. For instance, suppose that the players have played in four different local areas of the board, one stone in each of three areas and five stones in the fourth. The plays in the fourth area are likely to be less efficient than the other plays. Suppose also that one player has played only in the fourth area, while the opponent has played one stone in each. In the fourth area that player is three stones ahead, but by the first principle as well as the second, his stones are likely to be less efficient that the stones of the opponent.

Bill Spight · #3

Strength and thickness

It is, in my opinion, important for DDKs to learn to make strong shapes and to play thickly. (I speak from the opposite experience. At that level I did not play thickly.) While it is true that strong shapes can be slow and thus inefficient, weak shapes invite disaster. Once a player has made thick, strong shapes, she can then use those shapes to attack the opponent’s groups. The attacking moves need not be, and usually are not thick, but thinner and quicker. Thus the attacker avoids developing too slowly. Many players bemoan the fact that they do not understand how to utilize thickness. If they had started out playing thickly and using that thickness to attack, they would not have that problem.

Because of the danger of playing slowly when making thickness, handicap games, particularly high handicap games, offer a good practice field to learn thick play and attack. A little slowness will not lose the game. I have usually recommended that players who wish to advance take handicaps of 5 stones or more. For DDKs wishing to learn about thickness and attack, I recommend taking handicaps of 7 - 9 stones.

Hattori Inshuku, founder of the Hattori School, wrote a multivolume set about handicap go openings called Okigo Jizai. In his high handicap volumes he shows how to make thickness and use it to attack. Though it was written two centuries ago, and shows it, handicap go has not changed all that much since then, especially high handicap go.

Next I will use some high handicap variations from Okigo Jizai to illustrate how to make thickness and how to use it.

Bill Spight · #4

A nine stone game from Okigo Jizai.

Enjoy!

Bill Spight · #5

A seven stone game.

Krama · #6

At first diagram I would play black 28 at N15.

Joelnelsonb · #7

Thank you very much for this! There can never be too much information on fundamentals.

Bill Spight · #8

Krama wrote:

At first diagram I would play black 28 at N15.

That may work out. But you should be aware that that play, a kosumi, generates a weakness for Black, since Black cannot afford to give up the N-15 stone.

Hattori's hane undercuts White's potential base while making a base for Black. Such plays are big.

oren · #9

Bill, have you seen this book?

http://www.amazon.co.jp/%E7%AD%8B%E5%A0 ... 4062192357

It's an interesting discussion on shape with good examples as well.

John Fairbairn · **#10**

Quote:

It's an interesting discussion on shape with good examples as well.

Yes. I await the book. So far I've only seen this on Yoda's blog but if the early promise is sustained this will be a major contribution to go theory for amateurs. It's a kind of tewari, but it seems to work for all contact fights.

The key word is sujiba, which is word Yoda made up. It denotes certain key liberties.

Imagine an independent line of connected same-colour stones. Let's say two. These have six liberties, but sujiba refers to the liberties on the wall sides, four in this case.

Imagine two lines, one white, one Black, of three stones facing each other a space apart somewhere away from the edge. Each line of stones has six sujiba, and three of them are shared.

The basic premise is that it is inefficient to play on a sujiba, so if you look at a completed position and see any stones on sujiba, you have spotted a likely inefficiency.

Yoda's technique is to illustrate his theory mainly by re-examining standard proverbs. With the caveat that I've only seen a few examples, it is impressive stuff. One example is the crosscut. Weak amateurs tend to respond by playing Atari on both sides, but when the opponent extends out of Atari, the end result is that the atari-er has inefficiently played on sujiba. So the proverbial advice to extend one side from a cross-cut is justified. Obviously, just like tewari, sujiba is not a magic bullet, but it does seem that if you treat walking on sujiba like walking on the cracks in pavements, you are much less likely to trip up.

He mentions that amateurs probably haven't heard of this, apparently implying that pros already know about it, but not under any special name. Given Yoda's track record of revealing major insights into pro techniques, this too seems plausible.

Bill Spight · **#11**

oren wrote:

Bill, have you seen this book?

http://www.amazon.co.jp/%E7%AD%8B%E5%A0 ... 4062192357

It's an interesting discussion on shape with good examples as well.

Looks interesting. Thanks.

EdLee · **#12**

Question for those more proficient in Japanese:

The bottom marketing quote:
「3目強くなることを私が保証します」(依田紀基)
Does it say -- I guarantee a 3-stone improvement,
or a 3-point improvement ?

oren · **#13**

John Fairbairn wrote:

He mentions that amateurs probably haven't heard of this, apparently implying that pros already know about it, but not under any special name. Given Yoda's track record of revealing major insights into pro techniques, this too seems plausible.

It's been a while, since I read the intro but my memory says that he mentions this theory is something he came up with while trying to teach. He runs a school, and he knew what good shape was, but no one knew how to really teach it. Then he noticed how virtually all good/bad shape could be taught with tewari and sujiba (which boils down to places that make empty triangles for you or your opponent. It's not just empty triangles of one color but for both players). I've gone through about half the book so far.

I think first was the videos for Igo Shogi channel and then the book.

http://www.nicovideo.jp/watch/1417099940

oren · **#14**

EdLee wrote:

Question for those more proficient in Japanese:

The bottom marketing quote:
「3目強くなることを私が保証します」(依田紀基)
Does it say -- I guarantee a 3-stone improvement,
or a 3-point improvement ?

It does look like 3 point improvement. I guess he doesn't expect it to help much.

Bantari · **#15**

oren wrote:

EdLee wrote:

Question for those more proficient in Japanese:

The bottom marketing quote:
「3目強くなることを私が保証します」(依田紀基)
Does it say -- I guarantee a 3-stone improvement,
or a 3-point improvement ?

It does look like 3 point improvement. I guess he doesn't expect it to help much.

3 points improvement for a pro might be more than 3 stones improvement for us.

John Fairbairn · **#16**

Quote:

「3目強くなることを私が保証します」(依田紀基)
Does it say -- I guarantee a 3-stone improvement,
or a 3-point improvement ?

Three stones (though from what base in this case I don't know). I

This point confuses even Japanese people. In the ordinary language the counter 目 counts territory and 子 counts handicap stones. But among go players, 目 in such phrases is accepted as referring to handicap stones (a similar one is 定石を覚えて２目弱くなる - learn joseki and become two stones weaker), to the extent that some go players will even read the correctly written 三子 as sanmoku rather than sanshi. Similarly a line of three stones will generally be called sanmoku (as in sanmoku no mannaka wa kyuusho - the centre of three stones is a vital point, or nimoku no atama wa mizu haneyo - hane automatically at the head of two stones) but we see shi (which some regard as correct) in nishi ni shite suteyo - make a single stone into a two stones then sacrifice.

Speculating, shi also means "death" and "four" and references to four are often changed to yo to avoid unpleasant associations. Also sanshi would sound confusingly like "three or four", so go players, who obviously use these terms more than the average person, would have a couple of good reasons to avoid shi. In the case of 'nishi ni shite suteyo' I would imagine the euphony of repetition prevailed.

Go players also have a few grammatical usages, e.g. of de and wo, which defy standard grammar books and confuse Japanese non-go players.

EdLee · **#17**

Quote:

Three stones (though from what base in this case I don't know).

Thanks.

Marketing folks having good fun, as usual.

aiichigo · **#18**

When I think about shape in Go, I think more about things like the table, the mouth etc. Rather than the relationships between two moves, or haengma as it is called in korean.

Good shape are formations that are easy to make eyes and hard to attack.

Bad shape is formations which are hard to make eyes and easy to attack.

So to make an eye in the middle of a board, you need to control 3/4 of the corners of a 9x9 intersection area. This is why the mouth shape and the table shape, are good shapes. And why playing in your opponent's vital shape positions can be really effective.

Bad shape are things like the empty triangle, because it is over concentrated and has a shortage of liberties problems.

So to me, to teach good shape you should give details of good shape, and then give examples of the haengma used to make them. Also it isn't enough to make good shape, you need to do it in sente as much as possible. Perhaps also explaining bad shape, and why it is bad.

In my opinion, efficiency isn't really directly linked to good shape, the mouth shape, is good shape, but it is really over concentrated and can be inefficient, depending on the situation locally.

In the 9 stone game, move 15, isn't really an attack, the white stone is very weak, rather it is a trick move, often used in high handicap games, which typically makes the black player panic and think he is under attack, and play passively, or defensively, allowing white to get an advantage. 9 Stone games are tough, because they typically include a lot of less than good play, because white has around 150 points of advantage to erode, white has to play a lot of what would normally be considered overplays. In truth, none of black's stones are under attack, there are so many stones around, they are all very strong, and white has played a weak stone surrounded by black's. I don't think white's shape is particularly good either, it has the advantage that it is not surrounded, but it is a stick with no eye potential, the white stone played has no way of forcing black into white's stick. It does make the black player play a slow move, making a heavy group with bad shape, poor eye potential, so itmight have to run, but it is still strong enough that the surrounding white stones become weaker. White then makes a bad shape with his one point jump. A good shape for white is to make a two point extension, but because it is a Handicap game, he doesn't really have that option.

In the 7 stone game, at move 8, Black makes a bad shape, I don't think white was really threatening to play there, and threatening the cut at N3, just forces white to fix, and removes future potential of the weakness.

A play at Q6 would have made good shape, it would have helped settle black's group, and marked out more potential territory, but a better move might be to play at the bottom part of the board, white could concievably live with it's 4 stones, and not only that the bottom is wide, so White wants to build there, a better way to use the cut would be to play on the left of the group, if white fixes then black makes a base on the bottom.

The point here is that as a result of the bad shape move, black didn't really get black anything, and forced white to fix a weakness.

Every move you play should be at least as good if not better than your opponents. If you play a bad move and your opponent plays a better one, then they get ahead, if this happens enough you lose. In this exchange, black gets the worse deal, he helps white to fix a weakness, he helps white get some thickness to expand along the bottom, and he doesn't gain a lot more for himself. This is often a characteristic of moves that create bad shape.

All this said, sometimes you have to play bad shape, often if you have to play a bad move, that makes bad shape, your opponent will have had to play a bad move to make it happen. In the seven stones game, move 35 an empty triangle, leads to move 36 another empty triangle. This exchange is about even. White's next move at move 37, is bad shape, white gets to live in gote, and Black gets to play a really good move.

Leading up to this there were many poor moves or bad shape moves, this is important, because making bad shape or poor moves, typically leads to more bad shape. If you have to play an empty triangle, you probably played a bad move 10 or 20 moves ago. In this game, again because it is a handicap game, white has to play unreasonable moves to make up for the massive lead black has. White's move 25, is complete overplay, Black naturally cuts it off, and normally white would just sacrifice, but white has to play whereever it is possible to get points. Most high handicap games are a matter of trying to get black to play weak moves that leave bad shape, which white can then attack, and probably kill a group or two.

However getting your opponent to play bad shape usually requires you to play bad shape, so such a tactic isn't very useful in an even game, as they will likely take advantage of your moves, or leave you no opportunities to do so. So playing good shape becomes very important.

Bill Spight · **#19**

aiichigo wrote:

In the 7 stone game, at move 8, Black makes a bad shape, I don't think white was really threatening to play there, and threatening the cut at N3, just forces white to fix, and removes future potential of the weakness.

A play at Q6 would have made good shape, it would have helped settle black's group, and marked out more potential territory, but a better move might be to play at the bottom part of the board, white could concievably live with it's 4 stones, and not only that the bottom is wide, so White wants to build there, a better way to use the cut would be to play on the left of the group, if white fixes then black makes a base on the bottom.

The point here is that as a result of the bad shape move, black didn't really get black anything, and forced white to fix a weakness.

is still joseki. The SGF file below shows a couple of variations with comments.

Quote:

If you have to play an empty triangle, you probably played a bad move 10 or 20 moves ago.

Good point.

Quote:

However getting your opponent to play bad shape usually requires you to play bad shape

That does not accord with my experience.

snorri · **#20**

John Fairbairn wrote:

Imagine two lines, one white, one Black, of three stones facing each other a space apart somewhere away from the edge. Each line of stones has six sujiba, and three of them are shared.

The basic premise is that it is inefficient to play on a sujiba, so if you look at a completed position and see any stones on sujiba, you have spotted a likely inefficiency.

John, is this what you are describing?

Click Here To Show Diagram Code: [go]$$B sujiba? $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . s O s X s . . . . . . . . | $$ | . . . , . . s O s X s . . . . , . . . | $$ | . . . . . . s O s X s . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . |[/go]

Shape for DDKs -- and thickness, too!

Who is online