3-4 point low approach diagonal
Return to parent joseki article: 3-4 point low approach
The 3-4 point low approach, diagonal is a classic response to the low approach. It is also known as "Shusaku's diagonal". This move is steady and prevents White from pressing at the 5-4 point. Previously, it was perceived as a slow move in the mid 21st Century, since there is no urgent need for White to respond, but strong AI programs have caused a resurgence in its popularity.
|Table of contents
Moves are listed by frequency in professional games, which is sensitive to whole-board position. Bolded moves are currently considered joseki moves based on professional usage in games after 2018.
- t, 3-4 Point Low Approach, Diagonal, Tenuki - (joseki) (common) (AI favorite) (intermediate)
- a, 3-4 Point Low Approach, Diagonal, 2-Space Low Extension - (joseki) (common) (AI favorite) (beginner)
- i, 3-4 Point Low Approach, Diagonal, 3-Space Mixed Extension - (joseki) (modern) (beginner)
- b, 3-4 Point Low Approach, Diagonal, 3-Space Low Extension - (joseki) (traditional) (intermediate)
- c, 3-4 Point Low Approach, Diagonal, Small Knight Extension - (joseki) (traditional) (beginner)
- l, 3-4 Point Low Approach, Diagonal, 2-Space Mixed Extension - (situational)
- d, 3-4 Point Low Approach, Diagonal, 4-Space Low Pincer - (situational)
- e, 3-4 Point Low Approach, Diagonal, 3-Space Low Pincer - (situational)
- f, 3-4 Point Low Approach, Diagonal, Double Diagonal - (traditional) (obsolete)
- g, 3-4 Point Low Approach, 2-Space Low Pincer - (situational)
- h, 3-4 Point Low Approach, Diagonal, 4-Space High Pincer - (situational)
- j, 3-4 Point Low Approach, Diagonal, 3-Space High Pincer - (situational)
- k, 3-4 Point Low Approach, Diagonal, Knight's Move Cap - (rare)
- m, 3-4 Point Low Approach, Diagonal, 4-Space Low Extension - (traditional)
- b, 3-4 Point Low Approach, Diagonal, Tenuki, Two-Space Low Pincer - (joseki) (common) (intermediate) (AI favorite)
- a, 3-4 Point Low Approach, Diagonal, Tenuki, Knight's Move Press - (joseki) (common) (intermediate) (traditional)
- d, 3-4 Point Low Approach, Diagonal, Tenuki, Three-Space Low Pincer - (joseki) (intermediate) (traditional)
- c, 3-4 Point Low Approach, Diagonal, Tenuki, Three-Space Low Extension - (situational)
- e, 3-4 Point Low Approach, Diagonal, Tenuki, Four-Space Low Extension - (situational)
- f, 3-4 Point Low Approach, Diagonal, Tenuki, Four-Space High Extension - (situational)
- The most common White reply is tenuki (~37%). Often times, both players will tenuki from this corner because the pincer is not strictly sente. The two-space pincer is most popular (Diagram 1), but White can tenuki again since the approach stone is light and Black needs at least another move to finish the corner. If Black needs four stones to capture one of White's stones, hopefully White is able to get suitable compensation elsewhere on the board. Alternatively, White can move out into the center.
- In Edo-period games, in Diagram 2 was standard, with White probably responding at a. Later on, the press went out of fashion, but it is still seen in professional games games (see also: Basic Instinct in Joseki 1).
- In special circumstances, it may be possible for Black to play a three-space pincer (Diagram 3).
- Although it is not joseki, it is also common for professionals to play an extension along the side in response to something interesting on the left side of the board.
- White's most common local response is to create a base. In modern go, the most common and simplest base in the two-space low extension, which is solid with few weaknesses. Historically, it was less common because White feels slow, especially after Black extends to a or kicks with b. However, the AI revolution has elevated the two-space extension to status of a good move. By avoiding weaknesses, White will be able to play more aggressive in the future.
- In the 21st century, the three-space low extension was most popular (Diagram 2). The three-space extension discourages the kick at d, which would give White an ideal spacing.
- In exchange, the three-space low extension has an obvious weakness at z, but White is not afraid of an isolated invasion in the early game. The reason for this is that White can easily handle (Diagram 3). While Black cannot invade immediately, the presence of additional supporting stones could change the situation, and White must remain mindful of the weakness. In fact, there is no efficient way for White to fix the weakness in one move, so strong AI programs evaluate the three-space low extension less favorably.
- The small knight extension with is a traditional joseki that was more popular before the AI revolution. The small knight aims to build a framework on the upper side of the board, and it is most effective when used in conjunction with additionally friendly stones on the top side. If Black responds peacefully with at a, at m completes the joseki with White in gote. It is also possible for Black to fight with at b, which goes on to illustrate that the whole-board context is an especially important factor.
- In the post-AI era, White can play the three-space mixed extension if they hope to develop a fast framework. This is currently favored over the three-space low extension, although White still retains the invasion point at z. However, White has a better connection with the three-space mixed extension.
- The two-spaced mixed extension is also seen in the post-AI era, but it was poorly advised in traditional joseki dictionary. The counter-approach at z is effective and the jump at y is sente.
- The double diagonal (also known as the Aigosumi joseki) is Edo period move that is largely obsolete. However, it can be a valid method to develop a framework on the top side of the board when there are existing stones.
- The four-space low extension is a classical play, that isn't really part of the modern style except in special cases. It is often played in relation to stones on the upper right side. It is easy for black to invade and Black would continue with a or b.
- It is difficult to say whether pincers are considered joseki. They are rarely addressed in several major joseki dictionary, and practically speaking the pincer functions as if White chose to tenuki. Nonetheless, the pincers are surprisingly common in professional games, and they often serve as an extension from a position on the bottom left, a splitting move, or an early reduction stone intended to disrupt a framework.
- The hoshishita (point below a star point) is most common in professional games and is discussed in a traditional chinese joseki dictionary. This functions as a splitting move that also threatens a follow-up at z. However, Black has nothing to fear, since a and b are miai. Black can treat the stone as if White played tenuki.
- The three-space low pincer is another option seen in professional games. This pincer is more active than the leisurely four-space pincer, and there is a greater likelihood black might respond with a move like a, but still there is no significant concern for Black (who can tenuki yet again). Realistically, White may play this way to cool down the left side. It is difficult for Black to pincer the stone, and m feels cramped. Similarly, White doesn't have a compelling follow-up either.
- The two-space low pincer is realistically only played when White has support on the left side, such as a friendly stone at w. Without support, the stone is too close and vulnerable to a pincer after Black exchanges - in sente. This pattern was often seen in Edo period games.
Traditionally, the two-space extension was felt felt to be somewhat cramped if White kicks at y and approaches at (Diagram 1). To mitigate this, it was popularly believed that White should play a three-space extension to avoid becoming overconcentrated. If White approaches at , the one-space jump makes the three-space extension efficient spacing (Diagram 2).
However, strong AI programs have a different perspective. Modern go emphasizes sente and fast pace. The two-space extension is good because White can and should tenuki in response to , in order to take a large point (Diagram 1). White's group is strong and Black cannot do anything that threatens the vitality of this group in one move. While the extension is a big move, it should be gote.
We can compare this with the three-space extension, where the approach is much more severe given the defect at x. White must respond locally, and the result is that Black obtained the large point in sente (Diagram 2). The consequence of this is that Black is able to take the next large point with , which puts Black slightly ahead.
The concept of taking sente is critical for understanding the purpose of playing the two-space extension. In Diagram 1, responding locally to the approach is one of White's worst options. The slide at h is especially dubious, and it is heavily criticized traditional joseki dictionary because Black clearly obtains a better result once Black presses down with i. This sentiment is still true after the AI Revolution, and Shibano Toramaru reiterates this criticism in 2021. He also calls the jump out to j "slow-paced". Ultimately, the rationale for playing the two-space extension is so White can tenuki in the future. Perhaps White would have been better playing the three-space extension if White intends to respond to a local move.
Historically, the knight's move press was the most common move after tenuki, but it has declined dramatically in popularity ever since the AI Revolution. In the book Joseki Revolution?, Shibano Toramaru explains that is a move that primarily develops a moyo and allows Black to acquire solid territory, which is an unfashionable style of play that professionals are less inclined to pursue in recent years. In an example similar to Diagram 2, Shibano argues that the two-space pincer is often similarly good at developing a moyo. If Black tries to move out the stone, White will naturally enclose the top side.
Shibano's second point is that Black can tenuki in response to the small knight press (Diagram 1). This is in contrast with traditional thinking, whereby Black rarely chooses to tenuki after the knight's press (~15%). If Black responds normally at a, White has many forcing moves they can use to develop a menacing wall in sente. Therefore, Black can use this opportunity to disrupt White's moyo-centric plan, counting on the fact that is slow for White to capture the stone immediately.
Despite the reduced popularity of the small knight press after the AI revolution, it is important to note that it is often still a very good move. There are many board positions and circumstances where the press is the top choice of strong AI programs. Han Sangcho vs. Cho Hanseung (2022) (Diagram 1) is an example where is the best move and the two-space pincer at w could be considered a strategic mistake. In this situation, the pincer at w is unreasonable because White will have another weak group to handle once Black moves out the stone. Importantly, White's three-stone wall in the upper left corner is under attack and Black is very strong on the left side. The resulting fight is likely to become quite favorable for Black.
To describe it in simple terms, perhaps there isn't a very compelling reason for White to pincer when Black's left side is strong.
|Efficiency and aji of capturing
A second major consideration for choosing between the press and pincer is the efficiency of Black's stones if White chooses to tenuki.
Traditionally, one reason why the two-space pincer was less popular was because covering with (Diagram 1) leaves notoriously bad aji. Although the stone is provisionally captured, it can (A) come back to life if the outside becomes thin, (B) result in a cut and fight, or (C) give White many free forcing moves on the outside. It was believed that that Black would rather capture cleanly with after the knight's move press (Diagram 2), which takes cares of the aji and future problems.
Historically, sometimes Black would opt for a loose enclosure (Diagram 3), which allows White to live inside but eliminates the problem with bad aji.
Shibano Toramaru writes that two changes in thinking have occurred thanks to AI. The first is that it is slow for Black to capture cleanly with after the knight's press (Diagram 2), and strong AI programs prefer a loose capture with x or y. The second innovation is the small knight enclosure (Diagram 4), which makes it more challenging for the stone to live, yet also gives Black excellent shape on the outside. Shibano praises this move and cites it as one of the reasons why he is more inclined to play the pincer in recent years.
When White has significant potential towards the left side, White can play an extension on the fourth line to emphasize development. The typical extensions on the third line are defensive and easily pressed down, so they are not particularly good moves for expanding large moyos or frameworks.
The knight's move extension is an especially solid option, and White has no need to fear the approach at z because White is extremely thick in the area. Conventionally, z is thought to be an overplay when Black lacks regional support.
Interestingly, AI considers the parallel diagonal at b to be a reasonable move in this board position.
The two-space high extension at c is also an option, but players should note that the shape is thin and has more weaknesses compared to other similar moves.
White's low approach and Black's diagonal answer were a key feature of the nineteenth-century Shusaku fuseki. With the advent of komi go a pincer became the common answer instead of the diagonal, because it was considered faster for Black and hence a better way to overcome komi. Recently, the diagonal came into fashion again for both sides.
-  Frequency statistics were obtained from ext Waltheri's Go Pattern Search using the full database restrained to a local search (accessed August 2022). Common joseki in contemporary go have been re-ordered and placed at the top.
-  Kogo's Joseki Dictionary (accessed August 2022)
-  Dictionary of Basic Joseki (1977) - Yoshio Ishida
-  Fuseki Revolution (2021) - Shibano Toramaru
-  围棋定式大全 (2006) - Liao Yusheng