Good Points and Bad Points to Play

    Keywords: Books & Publications

Good Points and Bad Points to Play
series: NHK lecture
By: Sonoda Yuichi
Publisher: NHK, 2004
ISBN10 4-14-016120-5
222 pp.

Good Points and Bad Points to Play: This book is based on a series of NHK lecture courses taught from October 2001 through March 2003, plus some additional material. As Sonoda explains, although amateurs are typically most comfortable when making territory or capturing enemy stones, these two tasks are appropriate as direct goals only later in the game. Earlier on, a player's main task is to set up the conditions so that the territory or stones will eventually be there to collect. Unfortunately, since this earlier task is much less concrete, it is harder to carry out. A seasoned professional can often just look at a position and see what to do; but an amateur player cannot expect to have that level of judgement.

To meet this problem, Sonoda continues, he has distilled some of his own sense of the game into a set of practical rules for finding good places to play as well as places to avoid. Though it is unreasonable to expect a set of rules alone to lead unerringly to the 100% moves, these can, he says, be expected to lead to moves scoring at least 80%; and they definitely will prevent outright blunders.

The rules come in three general categories, as explained and illustrated in the first three chapters. Chapter one explains how to use consideration of local numerical superiority to decide which friendly stones are in need of reinforcement and which enemy stones are ripe for attack. This often gets applied in conjunction with Sonoda's favorite precept from his earlier Go Strategy, that "playing near live stones is small."

Chapter two explains how shape considerations dictate where not to play. For instance, everyone knows not to make empty triangles, or other elementary bad shapes; and with a little foresight, if we expect to play at some point on the board very soon, we know not to play first at nearby points which would combine with the anticipated move to make bad shape. Sonoda gives some examples of this sort, in conjunction with some not quite standard bad shapes.

But he also carries things a step farther: In certain kinds of tactical battles, we can point to some small collection of possible moves, some of which are very likely to be necessary; but precisely which will depend on the opponent's responses. This implies that there is a whole surrounding corridor of points that, if possible, should not be played until the situation is resolved, because they might combine with necessary moves to create bad shape. This is a sophisticated way to play! We are shown examples in which the flow of battle includes a secondary struggle to avoid being pushed into the bad corridor, while at the same time trying to push the opponent there.

Chapter three explains how to judge which open areas are largest and most worth moving into, early in the game; and also how these considerations determine the proper direction of play in the ensuing battles.

Most of the examples in the first three chapters are from two, three or four stone handicap games, though the book is definitely not about handicap go. Maybe the issues are most starkly highlighted in such a context. At any rate, in Chapter four Sonoda works through five of his own games, to show that even (some of!) professional play can be understood in these same terms.

In retrospect, Sonoda's rules are nearly all about proper direction of play. Of course, we also need good continuations! Sonoda provides plenty of illustrative sequences, some starting from plays not conforming to his rules which, no surprise, don't work out well; while others, starting from moves of which he approves, succeed marvelously against a variety of responses.

Some of the advice in the book, I think, is really aimed at stronger players, because in some of the more difficult sequences he comments only on the most perplexing parts, taking it for granted that the reader is comfortable with standard tactics. (Also, he doesn't explain how to pick the 100% moves from the 80%-plus candidates....) Nonetheless,there is plenty for single-digit kyu players. His choice of theme-diagram positions is 100% interesting, and he maintains an encouraging tone throughout. The book is truly eye-opening, and Sonoda keeps his promise, made in the preface, that the reader will be put in touch with some of the deeper aspects of Go and come away with increased self-confidence.

Here are some positions examined in the book, along with incomplete summaries [2] of some high points of Sonoda's analysis.

Local Numerical Superiority:

An ordinary position?  

This is an ordinary-looking four stone opening, where W5 and W7 are a standard tactic used by White to torment a weaker Black player. But Sonoda finds this position problematic: if Black starts with a four stone advantage and, after only three moves, finds himself in such an uncomfortable situation, then something has gone badly wrong. Even if Black is strong enough to mount a successful counterattack, this is still a careless way for Black to play. So where is the mistake?

The key to the analysis is to imagine the diagonal line connecting the two x's, from lower left to upper right. Above the line, Black is outnumbered four to two! Black should never have let this happen, and in fact the mistake is B's second move B4. For after W3 B's upper left star point is already outnumbered two to one, and so Black should play his next stone above the diagonal, as reinforcement. The pincers of a, b, c, and d are suggested as ideal locations for B4, and the consequences of these moves are thoroughly explored for twenty-five or so whole-board diagrams.

Fight hard in an area of strength.  

Here is a three stone position in which Sonoda has already analyzed the play in the upper right corner. Now White plays W1. How does Black respond? The keys to the position are that W's two upper left stones are locally outnumbered above the x-to-x line, so Black not only can but must fight; and that the upper right Black group is alive, so Black shouldn't play near it. For instance, B2 at a would lead to more Black territory along the top, a wasteful enlargement of the already secure upper right group. Instead, Black should make an aggressive pincer as shown.

If White responds with W3 and W5, defending with B6 at B8 would be too small, since that would be too near a live group. Playing B6 as shown instead helps both the lower left star and especially B2, which are becoming vulnerable (and which are farther from live groups, and hence worth defending). If White responds with W7, then with B2 reinforced B8 is no longer merely defensive, but also threatens the severe Bb; so W9 is urgent. Then B10 is a fine conclusion to this sequence, enlarging B's sphere of influence on top as well as along the left side.

Sonoda explores other White responses to B2, with sequences more elaborate than this one. Black always comes out ahead.

Shape Considerations:

A no-play zone.  

This is from a four stone game in which White played a double kakari against B's upper right 4-4 stone. Sonoda has thoroughly discussed the shape considerations underlying B's initial responses and has also explained the details of the ensuing battle. White and Black have become closely intertwined, and it is now B's turn. The key to picking a good move is that the lettered points, a,...,h, form a no-play zone.

As Sonoda explains, either a or b would be a premature forcing move, a terrible waste of resource. "Timing is paramount." So they mustn't be played now, though either might be played soon. That being so, neither c,d,e, nor f should now be played, because a, if played later, would make an empty triangle in conjunction with c or d, and b would make an empty triangle in conjunction with e or f. Also, if c later became a good move, then g and h would be redundant moves adjacent to one's own bamboo joint.

Sonoda's point is not that these moves should never be made but, rather, that since they are not yet necessary, it is better to hold them in reserve until a definite need arises. Furthermore, if Black doesn't want these points, then he should try to induce White to play on them, since this is better than having White play on the points that Black actually wants.

Sonoda identifies B1 as the best play for Black at this moment --presumably the two White splitting? stones must be controlled or attacked-- though the two z points aren't bad. He says that just avoiding a through h already earns a passing grade.

No-play continued.  

He gives a number of variations showing how his principle leads to good results for B, no matter how White responds to B1 from the previous diagram. He says, for instance, that while W1 here may look frightening, it is not really so hard to handle. Note that now B2, the former c, is OK, because it is a reasonable response to White and, in particular, turns W1 --played in the no-play zone-- into a peep at a bamboo joint. The present sequence continues with Wa, Bb, Wc, Bd, We, Bf, with Black strong on the outside and W's four outside stones heavy and floating. He particularly likes the timing of the combination B4, B6 and B8. In all variations, Black ends up strong on the outside, with White ineffectual.

"Don't think."  

This position is from a three stone handicap game. In response to the double kakari from the two white+circle's, B plays the attach and extend joseki; and then W plays white+square. How should B respond?

B thinks, "there will be a moyo along the right side" and plays B1 to keep W on the bottom. But, says Sonoda, with that move, it is already too late for B to get a good result! Even though the displayed sequence is often played and sometimes recommended, B ends up with a concentrated shapeless and eyeless group of stones --which needs further care-- bordering a very open moyo which is easily reduced and invaded with moves like Wa, as Sonoda demonstrates. Meanwhile, unlike B, W ends up with a robust group which is ready for action.

Part of the trouble is that B is counting on developments whose outcome depends strongly on what happens over the board. That's the "thinking" to which Sonoda objects.

Don't think (continued)  

In fact, the shape move --Sonoda's strong recommendation-- is B1, and following B3 it is now B who has a live, robust group which can take care of itself; so B can fight hard wherever necessary, without having to watch his back. For instance, if W4, then B5 is played on the side rather than as a pincer on the bottom (staying away from his own live group); and after B9, B has now set up a sphere of influence on the left.

If, say, W continues with W10, B can take control of the game with Ba, stealing W's base. Sonoda supplies about fifteen more moves to show how this would work. But if W plays 10 as a defensive move at b, then B plays to seal W's group in, starting with c, simultaneously strengthening his influence along the left side.

Too many possibilities are given to include here, but in all of them B does well by following the flow instead of aiming for a preconceived result.

Another elementary bad shape  

In this even game example both sides have played san-ren-sei followed by a B stone at the bottom, a W stone at the top, and then black+circle at the very center. Next W reduces with W1 and B responds with B2, resulting in a crosscut fight. But, according to Sonoda, B2 is not really good; and furthermore, W can now exploit the relation with the marked stone to place B into bad shape: After W9, if black+circle were one space lower, at a, it would be peeping at W's hanging connection. But as it is, it does nothing.

Sonoda decrees this to be excellent shape for W, and gives two other unrelated examples in which a stone which might have peeped at a hanging connection is displaced by one space. In all three examples, black+circle and its counterparts perform little or no useful function in subsequent play, so they are inefficient stones, while the owners of the hanging connection go on to get excellent results [1].

However, in all three examples, Sonoda stresses the importance of actually making the hanging connection and not playing W9 (or its counterparts) as solid connections at b. For otherwise, the more compact shape seems vulnerable to counterattack, in the examples given. A solid connection might be inefficient, if a wider, hanging connection can be made without hazard. But the disadvantage of the solid connection, in the examples given, seems to go beyond that. Just why this should be so is not at all clear to me a priori. But it seems easy enough to find ways to follow the advice, in games.

(By the way, Sonoda does not say what B should have played instead of B2. It seems likely to me that black+circle might already have been a loose move. But this is a kyu player theorizing.)

Playing in the biggest open area:

A fuseki example.  

In this even game fuseki, after the first four moves are made, where should B next play? Sonoda identifies the left side as most valuable because it is widest, the top and right as tied for second because they are next widest?, and the bottom as least valuable because it is narrowest and both corner stones are low. So B should want to play on the left, and closer to the top than to the less valuable bottom. This leads to B1.

If W responds with W2 then B must continue on the left, for the point of playing there in the first place was to get more from it than W does. Again, where? The key is that B must play so as to force W to stay near the less valuable bottom, while B's position is higher up on the left. So, for instance, the joseki beginning with B3 at a, inviting W to block at b and extend up the side, while B gets the corner and a bit of the bottom, would be terrible.

Instead, B must pincer at 3. If W responds submissively with W4, then after B presses at the top, the moves B9 and W10 followed by Ba initiate a standard joseki which gives B a solid position facing up the left and W a position facing along the bottom, as B would like to see happen


But, if instead of W4 from the diagram above, W plays the marked diagonal extension in the present diagram, then B must be implacable about pushing W towards the bottom. Playing the variation beginning with B1 at x, for instance, would be a big mistake. Instead B plays as shown, followed by Ba, Wb, Bc, Wd and Be.

Sonoda remarks that if the lower right W stone were at f instead of the 3-4 point, then the top and bottom would be equally large; and he supplies some diagrams showing how play would differ.

Large knight's move in largest side.  

In this even game fuseki, once the initial four stones (unmarked) are played, the left side is largest, the right side second largest, the top third, and the bottom the smallest. So black+circle enters on the left, closer to the top than to the smaller bottom.

Sonoda devotes several diagrams to showing that W ends up with too small a share if he responds right in the corner or along the top. Instead, W must vigorously contest possession of the right side with a pincer, such as white+circle. Then if B plays B1 as shown, following the crosscut this leads to a standard joseki, with W10 followed by Ba, Wb, Bc, and the connection of Bd. W has done very well, stretched out along the side, with B confined to the smaller top.

In fact, White has done so well that Sonoda finds this diagram problematical on Black's behalf. "Some people may think that this joseki has given an even result, but here it has not. This is not a 50-50 division." This time it is Black who has made a mistake, by playing B1 submissively right in the corner. Instead, B must fight by jumping with B1 at b. Then after White defends the top, maybe at x, Black presses White low with B3 at y.

In the variation explored in the book, White extends down and gets a fair amount of third line territory along most of the left side; but Black gets outer thickness. Then a ferocious battle breaks out in the upper left corner, and about thirty moves later White has a small group on the top but Black owns the corner and has a developing central moyo.

Now it is Black who has done a little too well. Sonoda's conclusion is that ultimately, white+circle was a bit too close to Black and therefore too vulnerable to pressing, allowing Black a center build-up. So instead, it should have been a gentler pincer, at y. This development is reminiscent of the position described under the heading of "Don't attack, don't defend" in the review of Sonoda's Go Strategy.

See also



Another elementary bad shape  

[1] Dieter: (opinion) I just wanted to add that the actual position of black+circle might be key to the decision of which connection to make. If black+circle is here instead of one line lower, the peep at a becomes much more attractive in terms of shape.

JG: (another opinion) I was surprised by B2 in the original position. Since W1 is weaker than the nearby black stones, it feels wrong to attach to it. I would want to keep playing on a large scale, maybe with a high kakari at b, with the long-term aim of isolating W1 and keeping it weak. Can anyone explain B2 in the original diagram?

FredK: Actually, Sonoda himself said that that B2 wasn't really good, though he didn't detail why, apart from the given sequence. I've rephrased the text a bit to make this clearer. I was wondering myself what B should do instead.

Copyright discussion

PeterHB: Nice page, but I wonder at what point copyright worries will prevent this page expanding. 10% of a 200 page book would be 20 pages. Perhaps this is nearing 4%? I don't know what a reasonable excerpt is. Perhaps there are some people familiar with copyright issues about? Or is this just fair use, not matching the original really, more commentary than copying? To my mind, this is still commentary and fair enough at the moment, but it may develop too far.

FredK: I was planning to add one more diagram. I am concerned about copyright issues myself, but consider: The book contains over 425 diagrams, usually two per page, of which 10 or 11 would amount to under 3%, leaving about 415 more. [2]Actually most of the diagrams displayed here are not reproduced fully from the book but show only parts of longer sequences; and any reader would really also want to see Sonoda's additional diagrams associated with each board position. Further, the discussion supplied with each diagram on this SL page is highly summarized, paraphrased, and incomplete. I'm trying to show the kinds of things the author treats and give some hints of his ideas, not reproduce his entire story. Also I want to convey how well he presents his ideas to an amateur readership.

My model for these reviews is the sort of "featured book review" that I read in various scientific journals or in Sunday editions of well-known newspapers. I'm deliberately using this longer format rather than the kind of short review posted, say, by David Carlton, because so many readers here don't read Japanese, and have no way to browse through the books or get an idea of what they're really about. I have the utmost admiration for these authors and most definitely do not want to tread on their rights.

Bill: Fair use aside, ideas are not copyrightable, expression is. Translation, while technically not the same expression, is derivative and falls under the original copyright.

I gather that each of the chapters is based upon certain ideas, which the diagrams given here exemplify, and that the text presents these ideas, rather than being translations of excerpts.

However, the headings suggest that excerpts are being taken from each chapter and translated. It would be clearer that that is not the case if they said something about the ideas instead.

 FredK: Altered as suggested.

Good Points and Bad Points to Play last edited by velobici on August 20, 2008 - 15:03
RecentChanges · StartingPoints · About
Edit page ·Search · Related · Page info · Latest diff
[Welcome to Sensei's Library!]
Search position
Page history
Latest page diff
Partner sites:
Go Teaching Ladder
Login / Prefs
Sensei's Library