11th Chunlan Cup

Uberdude · **#21**

Like so? Black does feel a bit dumb capturing only those 2 stones as below now doesn't he?

Click Here To Show Diagram Code: [go]$$Wcm42 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . O . . . . . . . . X . . . . | $$ | . X . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . . . . . . . | $$ | . . O O . . . . . , . . 0 9 . , O . . | $$ | . . . . . . . . . . . . 7 8 O . . . . | $$ | . . . . . . . . . . . 4 O 6 O X X . . | $$ | . . . . . . . . . . 5 1 X X X O . . . | $$ | . . . . . . . . . . 3 2 . . . . O X . | $$ | . . . . . . . . . . . . . . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Click Here To Show Diagram Code: [go]$$Wcm52 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . O . . . . . . . a X . . . . | $$ | . X . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . 3 5 . . . . | $$ | . . O O . . . . . , . . X O 2 , O . . | $$ | . . . . . . . . . . . 1 O X O 4 . . . | $$ | . . . . . . . . . . . X O X O X X . . | $$ | . . . . . . . . . . O O X X X O . . . | $$ | . . . . . . . . . . O X . . . . O X . | $$ | . . . . . . . . . . . . . . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

But is this the best way for black, maybe 55 could push at 56? This direct cut looks better if white solidly connects as black can poke his head out and white can't run the 2 stones, so we probably expect the a-b exchange latter, or maybe black can get fancy and kikashi with c-d before b. This is definitely better for black than the last diagram, but is it better for black overall? Maybe? Would white e be next?

Click Here To Show Diagram Code: [go]$$Wcm50 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . O . . . . . . . . X . . . . | $$ | . X . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . c a . . | $$ | . X X O . . . . . . . . . . 4 . . d . | $$ | . . O O . . . . . , . . 3 1 2 , O . . | $$ | . . . . . . . . . . . . O X O b . . . | $$ | . . . . . . . . . . . X O X O X X . . | $$ | . . . . . . . . . . O O X X X O . . . | $$ | . . . . . . . . . . O X . . . . O X . | $$ | . . . . . . . . . . . e . . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Or would white extend like this? Black now could choose to play a rather than the other atari to make 52 a dumb empty triangle later and save the centre cutting stones. But what is the next move now? If push at b will white hane or extend? Capture or c seem soft.

Click Here To Show Diagram Code: [go]$$Wcm50 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . O . . . . . . . . X . . . . | $$ | . X . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . 3 b . . . . | $$ | . . O O . . . . . , . . . 1 2 c O . . | $$ | . . . . . . . . . . . a O X O . . . . | $$ | . . . . . . . . . . . X O X O X X . . | $$ | . . . . . . . . . . O O X X X O . . . | $$ | . . . . . . . . . . O X . . . . O X . | $$ | . . . . . . . . . . . . . . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Uberdude · **#22**

Maybe we should take this analysis to some AlphaGo not Chunlan Cup thread...

macelee · **#23**

I quote Uberdude's second diagram here:

Click Here To Show Diagram Code: [go]$$Wcm52 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . O . . . . . . . a X . . . . | $$ | . X . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . 3 5 . . . . | $$ | . . O O . . . . . , . . X O 2 , O . . | $$ | . . . . . . . . . . . 1 O X O 4 . . . | $$ | . . . . . . . . . . . X O X O X X . . | $$ | . . . . . . . . . . O O X X X O . . . | $$ | . . . . . . . . . . O X . . . . O X . | $$ | . . . . . . . . . . . . . . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

I actually think black is not bad here. Eating two stones does not look very large at the first sight. But it has several benefits that can be easily underestimated: (1) the squeeze is gone so black actually gets many more points. (2) white's wall is not perfect with aji at L9. (3) most importantly, the attach at 'a' becomes less effective for white (yes, believe me that is related as I will explain). (4) following (3) if white's 'a' is not very serious, black can first reduce the bottom left corner at C3 or C4 as he has sente.

Click Here To Show Diagram Code: [go]$$Wc $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . 8 . . . . | $$ | . . . . . . . . . . . . 7 2 3 4 . . . | $$ | . . O . . O . . . . . . 9 1 X 6 . . . | $$ | . X . O . . 0 . . , . . . . 5 , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . . . . . . . | $$ | . . O O . . . . . , . . . . . , O . . | $$ | . . . . . . . . . . . . . O O . . . . | $$ | . . . . . . . . . . . O O X O X X X . | $$ | . . . . . . . . . . . O X X X O . O O | $$ | . . . . . . . . . . X X O . . O O X X | $$ | . . . . . . . . . . . . X X X O X O . | $$ | . . . O . . . . . , . . . . O X X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

This is the real game: white got a very solid shape after :w9:

, black had to reduce at :b10:

which as the game turned out was not an easy fight. How many of you have wondered why :b6:

could not capture that stone instead to prevent white from sealing the top side? Here is why:

Click Here To Show Diagram Code: [go]$$Bcm6 = $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . 1 . . . . | $$ | . . . . . . . . . . . . . X @ X 6 8 9 | $$ | . . O . . O . . . . . . . O X 2 4 5 . | $$ | . X . O . . . . . , . . . . O 7 X X . | $$ | . . . . . . . . . . . . . . O X O 0 . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . . . . . . . | $$ | . . O O . . . . . , . . . . . , O . . | $$ | . . . . . . . . . . . . . O O . . . . | $$ | . . . . . . . . . . . O O X O X X X . | $$ | . . . . . . . . . . . O X X X O . O O | $$ | . . . . . . . . . . X X O . . O O X X | $$ | . . . . . . . . . . . . X X X O X O . | $$ | . . . O . . . . . , . . . . O X X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

If

captures, first of all black cannot fight a ko there because white has many ko threats near the squeeze (assuming AlphaGo knows how to fight ko properly

). So black has to connect at :ws:

. Locally black can defend the corner without problem. However, white can get :w15:

in sente. As a result, this would happen ......

Click Here To Show Diagram Code: [go]$$Wc $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . X . X . . | $$ | . . . . . . . . . . . . . X X X O O X | $$ | . . O . . O . . . . . . . O X O O X . | $$ | . X . O . . . . . , . . . . O X X X . | $$ | . . . . . . . . . . . . . . O X O O . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . X 1 | $$ | . X O X . . . . . . . . . . . . . X O | $$ | . X X O . . . . . . . . . . . . . X O | $$ | . . O O . . . . . , . . . . . , O X O | $$ | . . . . . . . . . . . . . O O . . X O | $$ | . . . . . . . . . . . O O X O X X X O | $$ | . . . . . . . . . . . O X X X O . O O | $$ | . . . . . . . . . . X X O . . O O X . | $$ | . . . . . . . . . . . . X X X O X O O | $$ | . . . O . . . . . , . . . . O X X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

I will leave you guys to decode the sequence, but basically with the :w15:

sente in the previous diagram, white can connect from the first line!

To summarise, in Uberdude's diagram, it may look like that capturing two stones are not quite big, but in fact it is much bigger because white's top side indirectly affected and reduced in value.

So back to the Chunlan Cup game, we can see Mi Yuting's S10/R11 exchange actually makes sense!

Uberdude · **#24**

Nice sequence macelee, Go Seigen would be proud of that 1st line crawling! I wonder if any such sequences appeared in AlphaGo's playouts and that is why it didn't make the s10 t9 exchange. I don't think there was any time Lee could have exchanged s10 for r11 later in the game was there (too small), so it seems Mi has found the correct timing. Is it overly harsh on Lee to say that he should have found s10 in the game? Hindsight is nice, the games had a lot of pressure and 2 hours is not much time to think, but for a top top pro such as himself they should pay attention to the tiniest details and efficiencies.

macelee · **#25**

I am sure the first-line move was discovered by a Chinese pro during one of their regular group studies. I read it from some discussions on the Internet. Given the hint, even amateurs at our level can read through. But getting that strategy in the first place is much more difficult. I think it is highly unlikely that an AI program can 'plan' for something like that.

A pro player's thinking process would be something like: I read this variation; it doesn't seem to work; but if there's an extra stone there my sequence would work; let me find out a way to place that extra stone without my opponent noticing...... If a machine can do this sort of trick, our humanbeing would be wiped out soon.

Uberdude · **#26**

macelee wrote:

A pro player's thinking process would be something like: I read this variation; it doesn't seem to work; but if there's an extra stone there my sequence would work; let me find out a way to place that extra stone without my opponent noticing...... If a machine can do this sort of trick, our humanbeing would be wiped out soon.

Reminds me of Ma Xiaochun's 36 Stratagems book, all sorts of brilliant sneaky plans like that.

Uberdude · **#27**

Uberdude wrote:

Something I was thinking about from the 5th AlphaGo game was why Lee Sedol pushed with :b43:

as below. By loosing that liberty the n6 cut etc became more potent even though in the end white only got a squeeze (though actually given Mi's s10 attach I wonder if that squeeze really was a misread from AlphaGo or a sente sequence to stop Lee successfully playing that attach at a time when white would need to answer submissively as Park did, i.e. did AlphaGo see a line for black in which he played s10 and that reduced AlphaGo's winning chances so it played out the squeeze so Lee wouldn't have the chance to s10?).

Click Here To Show Diagram Code: [go]$$Wcm42 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . O . . . . . . . . X . . . . | $$ | . X . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . . . . . . . | $$ | . . O O . . . . . , . . . . . , O . . | $$ | . . . . . . . . . . . . . 3 O . . . . | $$ | . . . . . . . . . . . 5 O 2 O X X . . | $$ | . . . . . . . . . . . 1 X X X O . . . | $$ | . . . . . . . . . . 6 4 . . . . O X . | $$ | . . . . . . . . . . . . . . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

If black directly haned and this happened that looks like an unconditional improvement to me (actually the improvement might not be so much as white could still probe with the n6 cut or s7 hane and if black plays in a way that makes the o8 liberty sente then white fixes that cut in sente which makes a big difference to the strength of the centre moyo, and if plays to prevent o8 being sente white gets the same squeeze...).

Click Here To Show Diagram Code: [go]$$Wcm42 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . O . . . . . . . . X . . . . | $$ | . X . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . . . . . . . | $$ | . . O O . . . . . , . . . . . , O . . | $$ | . . . . . . . . . . . . . . O . . . . | $$ | . . . . . . . . . . . 3 O . O X X . . | $$ | . . . . . . . . . . . 1 X X X O . . . | $$ | . . . . . . . . . . 4 2 . . . . O X . | $$ | . . . . . . . . . . . . . . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

So did Lee fear white would trade with something like this? White abandons the top right moyo area (which makes sense as he hasn't yet invested building it with the a attach and subsequent gote building of thickness).

Click Here To Show Diagram Code: [go]$$Wcm42 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . O . . . . . . . a X . . . . | $$ | . X . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . . . . . . . | $$ | . . O O . . . . . , . . . . . , O . . | $$ | . . . . . . . . . . . . . . O . . . . | $$ | . . . . . . . . . . . 4 O 6 O X X . . | $$ | . . . . . . . . . . 5 1 X X X O . . . | $$ | . . . . . . . . . . 3 2 . . . . O X . | $$ | . . . . . . . . . . . 7 . . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Fan Hui gave a talk about AlphaGo vs Lee Sedol game 5 at the US Go Congress (https://www.youtube.com/watch?v=cmK2HwKMrco), which answered some questions I posed in the above. Lee's push rather than direct hane does indeed seem to be a mistake, and his fear of the double hane may well explain it. But AlphaGo would have simply extended with 3 as below if he played hane, not double hane! (possible continuation a-f shown too)

Click Here To Show Diagram Code: [go]$$Wcm42 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . O . . . . . . . . X . . . . | $$ | . X . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . . . . . . . | $$ | . . O O . . . . . , . . . . . , O . . | $$ | . . . . . . . . . . . . . . O . . . . | $$ | . . . . . . . . . . . . O d O X X . . | $$ | . . . . . . . . . c 3 1 X X X O . . . | $$ | . . . . . . . . f b a 2 . . . . O X . | $$ | . . . . . . . . . . . . e . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Fan also showed the 1st line connection from the lower right dead stones to the upper right corner aji depending how black answers the crosscut there, though didn't say if AlphaGo saw those lines or just post-game professional analysis. No mention of Mi Yuting's s10 attachment though. Let's see what is in their release of commented games in a few weeks...

ez4u · **#28**

Uberdude wrote:

Something I was thinking about from the 5th AlphaGo game was why Lee Sedol pushed with :b43:

as below. By loosing that liberty the n6 cut etc became more potent even though in the end white only got a squeeze (though actually given Mi's s10 attach I wonder if that squeeze really was a misread from AlphaGo or a sente sequence to stop Lee successfully playing that attach at a time when white would need to answer submissively as Park did, i.e. did AlphaGo see a line for black in which he played s10 and that reduced AlphaGo's winning chances so it played out the squeeze so Lee wouldn't have the chance to s10?).

Click Here To Show Diagram Code: [go]$$Wcm42 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . O . . . . . . . . X . . . . | $$ | . X . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . . . . . . . | $$ | . . O O . . . . . , . . . . . , O . . | $$ | . . . . . . . . . . . . . 3 O . . . . | $$ | . . . . . . . . . . . 5 O 2 O X X . . | $$ | . . . . . . . . . . . 1 X X X O . . . | $$ | . . . . . . . . . . 6 4 . . . . O X . | $$ | . . . . . . . . . . . . . . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

If black directly haned and this happened that looks like an unconditional improvement to me (actually the improvement might not be so much as white could still probe with the n6 cut or s7 hane and if black plays in a way that makes the o8 liberty sente then white fixes that cut in sente which makes a big difference to the strength of the centre moyo, and if plays to prevent o8 being sente white gets the same squeeze...).

Click Here To Show Diagram Code: [go]$$Wcm42 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . O . . . . . . . . X . . . . | $$ | . X . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . . . . . . . | $$ | . . O O . . . . . , . . . . . , O . . | $$ | . . . . . . . . . . . . . . O . . . . | $$ | . . . . . . . . . . . 3 O . O X X . . | $$ | . . . . . . . . . . . 1 X X X O . . . | $$ | . . . . . . . . . . 4 2 . . . . O X . | $$ | . . . . . . . . . . . . . . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

So did Lee fear white would trade with something like this? White abandons the top right moyo area (which makes sense as he hasn't yet invested building it with the a attach and subsequent gote building of thickness).

Click Here To Show Diagram Code: [go]$$Wcm42 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . O . . . . . . . a X . . . . | $$ | . X . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . . . . . . . | $$ | . . O O . . . . . , . . . . . , O . . | $$ | . . . . . . . . . . . . . . O . . . . | $$ | . . . . . . . . . . . 4 O 6 O X X . . | $$ | . . . . . . . . . . 5 1 X X X O . . . | $$ | . . . . . . . . . . 3 2 . . . . O X . | $$ | . . . . . . . . . . . 7 . . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Fan Hui gave a talk about AlphaGo vs Lee Sedol game 5 at the US Go Congress (https://www.youtube.com/watch?v=cmK2HwKMrco), which answered some questions I posed in the above. Lee's push rather than direct hane does indeed seem to be a mistake, and his fear of the double hane may well explain it. But AlphaGo would have simply extended with 3 as below if he played hane, not double hane! (possible continuation a-f shown too)

Click Here To Show Diagram Code: [go]$$Wcm42 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . O . . . . . . . . X . . . . | $$ | . X . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . . . . . . . | $$ | . . O O . . . . . , . . . . . , O . . | $$ | . . . . . . . . . . . . . . O . . . . | $$ | . . . . . . . . . . . . O d O X X . . | $$ | . . . . . . . . . c 3 1 X X X O . . . | $$ | . . . . . . . . f b a 2 . . . . O X . | $$ | . . . . . . . . . . . . e . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Fan also showed the 1st line connection from the lower right dead stones to the upper right corner aji depending how black answers the crosscut there, though didn't say if AlphaGo saw those lines or just post-game professional analysis. No mention of Mi Yuting's s10 attachment though. Let's see what is in their release of commented games in a few weeks...

What was the explanation for rejecting the double hane?

Uberdude · **#29**

ez4u wrote:

What was the explanation for rejecting the double hane?

Not much directly, other than extend was good enough and result to f was good for white too with a lone black stone in a sea of white centre.

trout · **#30**

Reminder !!!!!!!!!

round of 8;

Gu Zihao vs Kim Jiseok
Park Yeonghun vs Lian Xiao
Mi Yuting vs Ke Jie
Tan Xiao vs Tuo Jiaxi

round of 8 - 12/20/16
round of 4 - 12/22/16

trout · **#31**

round of 8;

Ke Jie defeated Mi Yuting by resign.
Park Yeonghun defeated Lian Xiao by resign.
Gu Zihao defeated Kim Jiseok by 0.5.
Tan Xiao defeated Tuo Jiaxi by resign.

round of 4 - 12/22/16

pookpooi · **#32**

Next round Ke Jie vs. Park Yeonghun
Tan Xiao vs. Gu Zihao

Uberdude · **#33**

Oh, did they change the pairings? Mr Kin's tournament table says it would be Ke vs Gu. http://igokisen.web.fc2.com/wr/cl.html. Btw Ke vs Mi was another example of the "wrong" close approach to Kobayashi that's popular these days.

pookpooi · **#34**

Uberdude wrote:

Oh, did they change the pairings? Mr Kin's tournament table says it would be Ke vs Gu. http://igokisen.web.fc2.com/wr/cl.html. Btw Ke vs Mi was another example of the "wrong" close approach to Kobayashi that's popular these days.

My source is http://live.sina.com.cn/zt/l/v/sports/chunlan8/

hyperpape · **#35**

I'm not sure, but I think I remember that sometimes when there is a draw that happens later, Mr Kin will print a table, and then shuffle it when the later rounds are revealed. It's one of the limitations of presenting a tournament table.

macelee · **#36**

hyperpape wrote:

I'm not sure, but I think I remember that sometimes when there is a draw that happens later, Mr Kin will print a table, and then shuffle it when the later rounds are revealed. It's one of the limitations of presenting a tournament table.

That is correct. Mr Kin builds those tables from the beginning of the tournaments. They often have to be adjusted at later time.

trout · **#37**

Semi Final;

Park Yeonghun defeated Ke Jie by resign.
Tan Xiao defeated Gu Zihao by resign.

Final,

Park Younghun vs Tan Xiao

TheCannyOnion · **#38**

trout wrote:

Semi Final;

Park Yeonghun defeated Ke Jie by resign.
Tan Xiao defeated Gu Zihao by resign.

Final,

Park Younghun vs Tan Xiao

Good to see Park doing so well. He's one of my favorite players. I really enjoy his calm and balanced style.

Lucian · **#39**

TheCannyOnion wrote:

Good to see Park doing so well. He's one of my favorite players. I really enjoy his calm and balanced style.

Indeed, a characteristic game for Park when he is in good shape. Not sure what was Ke Jie mistake in fuseki, but thereafter Park play looked so smooth and effortless.

I am happy to see the nimble Tan Xiao in his first world final. I checked on go4go but couldn't find a head to head Park - Tan, did they ever play together?

Uberdude · **#40**

I too like Park Yeonghun's calm territorial style, though I found it interesting that in his game with Ke Jie he did that fancy attach on the top right corner to make shape in sente and swallow up the invader, whereas Lian Xiao just did the simple attach on top of the invader (and all those stones subsequently died).

This is a great chance for Park to win his first international title in almost a decade: he won the Fujitsu aged 19 in 2004 and again in 2007. He had a bit of a resurgence last year (https://www.goratings.org/players/125.html) but lost in the final of the LG cup to Kang Dongyun (beat Park Junghwan and Tuo Jiaxi on the way).

11th Chunlan Cup

Who is online