Today's post's theme revolves around strategy based on the 3-3 point, as per Joachim's inquiry. The example game in question was played yesterday in the Pandanet European Team Championships: I defeated Catalin Taranu of Romania by resignation, getting revenge for my loss to him in the European Go Congress 2010. The game this time was very good, with few apparent mistakes for both sides. I received commentary for the game by An Younggil 8 dan professional right after the game; some bits of what I write here come from there. I play black.
More sharp-eyed readers might notice that I differentiate between writing about black in first person and in third person: when I write in first person, I am reflecting on my thoughts during the game, and when in third person I'm looking at the position now, after the game.
The game started with a fairly common joseki in the upper left corner. Usually I play 11 one line further along the side, but I decided to consider this situation a special case. A main feature of the 3-3 point is that it doesn't easily work together with surrounding positions, being there deeply stuck in the corner. I thus decided to play the narrow and solid extension of 11, considering the black groups on the left side to be wholly separate entities.
White then played the more unconventional corner enclosure of 12. Even after it, however, the lower right corner is still open to a black invasion. 12's function is rather to improve white's influence in the centre and on the right side. It is because of this function that I almost instantly played at A to counter white's plans - then I decided that just for once, I could do something different, and chose the black corner enclosure of 13 instead.
With white 14, I was again going to respond at B without thinking, B being the most normal answer. White might play at C then, and the game would still continue normally - however, I chose to play the more interesting move of 15 instead, building the black lower left corner and limiting the expansion of white's lower right corner. The black moves of 13 and 15 show quite well the way I tend to use the 3-3 point: preventing the opponent's expansion while taking profit for myself.
White's double corner approach of 16 is then inevitable, and the following, normal joseki was played without much thought. Taking the corner territory and leaving the gaps of white A and B on the right side seemed good enough for me.
White chose to play 32 and 34 to prevent a big black framework from emerging on the left side. I then played 35 and 37 to create some opportunities on the right side, and then took the corner with 39 and follows. A crucial problem with this plan was, that black was supposed to get one more living group on the right side anyway - after white 50, the black stones of 35 and 37 don't exactly seem useful.
An Younggil pointed out that the following sequence might be good for black:
With the sequence starting with 1, black makes sure that white's right side doesn't get too big, and then taking the corner with 11 feels good enough. White doesn't have a very good way to prevent this sequence from happening, either.
The black defense of 51 might have been a bit timid, if not exactly bad. White then went to force the black corner to live starting with 52. Quite funnily, though, black lives unconditionally after white 60. White A - black B - white C - black D ensures that black lives in a double ko. After 60, however, I was rather perplexed as to what play next. Here's a good opportunity for the reader to think for themselves as to what to do, and then compare their choice with mine.
I deemed the left side to be the last big are of the board, and went to stake claims starting with black 61 (threatening to connect under, playing at 62 next). The sequence up to 67 then is natural - I was a bit happy that white chose to push with 68 to 73, letting me ensure the large lower left territory. A remarkable note here is that now the 3-3 stone is perfect, preventing any possible invasion in the corner. For example, a black 4-4 stone instead would still hold the threat of a white 3-3 point invasion.
After white played 74 to build the centre, I chose 75 without counting if it was actually necessary - I merely judged to myself that it could not be captured. Might have been a misjudgement, who knows. Black A would be an easier way of reducing the white centre, at the same time building pressure on the white left-side group.
White's attack with 76 and 78 comes naturally. With black 79, I was actually planning on attacking the white group in the lower right corner, in a similar fashion as Fujisawa Shuko attacked Kato Masao's framework in an earlier game review I did.
White 86 might be the losing move.
If white played 1 here, black could still fight back, but even after 8 it is not yet clear that the black group will survive. Black does have some space for his group to live in, and there's already a possibility for one eye on the right side, but the situation is uncertain nevertheless.
The game proceeded as here, black getting some eye shape for his group. It actually seems like white is in more trouble, with problems in his shape at A and B. Yet, after white 86 in the last diagram, he didn't really have a chance to do anything else either.
White showed good fighting spirit, being ready start the ko at 99. Instead of rushing straight in and beginning the ko immediately, I played 101 first to create threats. White's decision to end the ko is probably the correct one, even though black seems fairly happy to capture three stones in the centre with black 107. Furthermore, the black lower right corner is not dead until white also plays 114 - in the meantime, black got three moves elsewhere on the board. My assessment of the situation was that if I I can now remove all uncertainties in the centre, I will win on points.
I originally played 17 to force white to play at 19, letting black connect underneath (playing between 18 and 21). White again shows good spirit with fighting back starting with 18, for if black can connect he will have too much territory. I decided not to take any unnecessary risks, and let the white group get away, getting the largest endgame move of 41. After this, there's not really anything white can do to turn the game around. The game continued for some moves still - the full sgf file can be found here.
Next up for me is the Finnish Championship final next weekend. There's going to be a game relay on KGS with the account FinFinal, as well as a EuroGoTV video relay. This game would imply that I'm in good shape for the decisive games, but of course I'll have to be careful not to slip up.