Difference between revisions of "Joseki"

From HexWiki
Jump to: navigation, search
(Cleanup)
 
(26 intermediate revisions by 3 users not shown)
Line 1: Line 1:
The word "joseki" is borrowed from the game of Go because of the strong analogy between Hex and Go corner play. There are certain sequences of moves played nearby corners of the board that repeat very often among different games. The corners are considered the key areas of the board (by the [[Advanced_(strategy_guide)#The_minimax_principle|mini-max rule]] ). Hence often playing joseki is [[Forcing moves|forcing]] for the opponent to keep playing locally. Knowing josekis narrows down the number of options to choose from. Nevertheless, that doesn't make player's life very easy - picking a correct joseki is as difficult as picking a correct global move. Therefore, this subject is addressed mostly to the advanced players.
+
The word "joseki" is borrowed from the game of Go because of the strong analogy between Hex and Go corner play. There are certain sequences of moves played near [[Board#Corners|corners of the board]] that often recur in different games. The corners are considered the key areas of the board (by the [[Advanced_(strategy_guide)#The_minimax_principle|mini-max rule]]). Hence, playing joseki often [[Forcing moves|forces]] the opponent to keep playing locally. Knowing josekis narrows down the number of options to choose from. Nevertheless, that doesn't make the player's life very easy - picking a correct joseki is as difficult as picking a correct global move. Therefore, this subject is addressed mostly to advanced players.
  
 
== Joseki's logic ==
 
== Joseki's logic ==
When player A makes a move in the corner he dominates the territory around this corner. Assuming that he picked his best strategy, he should stick to it and always defend his domination in the area. However, that doesn't mean that player B should give up that territory. Playing somewhere else will sometimes lead to losing the initiative around previously mentioned corner, but there is a direct counterplay for player B. A game progresses in that manner: player B counterplays, player A holds on to his domination, player B counterplays again etc... This usually lasts 2-10 moves by each player depending on how far from an edge was the first move by player A.
+
When player A makes a move in a corner, she dominates the territory around this corner. Assuming that she picked her best strategy, she should stick to it and always defend her domination in the area. However, that doesn't mean that player B should give up that territory. Playing elsewhere will sometimes lead to losing the [[initiative]] around the aforementioned corner, so instead, player B counterplays in the corner. A game progresses in this manner: player B counterplays, player A holds on to her domination, player B counterplays again etc... This usually lasts 2-10 moves by each player, depending on how far from an edge the first move by player A was. When the final position is reached, player A has a fixed escape on one side of the corner and player B has a fixed escape on the other side (the second player's escape is always of shorter range).
When the final position is reached player A has a fixed escape on one side of the corner and player B has a fixed escape on the other side (second player's escape is always of shorter range).
+
  
Playing a joseki can often be delayed. This is a good choice if one isn't sure which joseki is favourable for them. However, it's important to make sure if one gets initiative back to play joseki. If opponent can make [[Forcing moves|forcing]] moves in other areas of the board then a chance to delay joseki is gone.
+
Playing a joseki can often be delayed. This is a good choice if a player isn't sure which joseki is favourable for them. However, it's important to make sure that one gets the [[initiative]] back to play joseki. If the opponent can make [[Forcing moves|forcing]] moves in other areas of the board then a chance to delay joseki is gone.
  
== Examples ==
+
In the below examples, when we say "Red has an escape on the bottom" or "Blue has an escape on the top", we mean a [[ladder escape]] for certain ladders (e.g., 2nd row ladder, 3rd row ladder). When we say "Blue has an escape on the bottom", we mean that Blue ''denies'' Red a ladder escape there, i.e., Blue, as the [[ladder handling|defender]] of the ladder along the red edge, will "win" the ladder.
  
=== 4th line josekis ===
+
== 4th row josekis ==
 
'''A1)'''  
 
'''A1)'''  
Red wants an escape on bottom and Blue wants an escape on top:
+
Red wants an escape on the bottom and Blue wants an escape on the top:
 
<hexboard size="9x9"
 
<hexboard size="9x9"
 
   coords="hide"
 
   coords="hide"
   contents="B 3:h6 B 1:f7 R 2:h7"
+
  edges="bottom right"
 +
   contents="R 1:g6 B 2:g8 R 3:f8"
 
   /><br/>
 
   /><br/>
  
 
'''A2)'''  
 
'''A2)'''  
Red wants an escape on bottom, Blue wants an escape on top '''and Blue knows that Red wants an escape on bottom'''.<br/>
+
Red wants an escape on the bottom, Blue wants an escape on the top '''and Red knows that Blue wants an escape on the top'''.<br/>
The move #3 is an improvement, because in some cases Blue might get an additional free escape on bottom.<br/>
+
The move #3 is an improvement, because in some cases Red might get an additional free escape on the top.<br/>
There is risk involved. If Red made an error on move #2 and in truth he wanted an escape on top he can correct the error by playing a different move #4.
+
There is risk involved. If Blue made an error on move #2 and in truth he wanted an escape on bottom he can correct the error by playing a different move #4.
 
<hexboard size="9x9"
 
<hexboard size="9x9"
 
   coords="hide"
 
   coords="hide"
   contents="B 5:h6 B 1:f7 R 2:h7 B 3:g8 R 4:h8"
+
  edges="bottom right"
 +
   contents="R 1:g6 B 2:g8 R 3:h7 B 4:h8 R 5:f8"
 
   /><br/>
 
   /><br/>
  
 
'''B)'''
 
'''B)'''
Red wants an escape on top and Blue wants an escape on bottom:
+
Red wants an escape on the top and Blue wants an escape on the bottom:
 
<hexboard size="9x9"
 
<hexboard size="9x9"
 
   coords="hide"
 
   coords="hide"
   contents="B 5:g6 B 1:f7 R 2:g7 B 3:f8 R 4:g8"
+
  edges="bottom right"
 +
   contents="R 1:g6 B 2:g7 R 3:h6 B 4:h7 R 5:f7"
 
   /><br/>
 
   /><br/>
  
 
'''C)'''
 
'''C)'''
If both players want an escape on top this position might be reached.<br/>
+
This position might be reached if both players want an escape on the top.<br/>
Blue keeps a future option of playing in the cell marked with an asterisk.<br/>
+
Red keeps a future option of playing in the cell marked with an asterisk.<br/>
This is a very risky plan by Blue, because Red has additionally got an escape on bottom. Blue should only play this joseki if he is certain that Red cannot use the escape on bottom. Otherwise a reasonable option is to give up the strong escape on top and go for position '''B)'''.
+
This is a very risky plan by Red, because Blue has additionally got an escape on top. Red should only play this joseki if she is certain that Blue cannot use the escape on top. Otherwise a reasonable option is to give up the strong escape on the bottom and go for position '''B)'''.
 
<hexboard size="9x9"
 
<hexboard size="9x9"
 
   coords="hide"
 
   coords="hide"
   contents="B 3:g6 B 1:f7 R 2:g7 E *:e8"
+
  edges="bottom right"
 +
   contents="R 1:g6 B 2:g7 R 3:f7 E *:h5"
 
   /><br/>
 
   /><br/>
  
 
'''D)'''
 
'''D)'''
If both players want an escape on bottom we get an unusual position where Red plays the last move of the joseki.<br/>
+
If both players want an escape on top we get an unusual position where Blue plays the last move of the joseki.<br/>
Blue's plan is to get a cell marked with an asterisk for free in future, so he shouldn't play there immediately. What he should do depends on the play in rest of the board.
+
Red's plan is to get the cell marked with an asterisk for free in future, so she shouldn't play there immediately. What she should do depends on the play on the rest of the board.
 
<hexboard size="9x9"
 
<hexboard size="9x9"
 
   coords="hide"
 
   coords="hide"
   contents="B 1:f7 E *:g7 R 2:h7 B 3:g8 R 4:h8"
+
  edges="bottom right"
 +
   contents="R 1:g6 B 2:g8 R 3:h7 B 4:h8 E *:g7"
 
   /><br/>
 
   /><br/>
  
 
<br/>
 
<br/>
=== 5th line josekis ===
+
== 5th row josekis ==
  
'''A)'''
+
=== Common ===
 +
 
 +
'''A1)'''
 +
<hexboard size="11x11"
 +
  coords="hide"
 +
  edges="bottom right"
 +
  contents="R 1:h7 B 2:h8 R 3:g8"
 +
  /><br/>
 +
'''A2)'''
 +
<hexboard size="11x11"
 +
  coords="hide"
 +
  edges="bottom right"
 +
  contents="R 1:h7 B 2:h8 R 3:i7 B 4:i8 R 5:g8"
 +
  /><br/>
 +
'''A3)'''
 
<hexboard size="11x11"
 
<hexboard size="11x11"
 
   coords="hide"
 
   coords="hide"
   contents="B 1:g8 R 2:i8 B 3:i10 R 4:h10 B 5:i7"
+
  edges="bottom right"
 +
   contents="R 1:h7 B 2:h8 R 3:i7 B 4:i9 R 5:j8 B 6:j9 R 7:g8 B 8:f10 R 9:i8 B 10:h9 R 11:e9 B 12:c10 R 13:c11 B 14:f8 R 15:f9"
 
   /><br/>
 
   /><br/>
 
'''B1)'''
 
'''B1)'''
 
<hexboard size="11x11"
 
<hexboard size="11x11"
 
   coords="hide"
 
   coords="hide"
   contents="B 1:g8 R 2:i8 B 3:f10 E *:j7 E *:h8"
+
  edges="bottom right"
 +
   contents="R 1:h7 B 2:h9 R 3:i9 B 4:h10 R 5:i10 E *:h8,i8"
 
   /><br/>
 
   /><br/>
 
'''B2)'''
 
'''B2)'''
 
<hexboard size="11x11"
 
<hexboard size="11x11"
 
   coords="hide"
 
   coords="hide"
   contents="B :g8 R :i8 B :f10 R 1:j7 B 2:h9 R 3:i10"
+
  edges="bottom right"
 +
   contents="R 1:h7 B 2:h9 R 3:i9 B 4:h10 R 5:i10 B 6:i8 R 7:j6 B 8:j7 R 9:g9 B 10:e10 R 11:e11 B 12:g8 R 13:h8"
 
   /><br/>
 
   /><br/>
 +
 +
=== Case-specific ===
 +
 
'''B3)'''
 
'''B3)'''
 
<hexboard size="11x11"
 
<hexboard size="11x11"
 
   coords="hide"
 
   coords="hide"
   contents="B :g8 R :i8 B :f10 R 1:h8 B 2:i10 R 3:h10"
+
  edges="bottom right"
 +
   contents="R 1:h7 B 2:h9 R 3:i9 B 4:h10 R 5:i10 B 6:i8 R 7:j6 B 8:j7"
 
   /><br/>
 
   /><br/>
 
'''B4)'''
 
'''B4)'''
 
<hexboard size="11x11"
 
<hexboard size="11x11"
 
   coords="hide"
 
   coords="hide"
   contents="B :g8 R :i8 B :f10 R 1:h8 B 2:i10 R 3:h10 B 4:h7"
+
  edges="bottom right"
 +
   contents="R 1:h7 B 2:h9 R 3:i9 B 4:h10 R 5:i10 B 6:i8 R 7:j6 B 8:j7 R 9:g9 B 10:e10 R 11:e11 B 12:g8 E *:h8"
 
   /><br/>
 
   /><br/>
'''C1)'''
+
'''C)'''
 
<hexboard size="11x11"
 
<hexboard size="11x11"
 
   coords="hide"
 
   coords="hide"
   contents="B 1:g8 R 2:i8 B 3:i9 R 4:j8 B 5:j9 E *:h8 E *:h9"
+
  edges="bottom right"
 +
   contents="R 1:h7 B 2:h9 R 3:j9 B 4:j8 R 5:g9"
 
   /><br/>
 
   /><br/>
'''C2)'''
+
'''D1)'''
 
<hexboard size="11x11"
 
<hexboard size="11x11"
 
   coords="hide"
 
   coords="hide"
   contents="B :g8 R :i8 B :i9 R :j8 B :j9 R 1:h9 B 2:f10 R 3:g10"
+
  edges="bottom right"
 +
   contents="R 1:h7 B 2:h9 R 3:j6 E *:g10,h8"
 
   /><br/>
 
   /><br/>
'''C3)'''
+
'''D2)'''
 
<hexboard size="11x11"
 
<hexboard size="11x11"
 
   coords="hide"
 
   coords="hide"
   contents="B :g8 R :i8 B :i9 R :j8 B :j9 R 1:h9 B 2:f10 R 3:g10 B 4:i7 R 5:j5 B 6:k5 R 7:h7"
+
  edges="bottom right"
 +
   contents="R 1:h7 B 2:h9 R 3:j6 B 4:g10 R 5:i8 B 6:j9"
 
   /><br/>
 
   /><br/>
'''C4)'''
+
'''D2a)'''
 +
 
 +
The following joseki is preferable (for Blue) to D2, because by playing 4 before 6, Red does not get the cell marked "+". If Blue played the equivalent of 4 after joseki D2, Red would get "+".
 
<hexboard size="11x11"
 
<hexboard size="11x11"
 
   coords="hide"
 
   coords="hide"
   contents="B :g8 R :i8 B :i9 R :j8 B :j9 R 1:h9 B 2:f10 R 3:g10 B 4:i7 R 5:j5 B 6:k5 R 7:h7 B 8:h8"
+
  edges="bottom right"
 +
   contents="R 1:h7 B 2:h9 R 3:j6 B 4:i6 R 5:i7 B 6:g10 E +:h8"
 
   /><br/>
 
   /><br/>
'''D)'''
+
'''D3)'''
 +
<hexboard size="11x11"
 +
  coords="hide"
 +
  edges="bottom right"
 +
  contents="R 1:h7 B 2:h9 R 3:j6 B 4:h8 R 5:j9 B 6:j8"
 +
  /><br/>
 +
'''D4)'''
 
<hexboard size="11x11"
 
<hexboard size="11x11"
 
   coords="hide"
 
   coords="hide"
   contents="B 1:g8 R 2:i8 B 3:i9 R 4:h9 B 5:i7 E *:j7"
+
  edges="bottom right"
 +
   contents="R 1:h7 B 2:h9 R 3:j6 B 4:h8 R 5:j9 B 6:j8 R 7:g8"
 
   /><br/>
 
   /><br/>
 
'''E)'''
 
'''E)'''
 
<hexboard size="11x11"
 
<hexboard size="11x11"
 
   coords="hide"
 
   coords="hide"
   contents="B 1:g8 R 2:h8 B 3:g9 R 4:h9 B 5:h7"
+
  edges="bottom right"
 +
   contents="R 1:h7 B 2:h9 R 3:i9 B 4:i8 R 5:g9 E *:g10"
 
   /><br/>
 
   /><br/>
 
'''F1)'''
 
'''F1)'''
 
<hexboard size="11x11"
 
<hexboard size="11x11"
 
   coords="hide"
 
   coords="hide"
   contents="B 1:g8 R 2:h7 B 3:h8 R 4:i7 B 5:i8"
+
  edges="bottom right"
 +
   contents="R 1:h7 B 2:g8 R 3:h8 B 4:g9 R 5:h9"
 
   /><br/>
 
   /><br/>
 
'''F2)'''
 
'''F2)'''
 
<hexboard size="11x11"
 
<hexboard size="11x11"
 
   coords="hide"
 
   coords="hide"
   contents="B 1:g8 R 2:h7 B 3:h9 R 4:i8 B 5:i9 R 6:j8 B 7:j9"
+
  edges="bottom right"
 +
   contents="R 1:h7 B 2:g8 R 3:i8 B 4:h9 R 5:i9 B 6:h10 R 7:i10"
 
   /><br/>
 
   /><br/>
 +
 +
== Obtuse corner play ==
 +
 +
In the [[Board#Corners|obtuse corners]], play is not yet as codified as in the acute corners, and there are no obtuse corner josekis per se. Nevertheless, there are certain obtuse corner plays that are common, and some responses are better than others.
 +
 +
This section is under construction; more material should be added here.
 +
 +
The first move in an obtuse corner area is usually on the [[Board#Diagonals|short diagonal]], typically on the 4th or 5th row.
 +
 +
=== 4th row opening ===
 +
 +
Red stakes a claim in the obtuse corner with this move.
 +
<hexboard size="7x7"
 +
  coords="none"
 +
  edges="left bottom"
 +
  visible="all area(a1,e1,g3,g7,a7)"
 +
  contents="R 1:d4 E a:b5 b:c5 c:d5 d:e5 e:b6 f:e4"
 +
  />
 +
This move is quite strong; it is already connected to the red edge by [[edge template IV1d]], it gives Red 2nd-to-4th and 3rd-to-5th row [[Switchback#A4 switchback|switchbacks]] for ladders along the bottom edge, and it also blocks a significant part of the blue edge. Blue does not have any very good intrusions into the corner. For example, if Blue plays at a, Blue gets 2nd and 3rd row ladder escapes, but since Blue is not in Red's template, Red gains the [[initiative]]. If Blue plays at b or c, Red can push to f, gaining significant territory, while Blue gets at best a 2nd row ladder escape. If Blue plays at e, Red has a good response at b, limiting Blue to a 2nd row ladder escape. A reasonable move for Blue is d. Red can respond as follows:
 +
<hexboard size="7x7"
 +
  coords="none"
 +
  edges="left bottom"
 +
  visible="all area(a1,e1,g3,g7,a7)"
 +
  contents="R 1:d4 B 2:e5 R 3:c4 E x:d6 y:b4 z:c5 w:d5"
 +
  />
 +
Note that Red is connected to the bottom edge by [[edge template IV2a]]. Red also has an escape for 2nd row ladders along the bottom edge, and some territory. Blue has nothing very valuable, so it makes sense for Blue to make another move. There are two things Blue may hope to accomplish: take away Red's ladder escape, or create a ladder escape for Blue.
 +
 +
If Blue wants to take away Red's ladder escape while keeping the [[initiative]], it seems reasonable for Blue to play at x; however, in this case, Red can reconnect at y, gaining significant territory and probably leaving Blue worse off. A better move for Blue is z. This forces Red to reconnect at w, and ''then'' Blue can play x. Red still gains some territory, but less than if Blue had played at x immediately.
 +
 +
'''A1)'''
 +
<hexboard size="7x7"
 +
  coords="none"
 +
  edges="left bottom"
 +
  visible="all area(a1,e1,g3,g7,a7)"
 +
  contents="R 1:d4 B 2:e5 R 3:c4 B 4:c5 R 5:d5 B 6:d6 R 7:b5"
 +
  />
 +
The corner is now [[settled region|settled]]; neither Red nor Blue has a ladder escape, and both Red and Blue have gained some territory, with Red's territory being more useful.
 +
 +
Alternatively, if Blue's objective was to gain a ladder escape, then Blue plays as follows. Both players end up with 2nd row ladder escapes and some territory, with Red's territory being more useful.
 +
 +
'''A2)'''
 +
<hexboard size="7x7"
 +
  coords="none"
 +
  edges="left bottom"
 +
  visible="all area(a1,e1,g3,g7,a7)"
 +
  contents="R 1:d4 B 2:e5 R 3:c4 B 4:b5 R 5:c6"
 +
  />
 +
 +
== See also ==
 +
 +
* [[Corner move]]
 +
 +
[[category: Advanced Strategy]]

Latest revision as of 20:09, 9 March 2024

The word "joseki" is borrowed from the game of Go because of the strong analogy between Hex and Go corner play. There are certain sequences of moves played near corners of the board that often recur in different games. The corners are considered the key areas of the board (by the mini-max rule). Hence, playing joseki often forces the opponent to keep playing locally. Knowing josekis narrows down the number of options to choose from. Nevertheless, that doesn't make the player's life very easy - picking a correct joseki is as difficult as picking a correct global move. Therefore, this subject is addressed mostly to advanced players.

Joseki's logic

When player A makes a move in a corner, she dominates the territory around this corner. Assuming that she picked her best strategy, she should stick to it and always defend her domination in the area. However, that doesn't mean that player B should give up that territory. Playing elsewhere will sometimes lead to losing the initiative around the aforementioned corner, so instead, player B counterplays in the corner. A game progresses in this manner: player B counterplays, player A holds on to her domination, player B counterplays again etc... This usually lasts 2-10 moves by each player, depending on how far from an edge the first move by player A was. When the final position is reached, player A has a fixed escape on one side of the corner and player B has a fixed escape on the other side (the second player's escape is always of shorter range).

Playing a joseki can often be delayed. This is a good choice if a player isn't sure which joseki is favourable for them. However, it's important to make sure that one gets the initiative back to play joseki. If the opponent can make forcing moves in other areas of the board then a chance to delay joseki is gone.

In the below examples, when we say "Red has an escape on the bottom" or "Blue has an escape on the top", we mean a ladder escape for certain ladders (e.g., 2nd row ladder, 3rd row ladder). When we say "Blue has an escape on the bottom", we mean that Blue denies Red a ladder escape there, i.e., Blue, as the defender of the ladder along the red edge, will "win" the ladder.

4th row josekis

A1) Red wants an escape on the bottom and Blue wants an escape on the top:

132

A2) Red wants an escape on the bottom, Blue wants an escape on the top and Red knows that Blue wants an escape on the top.
The move #3 is an improvement, because in some cases Red might get an additional free escape on the top.
There is risk involved. If Blue made an error on move #2 and in truth he wanted an escape on bottom he can correct the error by playing a different move #4.

13524

B) Red wants an escape on the top and Blue wants an escape on the bottom:

13524

C) This position might be reached if both players want an escape on the top.
Red keeps a future option of playing in the cell marked with an asterisk.
This is a very risky plan by Red, because Blue has additionally got an escape on top. Red should only play this joseki if she is certain that Blue cannot use the escape on top. Otherwise a reasonable option is to give up the strong escape on the bottom and go for position B).

132

D) If both players want an escape on top we get an unusual position where Blue plays the last move of the joseki.
Red's plan is to get the cell marked with an asterisk for free in future, so she shouldn't play there immediately. What she should do depends on the play on the rest of the board.

1324


5th row josekis

Common

A1)

132

A2)

13524

A3)

131472951115104612813

B1)

12345

B2)

71812136923104511

Case-specific

B3)

71862345

B4)

718126923104511

C)

14523

D1)

312

D2)

315264

D2a)

The following joseki is preferable (for Blue) to D2, because by playing 4 before 6, Red does not get the cell marked "+". If Blue played the equivalent of 4 after joseki D2, Red would get "+".

431526

D3)

314625

D4)

3174625

E)

14523

F1)

12345

F2)

1234567

Obtuse corner play

In the obtuse corners, play is not yet as codified as in the acute corners, and there are no obtuse corner josekis per se. Nevertheless, there are certain obtuse corner plays that are common, and some responses are better than others.

This section is under construction; more material should be added here.

The first move in an obtuse corner area is usually on the short diagonal, typically on the 4th or 5th row.

4th row opening

Red stakes a claim in the obtuse corner with this move.

1fabcde

This move is quite strong; it is already connected to the red edge by edge template IV1d, it gives Red 2nd-to-4th and 3rd-to-5th row switchbacks for ladders along the bottom edge, and it also blocks a significant part of the blue edge. Blue does not have any very good intrusions into the corner. For example, if Blue plays at a, Blue gets 2nd and 3rd row ladder escapes, but since Blue is not in Red's template, Red gains the initiative. If Blue plays at b or c, Red can push to f, gaining significant territory, while Blue gets at best a 2nd row ladder escape. If Blue plays at e, Red has a good response at b, limiting Blue to a 2nd row ladder escape. A reasonable move for Blue is d. Red can respond as follows:

y31zw2x

Note that Red is connected to the bottom edge by edge template IV2a. Red also has an escape for 2nd row ladders along the bottom edge, and some territory. Blue has nothing very valuable, so it makes sense for Blue to make another move. There are two things Blue may hope to accomplish: take away Red's ladder escape, or create a ladder escape for Blue.

If Blue wants to take away Red's ladder escape while keeping the initiative, it seems reasonable for Blue to play at x; however, in this case, Red can reconnect at y, gaining significant territory and probably leaving Blue worse off. A better move for Blue is z. This forces Red to reconnect at w, and then Blue can play x. Red still gains some territory, but less than if Blue had played at x immediately.

A1)

3174526

The corner is now settled; neither Red nor Blue has a ladder escape, and both Red and Blue have gained some territory, with Red's territory being more useful.

Alternatively, if Blue's objective was to gain a ladder escape, then Blue plays as follows. Both players end up with 2nd row ladder escapes and some territory, with Red's territory being more useful.

A2)

31425

See also