Lose a ko to win the game
One technique to seal a game when you are in a winning position, is to create a ko in the opponent's large prospective territory, preparing to lose it and play elsewhere twice.[1] The prospective territory becomes effective territory, while you choose an area to play twice. This simplifies the game. Given your being ahead before the exchange, the simplification is very likely to result in a clear advantage. The idea of making a ko when ahead, for simplification, is perhaps not intuitive, since a ko feels inherently complex to the amateur, but we can see it a lot in professional play, in particular in the games of Go Seigen and Gu Li.
(examples to be added)
See also:
hhw: Would it be better to call this page "Force your opponent to start a ko that you can lose"? Since the idea is to start a ko when you're ahead. Then, as normal, you would win the ko if you gained more than your opponent's threat. But in this situation you would also be happy to make a threat smaller than the ko (or just play a big endgame move) if your lead is kept after the territory solidification that results from your opponent winning the ko and you playing twice.
unkx80: The following game between Takagi Shoichi (Black) and Lin Haifeng (White) may be an example. There is no commentary regarding the ko.
unkx80: cannot play at
, or it will result in White a, Black b, White
, resulting in a capturing race that Black cannot win.
unkx80: I guess that the aim of starting this ko is to get both and
for compensation.
unkx80: After the White cut, Black's framework at the bottom vanished in no time. is needed to settle the large dragon.
Dieter: As a minor remark, I believe responding to is much better than
.
Dieter: In order to understand the strategic concept, we must evaluate what happens if White does not use the ko. Suppose White kills locally, then leaves bad aji, because a now is sente (small reading exercise). Or will
be played at b?
So I think part of the purpose of the ko was to eliminate the bad aji at the top, by forcing Black to play the threat at (two diagrams ago).
unkx80: I don't claim to fully understand this game, but I suspect will be played at b. Then when White cuts at
, Black may give up the two stones.
Dieter: I understand the reasoning, but doesn't a clean capture of the upper left, as would happen if Black gives up the 2 stones, simplify the game even more?
Dieter: Here is a game I recently played, in which I was reminded of this strategem. It is probably not a good example, because I was already largely ahead, but on the other hand it achieves the goal of removing aji and pushing the game towards a more clearly defined win. Moreover, I consciously applied the idea, being one of my favourties:
I could play the monkey jump with a but I saw the opportunity to start a ko I could afford to lose.
My first threat is an obvious one, likewise Black threatens to revive his group with
. With
I play a minor threat, which however is worth about 10 points and gives two eyes to my group, so that I don't have to worry anymore about the aji around a. I suspect Black did not expect the snap-back at
, which makes the move 2 points bigger and sente.
(Incidentally, this is part of ko psychology. Play a gigantic threat first. Then play a threat which is much less severe, but probably just enough for you. The opponent may be lured in to relief that it's only going to be as bad as this. Then you get what you want. It is similar to negotiation tactics, where you irst make an outrageous demand, to then settle for what you really want. Especially when you are not ko master such devious tactics may be helpful.)
Black has made a few points in the corner, but White took a chunk out of the big territory, making his group 100% alive in the process. At the low dan level, this completely settles the game.
tapir: I appreciate the strategy, but I wonder whether White gained anything here and whether simple endgame moves would not do the trick. Because: there is no meaningful aji at a as is dead, White can simply draw back to connect.
Uberdude: This game of mine is pretty much a textbook example of this strategy. The ko at b12 on move 84 is incredibly heavy for black and light for white and enables me to get some large profit as the game moves from the middlegame to the oyose.
http://www.online-go.com/games/board.php?boardID=310287
[1] This can be correct strategy, even when you do not have a commanding lead.
Hyperpape: The interesting part is that in this case 1) the ko simplifies the game, which is usually not desirable when you are behind, and 2) it is no longer essential that the ko gains points, so long as you are still clearly ahead after it is played. Right?
Dieter: Right! In the example minue showed me, White was ahead. Black had some prospective territory, say, arising from a 4-4 corner enclosure. White invaded at 3-3, lost the ko and played two successive moves in the centre to harnass the advantage and become very thick. The ko exchange wasn't all that spectacular: we often think of such exchanges as swapping one territory for another. In this case, White's prospective advantage became definite by becoming thick, thanks to the successive moves.