Uberdude wrote:Good question Leon, and nice spot of g1 (which I didn't see during the game). Here's my idea:
- the ko is not just about life/death of black/white in the corner, but also about the eyes of the outside black group. Working backwards:
- white could have simply answered black 145 at c8 and black's corner would be dead rather than the approach ko it became in the game (for which white got a few moves to build his centre), but then black would get d6 in sente which builds some handy eyes in e6 area, and then presumably play something in the centre.
- Given that white didn't answer, and if we make the assumption this was a reasonable decision of Ke Jie and not just misguided human fighting spirit (which is not as simple as does bot agree: a bot may be confident it can play out the half point win with c8 answer whereas Ke might see that as too uncertain and actually prefer a more complicated game where he can demonstrate his superior skill), then that explains Kim's earlier decision to give up on the ko (for now) and take the q12 profit and attempt to force from the outside with d7.
- so if white plays g1 then he loses 1 point as eventually he'll capture the stones anyway when black fills dame, black still gets the d6 forcing move for eyes, and because of the double ko infinite threats black can win any other ko such as answering e9 at e8 which means that group is practically alive in sente and he can then turn to the centre. So white played the atari because he wanted to make black's outside group weaker and then if he wins the ko it's in big trouble without d6 sente (even more so after white f8)
- so Ke makes the ko bigger, thus Kim feels he can't answer the q10 threat (which makes Ke's previous centre moves during ko better as now black's group is weak so he can't be aggressive in the centre), and also creates weakness in g8 connection which makes the cut for desserts we saw later in game.
Hi, thanks for the answer, I think I can understand this a bit better now. Ko fights are quite difficult, aren't they?
The game, Park Junghwan was black. One interesting thing from a quick LZ review was how often it thought black should defend at g16 throught the middlegame. White g17 wouldn't cut with black g18 g16 combo but makes an option to scoop out the territory and eyes, and LZ sees g17 g18 and tenuki as a beneficial exchange.
Upset in game 2 Park Junghwan (white) made a massive overplay/blunder at 96 and Park Younghoon is now at 99% or ~28 points lead (says Yike bot)! At least white profit from the trade at top right is more certain (assuming L9 outside cutting stones don't die, how much of the AI judgement is based on black using that aji to stop white making much on lower side?) whilst the left side white can still try to make a mess and hope for black mistake.
Update: after a long think Park Y played g10 extension instead of saving the top left stones, claiming it is sente for L8 and resurrecting the dead group. If that's true it captures the 5 cetnre stones in sente for good aji, but Park J resisted and Younghoon played the wrong net, looks like he's trying his best to lose! What a reversal this will be if Junghwan pulls it off.
12 chunlan final 2 winrate.PNG (171 KiB) Viewed 10910 times
