Alex Weldon: While looking for real-world examples to use in the book I'm planning (Breaking Bad Habits . . . a kind of dictionary of common mistakes, and how to avoid making them), I came across a game between an 8k and a 9k on KGS. A simple life-and-death situation came up. This is a simplified, but functionally identical position:
This should be a piece of cake for an 8k. a in Diagram 1 is a perfectly straightforward way to get a ko, which I believe to be the best W can expect. b can also get a ko, but a less favourable one. However, that's not what W played. Diagram 2 shows the actual game.
is a tesuji in case the two marked liberties are filled: then W can play at a and B can't cut, due to shortage of liberties. But in a case like this, where there are outside liberties, it's not the kind of move you would come up with by reading. I think it's a pretty good assumption that W has seen many tsumego involving that tesuji and played it instinctively, without reading.
Just as studying joseki can be harmful, as people will tend to play the patterns without thinking about the whole board, perhaps tsumego can be damaging if overdone... players attempting to substitute recognition of common patterns for actual reading.
Thoughts?
dnerra: Opinions may vary on this -- my experience tells rather the opposite: These players have not done enough Tsume-Go! In a recent interview with the DGoZ, the korean baduk professor (and strong 7d) Lee Ki Bong recommended spending most of the time devoted to go study on life-and-death problems. Apparently, it is also what the korean inseis do most of their time.
Bill: Yes. First, good luck on your book! It sounds interesting. :-)
Second, I'm not sure that White would have been better off not to have learned some patterns. (Besides, he didn't exactly learn this pattern, did he? ;-))
Third, I think it's a hard problem.
E. g.,
Will an 8-kyu realize that allows Black to live?
(Will he even think of trying -
in the first place? ;-))
Robert Pauli: at
. . . and same ko as below.
I'm not so good at tsumego, myself, but I think this is an advanced problem.
Calvin: We should probably move solution attempts to a discussion page, but maybe white can just kill with , though I'm not completely sure. Still, I think your point is correct. If something looks like it can be killed, there is a temptation to try even if you can't read it out. Tesuji-like moves can be very bad if they don't work---worse than passing. Another point is that even if this or the other moves work, in this example, white has 11 stones in the area to black's six, and it's white's move so black presumably got five moves somewhere the rest of the board plus sente if white finds a killing move, so black might just say: "Congratuations white, you got it."
jfc: In response to , how about
? There are plenty of other options here...
Bill: Thanks, Calvin and Jonathan! I think that does it. :-)
Dave: WOW! The double ko death turns this into a great problem. I hope there isn't an easier way that would preempt this solution :-)
Alex Weldon: Wow. I clearly didn't read it carefully enough. Anyway, yeah. Fantastic problem, if that's really the best solution. Odd to stumble across it this way. We should definitely move it somewhere else in the library, as a problem, and meanwhile I'll look for a better example of misapplied tesuji resulting from applying tsumego patterns without reading. :-)
Bill: I have created Tsumego From Games 42 for this problem, and copied the attempts here to the Tsumego From Games 42 /Attempts page, with minor edits. I'm leaving the original here, pending approval of the move and edits.
unkx80: Hey ya... I did not get this solution myself too. Nice. =)
In an interview in DGoZ 4/04, Lee Ki-Bong (7d = 1-2p, teaching Baduk at the Myojing University in Seoul) surprisingly stressed following: