Tuesday, September 27, 2016

Round 2 of 2016 US Tak Open

With the second round of the Swiss stage complete, the participants are really beginning to distinguish themselves. While still pretty flexible, the top 32 is starting to shape out. Quite a number of the names at the top here are Playtak and tournament veterans, including Simmon, Kakaburra, Turing, SkippyThePenguin, Ally, and others. In addition, there are quite a few names that I'm not as familiar with. It will be exciting to see how these individuals continue to perform over the next round.



While a lot of the games were still one-sided road wins, as expected, there are a lot more matches that came down to tie-breakers. Perhaps the most notable of these was the match between Simmon and Kakaburra. I would certainly recommend for those learning tak to delve into those games in greater depth. For today, though, I wanted to reflect and review a piece of my second game with SultanPepper:



In particular, I want to focus on the position after the fourth move:

[1. a5 a1 2.c3 Cb3 3. b2 c2 4. d2 c1]




While it's still relatively early in the game, with an experienced player like SultanPepper, strong moves even at this stage can lead to victory. In reaction to this position, I had considered a few options. In the game, I wound up choosing to play 5. d3, connecting my stones and trying to develop a threat. I, however, forgot to take into account the strong response that black played:  5. b3>!

[5. d3 b3>!]
With this one move, black cuts off the white stones, establishes a strong control at the center, and gains a strong tempo lead. In contrast, white's control is spread out and lacks tempo. Notice, though, how a lot of the control black gains is specifically as a result of the line black has developed up the c-file. Had white anticipated this strong move and played accordingly, white could have developed a stronger response.

The most direct way to have addressed this move would have been to play 5. c3-:

[5. c3- {b3-}]
As can be seen, this directly cuts black pieces, forcibly connects white's, and negates the effect of b3>. There are two major downsides to this play, though. First of all, as a capture, it's a bit slow for this stage in the game. While it's motivated, it also puts white in a bit of an odd position trying to defend prisoners from black's capstone. Black can respond with a pretty powerful threat of b3- at some point, threatening the c2 stack. (And in fact, it was this threat that led me to be wary of this move).

In the interest of improving my own play and understanding of the game, I decided to give this position to TakticianBot - currently, the best Tak ai. The response I got back was a play at 5. d4:


[5. d4]
This play is primarily what inspired me to share this analysis here on the blog. Having seen this play, I can come up with explanations for it and would consider playing this in a future game. But, it would have been unlikely that I would have ever come up with it before. What's particularly interesting about the move is that it approaches the situation indirectly. Black can still play b3>, which would still be quite strong. But now, white has a presence in the northeast corner, making it easier to defend the north or east territories with a play at c4 or d4 (highlighted). In fact, c4 would also block the immediate tempo of black. Even if black opts to play something else, the structure of white's stones is quite flexible, allowing white to respond to a variety of moves and to shift the tempo and focus accordingly.

Feel free to share your own thoughts, games, comments, and questions.
Best of luck to the participants in the next round of the tournament.