Chapter Thirteen. Theory Applying to Effective Play

13.1 Doing the necessary, or losing the plot?

The central character in Pushkin’s Queen of Spades is led to his downfall by overriding his cherished principle, of ‘not risking the necessary to gain the superfluous’. In go, it is often hard to understand how to distinguish the two. One aspect of improving your strength is to shed all unnecessary plays. In a sense this is more important than making good shape. The gain in making the correct shape may only be a couple of points, compared with the secondbest play. Some misdirected moves are almost completely wasted.

It is common to characterise inferior moves as 80% or 50% of full value, and so on. An amateur 1 dan probably plays 90% moves, on average over the game; top players operate at around 98% or 99% efficiency. These figures do assume that perfect play in go resembles what can be seen in professional go. This is simply a hypothesis, extrapolating from the current state of knowledge, and it is hard to see how to test it.

💡 What are the most common causes of wasted plays? High on the list are:

• playing to save a group that is already alive;
• threatening a group with a play that isn’t in fact sente;
• capturing stones that have no strategic meaning;
• defending territory that is badly located (e.g. openskirted);
• carrying on in a set sequence through momentum alone.

General classes of mistakes are safety plays (nothing safe about playing badly), miscalculations about the burden of proof (if a forcing play isn’t clearly forcing and clearly required right now, it is quite likely to be bad), and misconceptions about which are the key areas or stones.

13.2 123 and use of threats

{{Dia 801}} {{Dia 802}}

Both of these sequences allow White to escape. But in the left-hand diagram Black gains more outside influence. Since the exchange of White 1 for Black 2 in that diagram isn’t necessary, it should be omitted.

💡 123 Principle

Don’t play 1-2-3, just play at 3.

{{Dia 803}} {{Dia 804}}

(Left) There is no good reason to play 1 here. It loses Black a liberty, and a ko threat; and also some potential for later play with Black at 2 or one point above. (Right) A cross-cut: see p.103 for Black 3 connecting solidly after Black 1 atari. A very common case of the principle is: ‘don’t play atari and then connect’. That can look like planless play. Black 3 in this diagram is part of a plan, to sacrifice one stone, with White A, Black B, White C, Black D. That builds strong shape for Black in one direction. Quite generally, if your play 1 has an obvious answer, you should have play 3 already lined up.

💡 One of the key proverbs is don’t go back to patch up.

Effective play normally aims to generates sustainable forward momentum.

13.3 Miai and ABC

{{Dia 805}} {{Dia 806}}

A common phenomenon of fighting. White 1 sets up one of two good shape continuations with 3 (see 15.1 for more in the right-hand case).

{{Dia 807}} {{Dia 808}}

The mechanism behind such plays is one aspect of the Japanese term miai. After playing 1 in this position, first seen in 9.1, Black will be content with either of the plays 3.

{{Dia 809}}

Summing up, Black’s play at C makes A and B into miai, a pair of points of which Black can be sure of one. We can enunciate a further useful principle:

💡 Don’t play either of A-B-C or B-A-C, just play C.

13.4 Double-purpose plays

Killing two birds with one stone is a proverb in many languages. If you want your stones to work harder for you, place them where their purpose in life (or death) isn’t limited to just one future direction of play.

{{Dia 810}} {{Dia 811}}

In the left-hand diagram Black attacks White single-mindedly. On the right Black 2 sets up subsequent plays at A to attack, or B to build a framework. This is more reasonable. The points A and B are like miai (13.3).

{{Dia 812}}

If White challenges Black on the upper side with 3, Black 4 has a double purpose (attack White, build up the top right), and Black 6 has three aims (attack the White groups to left and right, and avoid getting shut in).

{{Dia 813}} {{Dia 814}}

A well-masked trap in the endgame (White 4 should be at 7). Black 9 sets up two kos, both dangerous to White (who has to find the first ko threat).

13.5 Forcing: playing for definite effect

{{Dia 815}} {{Dia 816}}

A forcing play is one that the opponent will answer, in practice. Here White 1 is a professional-level forcing play. When Black answers at 2, White has miai points at A or B to live. Without this exchange it may be possible later for Black to kill White (right). But once it is played the White group is definitely alive. (From analysis of the 1999 British Championship.)

It is quite tempting therefore for Black to ignore White 1, when it is played. However Black’s shape then is full of cutting points. If White follows up by pushing at 2 in the left-hand diagram Black will have a very unfavourable fighting position.

💡 Forcing plays are highly effective if they achieve something definite, retain the initiative, and can be abandoned once played. You should question the value of a forcing play if any of the following might be true:

• ignoring it is a real option for the opponent;
• it will provoke an exchange of plays that doesn’t do anything clear-cut > for you, or even benefits the opponent;
• it might be better later on to play another way in that part of the board;
• it wastes a ko threat;
• it might be answered in a way that resists your intention, or leads the game down an unexpected road;
• you feel some obligation to save the stone played, or may be drawn into a local fight that loses the initiative.

All strong players seem to be generally agreed on matters of shape, but the same cannot be said about forcing plays. Play forcing moves early, and the game will have an abundance of fixed shapes, that have no further flexibility. Fixed shapes were a feature of the games of the great champion Sakata Eio.

13.6 Probes: information-led effects

A probe is a play that makes the opponent reveal information. It is a forcing play in a sense, but of a different kind. After a successful probe you should feel your opponent has made some sort of commitment or concession, about which you were previously uncertain.

{{Dia 817}} {{Dia 818}}

White 1 gets the answer Black 2, meaning that White may later be able to live in the corner with B. Black could have answered at A, B or C also. Having discovered Black’s intention to emphasise the outside, White may be able to turn elsewhere on the board. Whether White returns to play at B, before Black suppresses White’s stone with C, depends on the rest of the board.

The right-hand diagram shows a further probe White 3. If Black answers with 4, White can live by playing at D (then Black F, White G). If Black became strong in the centre it is conceivable that Black would later answer White 3 with Black D. Then White at 4, Black at E sends White running out into the centre. If this is a real possibility White can play 3 to test Black’s reaction, without necessarily making life in the corner in gote.

13.7 Counting and self-criticism

An ineffective way to play is start or continue a strategy that has no chance of winning the game, even if it works in its own terms.

This may seem to be so obvious as not to require saying. On the other hand, unless you count the game, you may fall into this trap without realising it. If you do count carefully enough to reveal that you are a little behind in a game, there is still the question of what you do about it. Playing on in the hope that the opponent will make a mistake is a practical strategy; but not one which will lead you to much improvement (except in the endgame). All in all, when offered a chance to pursue a plan which leads down a cul-de-sac, you should shun it. Each play of yours should aim to put you ahead of the opponent; so that simply playing a passive big point or simplifying matters by starting the endgame is not an option for the other side.

💡 Evaluating the effects of middlegame plans

Consider three kinds of vigorous middlegame play that have been seen earlier in this book: cross-cut fights, invasions, and reduction plays. These are typical of actions one may take when apparently behind in the game. They still require some counting in order to assess their results.

This is most clearly necessary in the case of reduction plays (see 9.3). If the deepest reduction play one can safely make still leaves the opponent enough secure territory to win, this plan must be rejected. Other possibilities to be considered are: reduce with a deeper play and hope for the best (amateurish), invade deeply and challenge the opponent to kill you outright, or try to build up strength on the basis of an attack in another part of the board first.

In the case of invasions of extensions on the side, which was the topic of Chapter 10, the point of view of counting throws up an instructive paradox. The territory defined by the group invaded might be only six or eight points. The creation of a small living group inside might be worth the same again: total about 12 to 15 points. This is the value of a large endgame play, no more. The value of the biggest opening points is twice this much, and plays in the middlegame rarely drop much below 20 points (and are often considerably more valuable).

One has, though, also to count negative values for any weak groups created, in the range 10 to 20 points. This number can be explained: assume the opponent will play the equivalent of one substantial endgame play against them, in sente, before they are settled. An invasion that creates a weak group is very different from one that simply affects territory.

With that in mind, it can be seen that the true assessment of cross-cut fights, such as were seen in 7.1, is mostly to do with valuations of up to four weak groups created in them. The first example on p.96 resulted in two settled groups for White, a small insecure corner and a weak central group for Black. It was therefore favourable for White.

It seems that an acute sense of positional judgement does naturally link to objectivity about the position on the board, and to finding an effective plan of play, one that has some chance of winning the game if it succeeds.