$$Bcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . . a 0 4 8 . |
$$ . . . . . 1 5 6 . |
$$ . . . . . . 9 7 . |
$$ . . . . . . 2 . . |
$$ . . . . . . . . . |
$$ . . . . . . 3 . . |
$$ . . . . . . . . . |
My opponent played this sequence. I was sure that I could play a in the game, but it finished horribly. What should I play?
This question may look simple at first, but refuting the deviation is pretty difficult - there are no original sources on this to my knowledge.
$$Bcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . . a 0 4 8 . |
$$ . . . . b 1 5 6 . |
$$ . . . . . . 9 7 . |
$$ . . . . . . 2 . . |
$$ . . . . . . . . . |
$$ . . . . . . 3 . . |
$$ . . . . . . . . . |
First off, let's try a different answer to white 6. Black could consider playing 7 at a or b, but I believe the result would be bad. It can be considered, though, if the board is right.
Hane at a could lead to a complicated fight, probably involving an exchange, but I could not tell if it is better for either player.
$$Bcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . . 3 O O . . |
$$ . . . . 1 X X O . |
$$ . . . . . . 2 . . |
$$ . . . . . . O . . |
$$ . . . . a . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
The idea of b is to either magari-turn at 3 if white connect on the right, switching to the top instead of the right side all of sudden. This would be very unusual after blocking at 5 initially, and locally white gets a superior result. I presume that (apart from extending on the top) black would like to add a stone at a later to enclose white and get a reasonable outside shape. Either way black is not thick.
$$Bcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . a 2 O O . . |
$$ . . . . 1 X X O . |
$$ . . . . . . . 3 . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
If white instead opts to push from behind on the third line once more (at a) before returning to play 7, it is better than joseki for black. If white attempts to resist by jumping to a, black will not hane at 3 next, but push through at 2 and cut, involving a ladder.
In conclusion of this section, while an option to consider, I would not regard a different answer to the hane a refutation at all.
$$Bcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . 2 . O O O . |
$$ . . . . 1 X X O . |
$$ . . . . . . X X . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
After the deviation is played out, black could nobi at 1, with white jumping at 2. Seemingly, this is even because we returned to joseki.
$$Bcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . . . O O 1 . |
$$ . . . . a X X . . |
$$ . . . . . . . . . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
However, instead of the usual move at a, black might consider playing at 1 instead. This leads to a different joseki, with emphasis on territory on the right edge. Losing this option is a loss for black. This in effect means that white's early hane is superior to joseki, which is impossible by definition. Accordingly, returning to joseki is not sufficient as a refutation of the deviation.
$$Wcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . . . a O . . |
$$ . . . . . X X 1 . |
$$ . . . . . . . . . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
As a side note, the hane is sometimes played by white in a similar shape precisely when white is afraid of a black hane there instead of the usual nobi at a. This is an (uncommon) joseki.
So sum this up, we still have not found a viable refutation.
$$Bcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . b 1 O O O . |
$$ . . . . a X X O . |
$$ . . . . . . X X . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
Another option is to hane at 1. In this shape, white has the options of cutting at a or clamp at b. Hane underneath is impossible since black would be most glad to just nobi-extend.
$$Wcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . 4 2 . . . . |
$$ . . . 1 X O O O . |
$$ . . . . 3 X X O . |
$$ . . . . . . X X . |
$$ . . . . . . O a . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
After the clamp, descending at 2 is very dangerous. The sequence to 4 is the basic pattern likely to follow, but the fight is easier for white since black has no good response to a.
$$Wcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . . X O O O . |
$$ . . . . 1 X X O . |
$$ . . . . . . X X . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
If white cuts, a fight involving a ladder will follow, possible leading to an exchange, as explained in a different answer. This can get pretty messy if both sides keep playing the strongest moves.
$$Bcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . a 2 . . . . . |
$$ . . c 1 b O O O . |
$$ . . . . . X X O . |
$$ . . . . . . X X . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
Black is not forced to hane, he can consider a keima, too. Usually, white would attach at 2, leaving black with the options of a, b, c or tenuki. Since white was lowered to the second line, I consider this a superior for black.
$$Bcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . c . . . . . |
$$ . . . 1 2 O O O . |
$$ . . . b a X X O . |
$$ . . . . . . X X . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
In case white plays 2 instead, bumping against the black single stone, black should consider either covering at a or the strong nobi at b.
$$Bcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . b . . . . . |
$$ . . a X O O O O . |
$$ . . . c 1 X X O . |
$$ . . . . . . X X . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
In case black covers at 1, it is difficult to clamp at a, because white lacks an as effective attack on the 5 (instead of 4) black stones.
Crawling on the second line at b once or maybe twice is an option, in both cases probably leading to white gaining sente. Locally, the result is very bad compared to joseki, but white finished in sente twice, which is very valuable, too.
There is also the option to cut at c. This fight should be acceptable for both.
$$Bcm1
$$ -------------------
$$ . . . . 4 . . . . |
$$ . . 3 2 . . . . . |
$$ . 5 . X O O O O . |
$$ . . . 1 6 X X O . |
$$ . . . 8 7 . X X . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
In case black plays strongly, white will aim to exploit the increased cut aji, likely starting with the hane at 2. This way, white gets life while getting enclosed by a thin black group.
However, black can consider different moves (e.g. 5 at 6, or even 3 at 6).
$$Wcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . X a O O O . |
$$ . . . . 1 X X O . |
$$ . . . . . . X X . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
As per the comments, there is a very strong variation with white cutting immediately. If black cuts at a, it will return to the situation that requires a black ladder.
$$Wcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . X 3 O O O . |
$$ . . . 4 1 X X O . |
$$ . . . 5 2 . X X . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
Without said ladder, black will have to hane from the outside at 2. The sequence to 5 will follow, leading to a fight. Black's shape seems relatively poor so this should be carefully examined before playing the keima.
If black specifically wants to avoid a fight, he can always chose to play the simple nobi (6 in the previous diagram) to revert to the usual joseki.
On the other hand, despite the vast number of possibilities that arise from the keima, I believe this to be the best option in case black wants to resist and does not mind a fight.
So we could say that black has gained far more options than he lost by white's deviation, thus refuting white's move.
However, likely the early hane is a only tiny mistake.
First, let's notice that b in the following:
$$Bcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . . a 0 4 8 . |
$$ . . . . b 1 5 6 . |
$$ . . . . . . 9 7 . |
$$ . . . . . . 2 . . |
$$ . . . . . . . . . |
$$ . . . . . . 3 . . |
$$ . . . . . . . . . |
reverts to a well known joseki where 10 is played before 6.
It seems possible for Black to play a, answering the cross-cut with 3 (see below), but he needs a ladder to answer 4:
$$Bcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . 3 1 O O O . |
$$ . . . . 2 X X O . |
$$ . . . . 5 . X X . |
$$ . . . . . . O 4 . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
So:
If Black has the ladder… Since the previous position (at move 3) is very advantageous for Black, white will most probably clamp instead of cutting (as proposed by shujaa). In this case black can consider playing like this, and add a shape fixing move of his choice (N16 or O15), depending on the surroundings.
$$Bcm1 8 at 1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . 4 . . . . |
$$ . . 5 2 1 O O O . |
$$ . . . 3 6 X X O . |
$$ . . . . 7 . X X . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
Or he might prefer to stick with the standard joseki, playing b.
If the ladder is favorable to white… then a is not an option at all.
Last Remark: In this joseki variation, instead of a or b, N16 could also be a good (and safe) alternative, and there is also O17 (but it requires the same ladder). Tennuki has to be strongly motivated, because white playing b is huge.
This is in Josekipedia, just with a different move order. (I don't know how to link to a specific josekipedia variation, but taking the moves as numbered in your question and reordering 1, 2, 3, 4, 5, 10, you can see your marked stone as a bad move, with 2 ways to punish. A hane at P17 is standard for a 3-3 invasion without an approach stone, but you cannot ignore white's stone at R14!
Though it can feel too easy on the corner, P16 is joseki. White gets a big corner, but black has a nice wall and strong influence on the side.
$$Wcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . . . . . . |
$$ . . . 9 . 3 1 7 . |
$$ . . . . 4 X 2 5 . |
$$ . . . . . . 8 6 . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
If, as in @Stéphane's answer, Black insists on P17 it is equivalent to the above (see below), except black's move by pushing prematurely is an example of aji keshi (in fact, almost identical to the illustrative example of aji keshi on Sensei's Library). If Black pushes down instead of connecting then white can cut at P16. Of course, if white does have the ladder than she can cut at P16 immediately instead of the O17 clamp, which is much more painful for black.
$$Bcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . 4 . . . . |
$$ . . . 2 1 O O O . |
$$ . . . . 3 X X O . |
$$ . . . . . . X X . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
N.B.: White usually plays 3 (Q17) in my diagram before 5 (S16) just in case black brashly makes the P17 mistake. However, if black already had a partial enclosure with a stone at 017 then S16 would first is correct.
$$Wcm1
$$ -------------------
$$ . . . . . . . . . |
$$ . . . . 7 . . . . |
$$ . . . X . . 1 5 . |
$$ . . . . . X 2 3 . |
$$ . . . . . . 6 4 . |
$$ . . . . . . O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
$$ . . . . . , . . . |
Here white is alive though the corner is much smaller and doesn't extend as well into the top.
Black also has the option of blocking on the top if White chooses to hane underneath first. The variation below is from Josekipedia and listed as "even", though no source is cited.
$$Wcm1
$$ -------------------
$$ . . . . . . 9 . . |
$$ . . . . 8 7 4 3 . |
$$ . . . 0 . 2 O 5 . |
$$ . . . . . X X 1 . |
$$ . . . . . . . . . |
$$ . . . . . 6 O . . |
$$ . . . . . . . . . |
$$ . . . . . . X . . |
$$ . . . . . . . . . |
One would wonder here why Black would initially block the right side and then (seemingly inconsistently) switch to blocking on the top. Black should not do so without a reason.
a is not the joseki move b is do tewari black starpoint white kakari black pincer white invade 3-3 black blocks white nobi (move 10) black b then black hane and connect before jump
