Oiotoshi / Discussion

Path: <= Mistake =>
Sub-page of Oiotoshi

unkx80: I moved this lengthy discussion from the parent page. However, the contents are useful as an example of how not to set up an oiotoshi, probably on a new page.

Example 4

(modified from The Second Book of Go)

Oiotoshi

starts the oiotoshi. If White tries to run away at , is atari again, and this time running away at doesn't help White at all since now the whole group is captured.

So White shouldn't have run away at , but instead connected with .

Oiotoshi tesuji

Using the previous example as reference, it is easy to see that if White saves the stones here, Black can capture , blocking off White's progress along the top side. can also be called oiotoshi. (I'm not sure whether this example is relevant enough (and correct)! Comments, please. -- Jan)

BillSpight: is premature. It requires the atari play below.

Oiotoshi tesuji (not premature)

I don't think this is premature. Doesn't this work? (Thomas Nordhaus)

Oi-Otoshi dissent

Bill is right, in the starting position the tesuji is premature and brings almost nothing. (Capturing one stone will be a bigger sente than if hadn't been played, but on the other hand, one ko threat has been lost.)

White should of course connect at a instead of and Black can capture only one stone which can be captured without sacrificing at anyway.

White connects

Bill:

Normal play

After , is White's correct response. If White does not play there, Black can throw in at .

After , and are miai. Black can count 1/3 point in the ko (1/3 of ).

Black error

fills at .

In this case Black has lost 1 point for the throw-in stone at , and does not have 1/3 point in ko. gives away 1 1/3 points.

Cruncher: Well, after playing , the position is identical to Jan's original "Second Oiotoshi" diagram, and the idea of that example was to show that playing in the "Second Oiotoshi" diagram is the wrong move.

...so does that mean that all of you are right? ;-)

Bill: It looks to me that in that diagram is in this one. So I'm not sure what you are referring to.

Cruncher: I actually meant "...is identical to the starting situation in Jan's diagram...", that is, before is played there. But I guess the keyword here that I missed is tesuji: playing the premature doesn't really change anything and just recreates an equivalent of the original example - where white can still escape from.

Bill: The throw-in of is worse than zokusuji. Compare the results with and without it.

Yet, this "worse than zokusuji" is absolute sente through to its conclusion. White can only either completely obey black or accept at least the loss of all 4 stones.

dnerra: This claim misses that as suggested by black also threatens to capture 4 stones. I agree with all of Bill's comments.

Bill: We don't know the rest of the board, so we cannot say that the play is sente. In the position as shown, I would think that , next, or - where is the throw-in, would both be gote, since securing Black's stones would be quite large by comparison to taking or saving a few stones.
White keeps his options open.
Black takes in gote

Bill: Taking one stone with , above, would be bad play by Black until late in the game.
Miai

In fact, and are miai. Either Black plays as a prelude to taking the ko with , or the play goes - , .

White later improves the situation while threatening to save all of his stones without ko.
White threatens to start a ko that would save all his stones and break into black's territory

Nobody said this was a koyose sequence, Bill. Both moves have merit, and the rest of the board matters. Closing the side is huge, and finishing the job cleanly might be bigger than 1 1/3 points.

Bob Myers: Umm, this is no ko; Black simply ataris to the right of 1.

Bill: Well, 24.79.78.96, I do not believe that there is a case where a Black throw in is better than the simple atari, except perhaps as a losing ko threat, which is very unlikely. In all normal cases the simple atari is as good or better.

See Throw in or Not.

Path: <= Mistake =>
Oiotoshi / Discussion last edited by BillSpight on May 30, 2004 - 17:06