: ( 503)
KataGo's atari can be played successfully only now – at just one single moment in time – typical for the genius of Inoue Dōsetsu Inseki.

Please refer to "KataGo's Unbelievable White Atari on the Left Side" ( 1093a) to see why this atari does not work when played earlier or later.

Please note that we inserted the Hanezeki Exchange - (played before the Crosscut Sequence), as found in 2017 to be the correct timing for these moves.

: (A 548ka; 97 548kb)
It is best for Black to force White to connect at , immediately giving atari at his nakade stones.

Black must not answer White's atari by capturing at .

Black might want to immediately complete the nakade by playing at , in order to keep his nakade stones at two liberties, but this move will result in a change in the order of moves only.

:
Immediately after completed the nakade, White captures Black's single stone.

: (A 548kc; B 548kd)
Black cannot afford to play the large endgame move at in the top left corner, because this would put him behind by one liberty in the Main Semeai, resulting in a lost Punishment Semeai.

Preceding the move at by the atari of will not help either, as White has a surprising local response at her disposal.

:
After the tsuke of Yamada Shinji, KataGo preferred White's crosscut of to drawing back with a move at .
Despite this crosscut might be a standard endgame technique, it has been overlooked by amateurs and professionals alike.

: (106 548ke)
Black's atari in the corner starts KataGo's preferred continuation.

Black might want to resist more forcefully, by giving atari with instead, but will suffer by one point in the final score.

Please note that Black might not be able to act as forcefully as he wants to do in the top left corner. He has to play attention to keep his ability of occupying White's marked liberties without any delay; otherwise White will choose, and win, the Punishment Semeai.

: (A 548kf)
Black might want to capture at , instead, but will not benefit from this variation either.

: (B 548kg)
For a long time, we considered blocking at the top () and blocking at the bottom () to be miai.
However, we now think that blocking at the bottom is one point worse than blocking at the top.

:
As usual, Black ignores White's atari of at the bottom to gain points at the top.

: (123 548kh)
This is KataGo's surprising bad-shape move in the top left centre.

White might choose the "usual" continuation at to reach the same final result of the game – in the Capture Variation – but she will suffer by one point in the Semeai Variation.

:
Black forces White to connect at , so losing two Main Semeai Liberties.

: (A 548kj; B 548kk)
Black plays tenuki to capture White's single stone at the right.

Black can also immediately block with to reach the same final scores in both Main Variations.

KataGo preferred the solid connection at , but which is one point worse (in the Capture Variation), according to our estimations.
Please note that this tiny difference between "W + 2" and "W + 1" (under territory scoring) is invisible for KataGo, due to the usage of area scoring.

: (A 548kl)
After Black played elsewhere, this wedge is best for White.

White might also push at to reach the same result of the game (in the Capture Variation), but she will suffer by one point in the Semeai Variation.

( 132 548klk; A 548klp)
White attaches at the centre of three stones, threatening to cut below.

Dany found that playing the atekomi of , instead, will benefit White by one point in both Main Variations.
The development that follows includes a ko fight in the left centre (= "Dany's Ko"), which Black will be unable to win.

If White ever wanted to play the central attachment, she would have to precede it with her kikashi of , , in the lower left corner. However, Black would be able to win the Semeai Variation in this case.
Please note that threatening with without having played the exchange of , beforehand would be a grave mistake by White, resulting in a drawn game.

: (134 548km)
Black occupies one of White's Main Semeai Liberties in the lower left corner.

Black can also capture at , instead, to reach the same final scores in both Main Variations. These two moves are miai.

: (137 548kn)
White solidly connects her single stone at the upper edge.

She can also play at in the centre, instead, to the same effect.

: (A 548ko)
Black plays from above.
This is a suggestion of Joachim, and it is a kind of move that has a similar effect like Michael Redmond's kikashi in the bamboo-joint. It's just played on the other side.
Black gains one additional liberty at .

KataGo's preferred line of play uses the move at from below, favouring White by one point in the Semeai Variation (under territory scoring).
Please note that KataGo is unable to distinguish the difference between both moves; their final results of the game are identical under area scoring that is used by AI.

(B 548kp)
White occupies a liberty of Black's large group.

Connecting at , instead, forcing Black to capture with , will not affect the final score of the game, but worsen the Semeai Variation for White.

:
The Main Semeai continues.

:
Due to his just won liberty, Black is able to destroy two points of White's territory (in the Capture Variation) by connecting at – also losing two Punishment Semeai Liberties – instead of only one, e.g. with the forcing exchange of , .

...

...

: (A 548kq)
White captures the hanezeki's tail, entering the Capture Variation.

Despite the fact that White – compared to Our Solution (2017) – added four stones to her large group at the left side (), but Black only two to his large group at the right (), what seemed to confine White to the Capture Variation, she might also choose the Semeai Variation to get the same final result as here with the Capture Variation.

...

...

...

...

This is the final position of the Capture Variation.

Capture Variation:

Compared to the superordinate variation, ...

• Black lost two points of territory in the upper left corner.
• White gained two points of territory at the upper edge.
• White lost one point of territory at the top.
• Black lost three points of territory in the left centre.
• Black captured White's wedging stone, for a gain of one point.
• Black lost three points of territory in the lower left, and did not capture one of White's stones, for a cumulated loss of five points.
• White captured one Black stone in the lower left.
• White did not get one point of territory in the upper right centre.
• Black captured one White stone for a gain of two points there.
• In the lower right, White did not capture one of Black's stones, for a loss of two points.

In total, White gained six points.

Black + 5 => White + 1

White wins the Capture Variation by one point!