It is currently Thu Jun 20, 2019 3:55 am

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 64 posts ]  Go to page Previous  1, 2, 3, 4  Next
Author Message
Online
 Post subject: Re: Let's study AlphaGo's opening book
Post #41 Posted: Thu Feb 01, 2018 2:20 pm 
Judan

Posts: 5882
Location: Cambridge, UK
Liked others: 335
Was liked: 3141
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Over on another thread when talking about the magic sword an ponnuki joseki (or not) I posted about this AlphaGo opening book position:

Uberdude wrote:
[About what next at upper left, with lower right as a 3-4] Black wouldn't play a local move, q9 high Chinese looks nice to make a moyo. Locally black e18 is a fairly large move that stops white sliding so makes sense when you are trying to make the top side territory, but white's unlikely to answer it until endgame. Maybe this is why AlphaGo thinks this slide of 16 is a good move for white, I was surprised as it seems too small to play now but maybe it thinks black e18 is a good move if it tenukis given 15 is starting to develop the top side (but it's still very open!). That we only see the win% for a single move is rather frustrating in positions like this, does AG actually think 16 is significantly the best move on the board, or is it about the same or even worse than a/b/c?

Click Here To Show Diagram Code
[go]$$Bcm15
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . 2 . . . . . . . . . . . . . |
$$ | . . . O . . . . . . . . . 1 . . . . . |
$$ | . . . O X . X . . , . . . . . X . . . |
$$ | . . O X . X . . . . . . . . . . . . . |
$$ | . . O . X . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , b . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . a . . c . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


Well, I sort of found an answer that AG does indeed think e18 is a move worth playing very soon if white (or black below) doesn't slide there. Take a look at this position (rotated from the tool to match above). Note that at move 17 black is already down at 33%, most of that loss comes from the usual human joseki move at a cutting the big knight which AG really hates, and following moves are AG against itself fluctuating around 33-35%. White encloses with 18, black makes a shimari for his mini-Chinese, clearly a big move. White the immediately plays the e18 move, so is evidently saying this and o17 were kind of miai to make a nice top side and happy to get both (but black did get a shimari in exchange which is clearly huge too). Black does then reduce with a shoulder hit as I thought might happen, and then b/c next.

Click Here To Show Diagram Code
[go]$$Bcm17
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . 4 . . . . . . . . . . . . . . |
$$ | . . . X . . . . . . . . 6 2 . . . . . |
$$ | . . . X O . O . . , . . 5 . . O . . . |
$$ | . . X O a O . . . . . . . . . . . . . |
$$ | . . 1 . O . . . . . . . . . . . c . . |
$$ | . . . . . . . . . . . b . . . . . . . |
$$ | . . . X . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . O . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X , 3 . . . . , . . . . . O . . . |
$$ | . . . . . . . . X . . . . X . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


This post by Uberdude was liked by: Bill Spight
Top
 Profile  
 
Online
 Post subject: Re: Let's study AlphaGo's opening book
Post #42 Posted: Fri Feb 02, 2018 7:33 am 
Judan

Posts: 5882
Location: Cambridge, UK
Liked others: 335
Was liked: 3141
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Here's an interesting thing I found in whether AG thinks it's good to play the avalanche joseki or not. In general AG doesn't like the avalanche much and prefers to hane at a, as here, but interestingly how much worse the avalanche (b) is changes a lot based on the location of black's lower right corner stone. If it's at 4-4 then avalanche is only 1% worse, but with the marked 3-4 it's 5.5% worse. Looking at pro games, in both positions hane is most common (and has a slightly better win ratio) but avalanche is actually a relatively more common 2nd choice with the 3-4 where AG thinks it's worse.

Click Here To Show Diagram Code
[go]$$B
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . . . . . . . . . . . |
$$ | . . . b . . . . . , . . . . . , X . . |
$$ | . . O X . . . . . . . . . . . . . . . |
$$ | . . a . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . A . . . |
$$ | . . . . . . . . . . . . . . . A . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


Why the big difference? The answer seems to be the effect it has on a ladder in the small avalanche joseki: with the 4-4 this ladder doesn't work for white, but with 3-4 it does. So with 3-4 AG thinks white should play the small avalanche and it is significantly better than other choices (black win only 38%, large avalanche is 45%, descend is 42%), this is the punishment for black's bad choice to avalanche. Expected continuation is below, black has to make hanging connection at 6 rather than extend and usual continuations continue with black's win around 38%.

Click Here To Show Diagram Code
[go]$$Wm10
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 3 5 . . . . . . . . . . . . . . |
$$ | . 4 2 O X . . . . . . . . . . . . . . |
$$ | . . O X . 6 . . . , . . . . . , X . . |
$$ | . . O X . . . . . . . . . . . . . . . |
$$ | . . . 1 . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


My understanding is black would like to extend for 15, but this doesn't work due to the following ladder which just misses the lower right 3-4.
Click Here To Show Diagram Code
[go]$$Bm15
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O O 2 . . . . . . . . . . . . . |
$$ | . X X O X 1 6 . . . . . . . . . . . . |
$$ | . . O X 5 7 9 0 . , . . . . . , X . . |
$$ | . . O X 4 8 . . . . . . . . . . . . . |
$$ | . . 3 O . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


However, with the 4-4 at lower right this ladder doesn't work for white, so black can play the extend (45% for black vs 39% for hanging connection) and white's can't crawl (f18) because then black's cut (c14) would work and no ladder:
Click Here To Show Diagram Code
[go]$$Wm10
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 3 5 . . . . . . . . . . . . . . |
$$ | . 4 2 O X 6 . . . . . . . . . . . . . |
$$ | . . O X . . . . . , . . . . . , X . . |
$$ | . . O X . . . . . . . . . . . . . . . |
$$ | . . . 1 . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


Here is one expected continuation. Some comments: instead of 19 cut AG slightly prefers making a shimari at top right. White's push of 22 used to be defending at 2-2 in the corner but I surmise AG doesn't like allowing black to get the block at 22 in sente, indeed in the last few years this crawl has become a common joseki in pro games so AG seems to confirm they are probably right to prefer it. But then in a departure from normal human joseki (though only 5 hits at this point for corner in waltheri, it's more normal (20 hits) to have had hanging connection instead of extend at f18) AG sacrifices the 2 stones to get the key influence point of f14 in sente ...
Click Here To Show Diagram Code
[go]$$Wm16 Using the f18 extend
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O O . . . . . . . . . . . . . . |
$$ | . X X O X X . . . . . . . . . . . . . |
$$ | 8 3 O X . . . . . , . . . . . , X . . |
$$ | . 9 O X . 2 . . . . . . . . . . . . . |
$$ | . 5 4 O 1 0 . . . . . . . . . . . . . |
$$ | . 7 6 . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


.. and then capture the corner with bad aji. They give 3 moves for white 32 at top right, obviously eyeing that h17 cut.
Click Here To Show Diagram Code
[go]$$Wm26 cont.
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 2 O O 3 5 6 . . . . . . . . . . . |
$$ | . X X O X X 4 . . . . . . . . . . . . |
$$ | X O O X . . . . . , . . . . . 7 X . . |
$$ | . O O X . X . . . . . . . . . . . . . |
$$ | . O X O O X . . . . . . . . . . . . . |
$$ | . O X . . . . . . . . . . . . . . . . |
$$ | . . 1 . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


So because black is able to extend AG thinks white shouldn't have played the small avalanche (44.9% for black) but just descended (42.6%, Dinerstein style vs me!); large avalanche is at 44.7%. Expected continuation is white pushes 4 times to get sente then high approaches top right, as in Master game vs Jiang Weijie with the famous push through table shape.
Click Here To Show Diagram Code
[go]$$Wm10 White no ladder so shouldn't small avalanche
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 1 . . . . . . . . . . . . . . . |
$$ | . . . O X . . . . . . . . . . . . . . |
$$ | . . O X . . . . . , . . . . 7 , X . . |
$$ | . . O X . . . . . . . . . . . . . . . |
$$ | . . 3 2 . . . . . . . . . . . . . . . |
$$ | . . 5 4 . . . . . . . . . . . . . . . |
$$ | . . . 6 . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


So in conclusion (in this position at least), if white can play the small avalanche and black can't extend because white's ladder works then it's good for white so black shouldn't avalanche, if white doesn't have the ladder then black's avalanche isn't so bad (but hane still a little better) and white can't punish with small avalanche so should just descend.


This post by Uberdude was liked by 5 people: Bill Spight, dfan, Knotwilg, lightvector, wolfking
Top
 Profile  
 
Offline
 Post subject: Re: Let's study AlphaGo's opening book
Post #43 Posted: Sat Feb 03, 2018 2:44 am 
Dies with sente

Posts: 118
Liked others: 34
Was liked: 14
Rank: KGS 6 dan
You were right. Good job using the tool properly!
https://youtu.be/Nh9wnfnzkYo?t=6m

Top
 Profile  
 
Online
 Post subject: Re: Let's study AlphaGo's opening book
Post #44 Posted: Sat Feb 03, 2018 1:23 pm 
Judan

Posts: 5882
Location: Cambridge, UK
Liked others: 335
Was liked: 3141
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
johnsmith wrote:
You were right. Good job using the tool properly!
https://youtu.be/Nh9wnfnzkYo?t=6m

Ha, that's serendipitous. I also had a look at what difference it makes with a black 4-4 at top right. If bottom right is 3-4 so white can punish with small avalanche as ladder works then it's a 5% mistake for black. If both black corners are 4-4s then AG actually prefers avalanche to hane for a change (by a smidgen), and recommends white's extension to invite large avalanche but black declines with the atari under and hanging connection.

So AG seems to say starting the avalanche is a serious (~5%) mistake when your opponent has the ladder for the good small avalanche variation with extend, and in terms of direction on the top side it's slightly better with a 4-4 at top right than with a 3-4 (in the aiming for mini-Chinese direction). Not looked at relationship to white's lower left yet.

Top
 Profile  
 
Online
 Post subject: Re: Let's study AlphaGo's opening book
Post #45 Posted: Mon Feb 05, 2018 11:21 am 
Judan

Posts: 5882
Location: Cambridge, UK
Liked others: 335
Was liked: 3141
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Here's another move in a human joseki, very popular around 2010-2012, that AlphaGo thinks is a big mistake; I don't know if there was a consensus among pros it was uneven or just changing fashions. The outside attachment of 9 is a Korean-style joseki, and the connection of 12 one main line (other is push on top and then usually white gets the 3-3 but black gets a better position on the side). Push of 14 and following moves are solid and simple, I think maybe pros decided this joseki was good for black so tried more adventurous moves for white 14 like q10/11; AG likes q10 (but not q11) 3% more than q8, and even better by 1% is n4 shoulder hit to get some support in sente and then q10.

Click Here To Show Diagram Code
[go]$$Bcm9 Popular human joseki and opening around 2010
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . X . . . . |
$$ | . . . O . . . . . , . . . . . , X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , 7 . . |
$$ | . . . . . . . . . . . . . . . 0 8 9 . |
$$ | . . . . . . . . . . . . . . . 6 3 . . |
$$ | . . . . . . . . . . . . . . . 2 1 . . |
$$ | . . . . . . . . . . . . . . . 4 O . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . O . |
$$ | . . . . . . . . . . . . . X . . 5 . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

When black pushes up at 19 he is prepared to sacrifice the 2 stones to build a nice box in front of his shimari. Cutting with 20 is the move AG thinks is very bad, 8% worse than p10 hane. Josekipedia says (source?) this move is solid but slow, I'll check Takao's dictionary later. Iirc this move is in some Kim Sung Rae opening books as an okay move. The descent to 22 is a tesuji: if black goes after the 2 stones then when white cuts and extends at s10 then black can no longer live inside like he could if he captured one stone, so the usual continuation is black sacrifces the 3 stones but can make a hanging connection in sente on the outside to build that box. White usually solids captures the 3 stones with 24 rather than t9 atari for ko because losing the ko is more costly for white to lose than black (even if black doesn't have threats yet he could tenuki start the ko later when he has big threats). But this means black gets a nice atari at t9 later which is good for eyeshape and this makes it much harder for white to invade the a area compared to if black had played s10 solid connection out of fear of ko.

Click Here To Show Diagram Code
[go]$$Bcm19 Solid but slow joseki for white
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . X . . . . |
$$ | . . . O . . . . . , . . . . . , X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . a . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . 5 . |
$$ | . . . , . . . . . , . . . . . 1 X . . |
$$ | . . . . . . . . . . . . . . . O O X . |
$$ | . . . . . . . . . . . . . . . O X 2 4 |
$$ | . . . . . . . . . . . . . . . O X 3 . |
$$ | . . . . . . . . . . . . . . . O O 6 . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . O . |
$$ | . . . . . . . . . . . . . X . . X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

I presume AlphaGo thinks the above sequence is too good for black, so as white would fearlessly atari for 24. AG as black has a nice counter-strategy: invading the top left 3-3 point first to generate large ko threats, and the best-for-both AG sequence is as follows with black win % rising slightly from 57.1 at move 20 to 58.6 at move 35:

Click Here To Show Diagram Code
[go]$$Wm20 AG's large ko threat generation sequence
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 4 5 . . . . . . . . . . X . . . . |
$$ | . . 6 O . . . . . , . . . . . , X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 7 . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . 8 . |
$$ | . . . , . . . . . , . . . . . X X . . |
$$ | . . . . . . . . . . . . . . . O O X 9 |
$$ | . . . . . . . . . . . . . . . O X 1 3 |
$$ | . . . . . . . . . . . . . . . O X 2 . |
$$ | . . . . . . . . . . . . . . . O O . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . O . |
$$ | . . . . . . . . . . . . . X . . X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

Note AG's good technique of 27 as a probe asking "do you want to ko?" before playing the 29 threat which is lossy but creates larger threats before starting the ko. I had previously wondered if it would be able to discover such high-level ko threat amplification techniques, which are easier to understand with logical reasoning, itself through self-play; the answer appears to be yes.

Click Here To Show Diagram Code
[go]$$Bm29 AG's fights ko with ko threat amplification first
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 5 . . . . . . . . . . . . . . . |
$$ | . . X O . . . . . . . . . . X . . . . |
$$ | . . X O . . . . . , . . . . . , X . . |
$$ | . . 2 1 7 . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . X 6 |
$$ | . . . , . . . . . , . . . . . X X 4 3 |
$$ | . . . . . . . . . . . . . . . O O X O |
$$ | . . . . . . . . . . . . . . . O X O O |
$$ | . . . . . . . . . . . . . . . O X X . |
$$ | . . . . . . . . . . . . . . . O O . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . O . |
$$ | . . . . . . . . . . . . . X . . X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

AG also suggests white can block on the left after top left 3-3 (just 0.5% worse, not so meaningful) and then gives the following sequence with white extending solidly to avoid the big ko threats, getting sente to capture on the right and then black, of course, doing another 3-3 invasion with his sente:

Click Here To Show Diagram Code
[go]$$Wm24 AG block other side, no ko
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . 6 . . . . . . . . . . . . . |
$$ | . . X 2 4 5 . . . . . . . . X . . . . |
$$ | . . 1 O 3 . . . . , . . . . . , X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . 8 . |
$$ | . . . , . . . . . , . . . . . X X 7 . |
$$ | . . . . . . . . . . . . . . . O O . 9 |
$$ | . . . . . . . . . . . . . . . O X O O |
$$ | . . . . . . . . . . . . . . . O X X . |
$$ | . . . . . . . . . . . . . . . O O . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . O . |
$$ | . . 0 . . . . . . . . . . X . . X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


P.S. Changing the top left corner to a 3-4, AG still says the cut of 20 is a mistake, but only 6% this time. Also in this line we get to see how much worse it thinks white being a chicken and capturing the 3 stones vs ko is if black does immediate handing connection: 2% (a is black win 58.9, b is 56.7, black is already 10% better than the empty board). So if white avoid the ko black just approach the top left, if ko then AG does a similar 3-3 invasion to generate threats on the lower left 4-4. I've not managed to find this right side joseki with no white 4-4s on the left yet to see if/how it would generate threats then (but that fuseki would be rather weird so probably never happens as adding a move to 3-4 corners bigger than slide, and black more likely to approach than shimari).

Click Here To Show Diagram Code
[go]$$Wm24 Avoid ko with a vs b: -2%
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . . . . . . X . . . . |
$$ | . . . , . . . . . , . . . . . , X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . X . |
$$ | . . . , . . . . . , . . . . . X X . . |
$$ | . . . . . . . . . . . . . . . O O X b |
$$ | . . . . . . . . . . . . . . . O X O O |
$$ | . . . . . . . . . . . . . . . O X X . |
$$ | . . . . . . . . . . . . . . . O O a . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . O . |
$$ | . . . . . . . . . . . . . X . . X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


This post by Uberdude was liked by 4 people: Bill Spight, Gomoto, lightvector, zermelo
Top
 Profile  
 
Online
 Post subject: Re: Let's study AlphaGo's opening book
Post #46 Posted: Tue Feb 06, 2018 1:56 pm 
Judan

Posts: 5882
Location: Cambridge, UK
Liked others: 335
Was liked: 3141
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Is anyone reading this? Anyway, even if not I'll keep doing it as it's also for my own enjoyment and documenting interesting things. :)

A shape I've noticed AlphaGo likes to play a lot is the 5th line "peep" (I don't know what to call it) against a large knight move of the opponent from 3rd to 4th line. It's not a new move, humans play it too, but AG does seem to like it a lot and plays it where I think human pros wouldn't (either because they don't consider it or do but dismiss it as aji keshi). Here's an example from AG ancient history, game 2 vs Lee Sedol; this one is a reduction in middle-game position so a fairly typical human usage, in the AG Master teaching tool it's played in looser early opening positions.

Click Here To Show Diagram Code
[go]$$B
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . X . . . . . . . . . . |
$$ | . . X , X . . . . , . . . X . X . . . |
$$ | . . . . . . . . . . . . . . . . X . . |
$$ | . . . . . . . . . . . . . . . O O . . |
$$ | . . W . . . . . . . . . . . . . . . . |
$$ | . . . . 1 . . . . . . . O . . . . . . |
$$ | . . . 2 . . . . . . . . . . O O . . . |
$$ | . . . W . . . . . , . . . X X , . . . |
$$ | . . . O X . . . . . . . . . . . . . . |
$$ | . . O X . X . . . . . . . . . . . . . |
$$ | . . O O X X . . . . . . . . . . . . . |
$$ | . O X X X O O . . . . . . . . . O . . |
$$ | . O O X O X O X . . . . . . . X O . . |
$$ | . X X O O . X O . X . . X . X , O . . |
$$ | . X O X . O O . . . . . . X O O . . . |
$$ | . . . X . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]



Here's the same shape with 13 in a micro-Chinese variation. Some brief comments on preceding moves:
- approach of 5 is 45.3%, all 4 shimaris at top left are better with an amusing symmetry: high is 1% better than low, big is a half % better than small.
- answer at 6 is best local move, all pincers worse, but best (by >1%) is to approach top left high or low.
- 7: AG approves of micro-Chinese, best move, shimaris are close behind, mini-Chinese one space lower is -1.5%.
- 8: as always, AG not a fan of wedge (-2.3%), prefers to approach top right from inside or best is outside and go for a mutual moyo type game.
- 9: AG's best, and all 3 normal checking moves on this side (m/n/o17) are better than ones on the left of the wedge.
- 10: AG not impressed by this cool-looking 2nd-line that was fashionable for a while, -4% on normal 2-space extension which is best.
- 11: AG agrees with the humans who play this kosumi in preference to the high shimari answer (and it is human 1st choice 16 to 9 and bigger win %), by 2%.
- 12: normal
- 13: The topic of this post as the move AG likes, but checking human games with this position it is also pros favourite move, 5 hits in waltheri (all in 2012-13)! This obviously aims at splitting white at j17, and then also covering at g15 if white defends the top. The most common human response is a, which defends j17 and also makes the g15 attachment ineffective.

Click Here To Show Diagram Code
[go]$$Bcm5
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . 6 . . . . . . . . . . . . . |
$$ | . . . X . . . . . 4 . . . 5 . . . . . |
$$ | . . . , 7 . 8 a . , . . . . . X . . . |
$$ | . . . . . . . . 9 . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 3 . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 1 . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . O . . . |
$$ | . . . . . 2 . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


AlphaGo tool gives 2 responses for white, the obedient one at j16 (1 human game), and resisting with d15 (no humans). j16 has the advantage of taking a liberty from the black stone and better for coming out to right centre (e.g. k15), downside is g15 is still a decent black move (the human did it immediately). AG 3-3 invades the lower right first:

Click Here To Show Diagram Code
[go]$$Wcm14 AG continuation if defend: 3-3...
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . O . . . . . . . . . . . . . |
$$ | . . . X . . . . . O . . . X . . . . . |
$$ | . . . , X . O . 1 , . . . . . X . . . |
$$ | . . . . . . . . X . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . 0 . |
$$ | . . X . . . . . . . . . . . . . 8 7 . |
$$ | . . . . . . . . . . . . . . . 9 5 6 . |
$$ | . . . O . . . . . , . . . . . O 4 . . |
$$ | . . . . . O . . . . . . . . . 3 2 . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


Click Here To Show Diagram Code
[go]$$Wcm24 ... then block at 27
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . O . . . . . . . . . . . . . |
$$ | . . . X . . . . . O . . . X . . . . . |
$$ | . . . , X 9 O . O , . . . . . X . . . |
$$ | . . . . . . 4 8 X . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . 6 5 . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . 7 . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . X . |
$$ | . . X . . . . . . . . . . . . . X . 2 |
$$ | . . . . . . . . . . . . . . . O O X . |
$$ | . . . O . . . . . , . . . . . O X 1 . |
$$ | . . . . . O . . . . . . . . . O X 3 . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


Something I noticed is white plays the double hane joseki here, which is not rare for AG but jump is more common. I think this could be because j15 broke the ladder: we don't get a win% for jump but my suspicion is AG thinks that is better for white than the double hane if white has the ladder. So the 3-3 (and j15 to set it up) could be seen as urgent and wanting to get the good outside ponnuki result whilst the ladder works for black, but tbh I don't really see why getting that outside ponnuki in sente is a good exchange: AG likes corners and W got the corner.

Click Here To Show Diagram Code
[go]$$Wcm18 The joseki that needs ladder
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . O . . . . . . . . . . . . . |
$$ | . . . X . . . . . O . . . X . . . . . |
$$ | . . . , X . O . O , . . . . . X . . . |
$$ | . . . . . . . . X . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . a 7 . |
$$ | . . X . . . . . . . . . . . c 0 1 4 . |
$$ | . . . . . . . . . . . . . . . 3 2 . . |
$$ | . . . O . . . . . , . . . . . O X . . |
$$ | . . . . . O . . . . . . . . . O X . . |
$$ | . . . . . . . . . . . . . . 9 8 5 6 . |
$$ | . . . . . . . . . . . . . . . . b . . |
$$ +---------------------------------------+[/go]


Instead of obediantely answering j15, AG suggests resisting and mutual destruction is a little (0.6%) better. Is there any meaning to interposing 19-20 during the top side sequence, or just random bot flitting?

Click Here To Show Diagram Code
[go]$$Wcm14 AG continuation if resist...
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . O . . . . . . . . . . . . . |
$$ | . . . X . . . . 4 O . . . X . . . . . |
$$ | . . . , X . O 8 5 9 . . . . . X . . . |
$$ | . . . 1 2 . . 0 X . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 3 . . . . . . . . . . . . . . . |
$$ | . . X 6 7 . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . O . . . |
$$ | . . . . . O . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


Here I can see the benefit of making the 27-28 exchange before taking gote at 29, black probably doesn't want white to extend at 27 if at 29 immediately.
Click Here To Show Diagram Code
[go]$$Wcm24 ... cont
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . O 6 . 3 . . . . . . . . . . |
$$ | . . . X . . 2 1 . O . . . X . . . . . |
$$ | . . . , X . O X O O . . . . . X . . . |
$$ | . . . O X . . X X . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . 5 . . . . . . . . . . . . . |
$$ | . . X X O . . . . . . . . . . . . . . |
$$ | . . . . 4 7 . . . . . . . . . . . . . |
$$ | . . . , 8 . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . O . . . |
$$ | . . . . . O . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


This post by Uberdude was liked by 3 people: Baywa, Bill Spight, wolfking
Top
 Profile  
 
Offline
 Post subject: Re: Let's study AlphaGo's opening book
Post #47 Posted: Tue Feb 06, 2018 9:01 pm 
Lives with ko

Posts: 258
Liked others: 46
Was liked: 204
Rank: maybe 2d
Uberdude wrote:
Is anyone reading this? Anyway, even if not I'll keep doing it as it's also for my own enjoyment and documenting interesting things. :)


I'm reading this! Please keep it up, this thread is awesome. :bow:

Top
 Profile  
 
Offline
 Post subject: Re: Let's study AlphaGo's opening book
Post #48 Posted: Tue Feb 06, 2018 9:04 pm 
Lives in gote

Posts: 395
Liked others: 1
Was liked: 103
Rank: KGS 2k
GD Posts: 100
KGS: Tryss
Uberdude wrote:
Is anyone reading this?


I am :mrgreen:

Top
 Profile  
 
Offline
 Post subject: Re: Let's study AlphaGo's opening book
Post #49 Posted: Tue Feb 06, 2018 10:32 pm 
Dies in gote

Posts: 28
Liked others: 1
Was liked: 1
Rank: OGS 2k
KGS: Mikebass14
Tygem: Mikebass14
OGS: Mikebass14
Me too! Maybe too fast to really absorb, but enjoying it all the same. Thanks!

Top
 Profile  
 
Online
 Post subject: Re: Let's study AlphaGo's opening book
Post #50 Posted: Wed Feb 07, 2018 5:00 am 
Judan

Posts: 5882
Location: Cambridge, UK
Liked others: 335
Was liked: 3141
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Ok, thanks for encouragement!

I'll try to find an example of this shape that humans didn't do for the next post. Btw, does anyone know if it has a name, either English or C/J/K?

Top
 Profile  
 
Offline
 Post subject: Re: Let's study AlphaGo's opening book
Post #51 Posted: Wed Feb 07, 2018 5:58 am 
Oza

Posts: 2329
Liked others: 15
Was liked: 3346
Quote:
I'll try to find an example of this shape that humans didn't do for the next post. Btw, does anyone know if it has a name, either English or C/J/K?


It's called nozomi (gazing from a distance) in Japanese.


This post by John Fairbairn was liked by: Uberdude
Top
 Profile  
 
Online
 Post subject: Re: Let's study AlphaGo's opening book
Post #52 Posted: Wed Feb 07, 2018 6:43 am 
Judan

Posts: 5882
Location: Cambridge, UK
Liked others: 335
Was liked: 3141
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
John Fairbairn wrote:
It's called nozomi (gazing from a distance) in Japanese.

Thanks: that's nice, fits my feeling of wanting to call it a peep but a bit further away.

Top
 Profile  
 
Online
 Post subject: Re: Let's study AlphaGo's opening book
Post #53 Posted: Wed Feb 07, 2018 3:06 pm 
Judan

Posts: 5882
Location: Cambridge, UK
Liked others: 335
Was liked: 3141
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Here's an example of AG wanting to play this nozomi from one of the master games from 2017 new year. The human pro played a instead for 14 which is a classic large extension eyeing the attachment at b in the corner next. AG's move destroys this possibility, but it thinks it's 2.6% better (black is already at 54%, mostly because AG didn't like the c11 tenuki). What does it gain? Importantly it keeps sente which AG likes (and then uses to enclose lower left), but being sente is nothing good in itself, why not just enclose directly? I thought it might make the ladder good for the 3-3 invasion joseki from before, but it doesn't, plus that joseki probably wouldn't happen with r10 in place; maybe it changes another ladder. It helps c11 and develops the centre a little bit, plus I get the feeling it makes black c13 smaller, and also has a potential followup at g16 to reduce the top and develop the centre, but those all seem quite vague to me. I wonder what the win% would be for f3 directly, I suspect much the same: it's not an essential exchange to play now, but who knows...

Click Here To Show Diagram Code
[go]$$Wcm10 AlphaGo (black) vs Yang Dingxin 5p
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . b . . X . . . . . . . O X X . . . |
$$ | . . X 6 . . . . . , . . 4 2 O , X . . |
$$ | . . . . 5 . . . . . . . . . . . . . . |
$$ | . . a . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 1 . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , 3 . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . O . . . |
$$ | . . . . . 7 . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

Top
 Profile  
 
Offline
 Post subject: Re: Let's study AlphaGo's opening book
Post #54 Posted: Wed Feb 07, 2018 3:18 pm 
Dies with sente

Posts: 118
Liked others: 34
Was liked: 14
Rank: KGS 6 dan
I'm reading too. From time to time :) it's very tiring to make these ASCI diagrams (or whatever the name is) or I don't know the easy way. So I can't exchange my opinions easily and end up not writing anything... Are screenshots instead of diagrams desirable on this forum? Something like this:
Attachment:
sshot-1.jpg
sshot-1.jpg [ 57.71 KiB | Viewed 3681 times ]
By the way: Zen 7 is proposing only this "bad" move for white, and it's the only move it considers :o. While the most recent LZ network d6f3a68b plays the correct move. I noticed many good moves like this by LZ that AG taught us.

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


Last edited by johnsmith on Wed Feb 07, 2018 4:31 pm, edited 3 times in total.
Top
 Profile  
 
Online
 Post subject: Re: Let's study AlphaGo's opening book
Post #55 Posted: Wed Feb 07, 2018 3:39 pm 
Judan

Posts: 5882
Location: Cambridge, UK
Liked others: 335
Was liked: 3141
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
johnsmith wrote:
it's very tiring to make these ASCI diagrams (or whatever the name is) or I don't know the easy way.

Thanks to Herman: http://hiddema.nl/diagrammer/


This post by Uberdude was liked by 2 people: luigi, zermelo
Top
 Profile  
 
Online
 Post subject: Re: Let's study AlphaGo's opening book
Post #56 Posted: Wed Feb 07, 2018 4:27 pm 
Judan

Posts: 5882
Location: Cambridge, UK
Liked others: 335
Was liked: 3141
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
And another from the Master series. In the real game Master played j3 for 28 (best, Park Junghwan as black was already down at 40%), but the teaching tool version also likes (-0.2%) the nozomi at o15 on top, expecting black to answer at p17 and then come back to j3. With j3 directly it then approves of Park's c5 attachment etc (miai with the extension on lower side I think). In the real game Master played 34 at f8, an elegant shoulder hit that Inseong praised in one of his videos. AG teaching tool version suggests f7 or the o15 nozomi again. So it doesn't really matter if it was before or after j3, but it does seem to think it's a key point. I think many humans (myself included) would be somewhat reluctant to exchange it for black's defence as there was previously some chance to live or ko at 3-3 (and indeed in the real game Master did live at 3-3 later).

Click Here To Show Diagram Code
[go]$$Wcm28 AG Master (white) vs Park Junghwan 9p
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . O . . . . . . . . . . . . . . |
$$ | . . . X O . . O . X . . X . 8 . . . . |
$$ | . . . X . . . . . , . . . . . X . . . |
$$ | . . X . . O . . . . . . . 7 . . . . . |
$$ | . . X O . . . . . . . . . . . . X . . |
$$ | . . X X O O . . . . . . . . . . . . . |
$$ | . . . O X . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . 6 . . . . . . . . . . . . . . |
$$ | . . 4 . . . . . . . . . . . . . O . . |
$$ | . . 2 3 . . . . . . . . . . . . . . . |
$$ | . 5 O , . . . . . , X . . X . O . . . |
$$ | . . . . O . . . 1 . . . . X O . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


But I think why AG likes this exchange become clear in the followup at m16, it wants to press black low, if block at m17 then k16 cover is nice, so black resists. In fact many moves earlier when Park defended at n17 AG preferred (by <1%) him to jump at k15 instead, maybe this future nozomi-press overconcentration combination is one reason why.

Click Here To Show Diagram Code
[go]$$Wcm36 nozomi continuation
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . O . . . . . . . . . . . . . . |
$$ | . . . X O . 5 O . X . 3 X . X . . . . |
$$ | . . . X . . . 4 2 , . 1 . . . X . . . |
$$ | . . X . . O . . . . . . . O . . . . . |
$$ | . . X O . . . . . . . . . . . . X . . |
$$ | . . X X O O . . 6 . . . . . . . . . . |
$$ | . . . O X . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . X . . . . . . . . . . . . . . |
$$ | . . X . . . . . . . . . . . . . O . . |
$$ | . . X O . . . . . . . . . . . . . . . |
$$ | . O O , . . . . . , X . . X . O . . . |
$$ | . . . . O . . . O . . . . X O . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


This post by Uberdude was liked by 2 people: Bill Spight, zermelo
Top
 Profile  
 
Offline
 Post subject: Re: Let's study AlphaGo's opening book
Post #57 Posted: Wed Feb 07, 2018 4:44 pm 
Dies with sente

Posts: 118
Liked others: 34
Was liked: 14
Rank: KGS 6 dan
Park also quickly adopted AG's joseki in today's game against Ke Jie 9p. Instead of 'a', he continued with 1 (modified AG joseki).
Click Here To Show Diagram Code
[go]$$Wc Ke(black) vs Park(white)
$$ --------------------+
$$ . . . . . . . . . . |
$$ . . . . . . . . . . |
$$ . . 6 4 3 1 O O . . |
$$ , . . 5 2 X X X a . |
$$ . . . . . . . . . . |
$$ . . . . . . . O . . |
$$ . . . . . . . . . . |
$$ . . . . . . . X . . |
$$ . . . . . . . . . . |
$$ , . . . . . , . . . |[/go]

Top
 Profile  
 
Offline
 Post subject: Re: Let's study AlphaGo's opening book
Post #58 Posted: Thu Feb 08, 2018 3:16 am 
Oza

Posts: 2329
Liked others: 15
Was liked: 3346
Quote:
But I think why AG likes this exchange.... this future nozomi-press overconcentration combination is one reason why.


While not in any way demurring from your conclusion, I again raise the likelihood that this has all been considered before by pros, especially in the New Fuseki theorising. 5th line plays attracted them - recall Kitani's famous game where he started with three of them, like a high sanrensei. Among the main reasons were adaptability and potentiality for territory.

For human pros, adaptability was perhaps the bugbear. They soon found themselves incapable of controlling the complications in the centre. But that did not mean they gave up their views on the ideal theory. I think what we are seeing with AG is proof of that concept because AG can control the complications. Under that view, should humans try harder to cope with centre complications, or just give up, the way New Fuseki-ites did?

The answer to that may lie in the second point: potentiality for territory.

One of the main but counter-intuitive strands in New Fuseki thinking was that 5th line plays were good for making territory. You need to refer to the diagrams in Shin Fuseki-ho to get a firm grip on that, but, to boil it down, the main conclusion was that a box is geometrically much more efficient than a tray for garnering territory. This was manifested in two ways.

One was that 5th line plays can tend to end up in box shapes for the 5th-line player. It's not New Fuseki, but there's an interesting example of such innovative thinking by Dosaku in his earliest datable game (1665). Here he played a long series of presses (perhaps AG copied this :)) to get a long wall and then sealed off a 5th line territory around it - not using his thickness to attack, but to make territory. In fact there were several other early presses in this game by both sides, so what we may be seeing is the emergence of a strand of theory that was new at the time. It may even be considered pre-thickness and that it led to the concept of centre thickness - another feather in Dosaku's cap?

But the other way the box-tray dichotomy came up was the realisation that pressing your opponent down into a tray shape (using presses and shoulder hits mainly) was, in theory, a good thing for you (and bonus points if your pressing moves got a box side for yourself, of course). This didn't quite get off the ground. One reason must have been the ensuing difficulties, already mentioned, in dealing with nebulous centre positions, but I suspect lack of komi also played a big part. White would surely have been reluctant to give Black solid territory, even in a tray shape, when Black had the cushion of no komi, and at the same time White, having to rely on seeking complications to overcome Black's first-move advantage, would have been reluctant to settle positions and so remove aji.

The introduction of komi did not solve White's problems at once, because first it was only 4.5 points and then only 5.5. points. The big komi is a relatively recent development.

On that reading, modern pros have made quite the same advances in strategy/theory that they did in tactics/fighting, and AG is reminding them to go back and think more about what previous theoreticians, from Dosaku to Kitani and Go Seigen, thought about.

That and taking more of an interest in how AG controls the later centre game (box shape potentiality?) seems more likely to me to help humans improve, rather than quibbling over joseki minutiae. Uberdude's nozomi observation seems like a very good start.

Top
 Profile  
 
Online
 Post subject: Re: Let's study AlphaGo's opening book
Post #59 Posted: Mon Feb 12, 2018 7:49 am 
Judan

Posts: 5882
Location: Cambridge, UK
Liked others: 335
Was liked: 3141
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
A brief interlude in the nozomi theme with something I found remarkable in the AG vs Lee Sedol game 2 opening.
Click Here To Show Diagram Code
[go]$$Bcm11 Ok for B, 46.8% after 13
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . 3 . . . . . . . . . . |
$$ | . . X , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X . . . . . . . . . . . . . 2 . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . 1 . X , O . . |
$$ | . . . . . O . . . . . . . X O O . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

AG made the hanging connetion of 11, Lee played the usual joseki one-space extension to 12 and then AG tenukid to make a Chinese opening rather than extend on the lower side. The AG Master version which made this teaching tool has some different evaluations to the weaker AG Lee version, but it gives black win% in this position as 46.8%. Something I remember being taught (though how accurately does this reflect the current thinking of top pros?) is that making the connection makes the 2 black stones turn into 3 stones and become heavier: if just 2 you can tenuki and treat them lightly if white cuts, but by connecting you are commiting more to making a group out of them so extendind on the lower side would be normal (and white's extenion helped white on the right side and made it hard to develop there). However, AG strongly disagrees with this, saying the immediate tenuki to make a Chinese is a whopping 9% worse, and white punishes by cutting.
Click Here To Show Diagram Code
[go]$$Bcm11 Bad for B, 37.6% after 11
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . 1 . . . . . . . . . . |
$$ | . . X , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . X , O . . |
$$ | . . . . . O . . . . . . . X O O . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

So why the huge difference? To start with AG teaching tool actually thinks the hanging connection was bad (-1% on solid connection) and black is already down at 42%: high approach was bad (better low), and approach on left bad (better 3-3 or connect). But then 12 joseki extension is a big mistake, -5.5%, better to approach the top left 3-4. If black made the solid connection then AG thinks extend is a good move; this is something that has changed as AG evolved, it now generally considers the hanging connection soft as it's gote whereas the solid connection has a more severe follow up of the attachment and then hane (r5 and s6) that Go Seigen liked. So I think we can understand this as AG views the hanging connection for extend as a good exchange for black, making the 2 stones stronger in sente as white made a soft answer, rather than a bad exchange making them heavier. Also AG Master is not impressed by the famous q5 peep: it thinks it is a mistake (45.5%, c4 attach best at 48%) because white should resist with the o5 counter peep but if white meekly connects at r5 (as Lee Sedol did) then black goes up to 49%. I.e. peep was an overplay that Lee didn't punish.

Top
 Profile  
 
Offline
 Post subject: Re: Let's study AlphaGo's opening book
Post #60 Posted: Mon Feb 12, 2018 11:43 am 
Honinbo

Posts: 8493
Liked others: 2484
Was liked: 2947
Click Here To Show Diagram Code
[go]$$Bcm11 Bad for B, 37.6% after 11
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . 1 . . . . . . . . . . |
$$ | . . X , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . W . . . . . , . . . 2 X , O . . |
$$ | . . . . . W . . . . . . . X O O . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]


I wonder if one reason for the bad marks for :b11: has to do with the presence of the :wc: stones, in that they make the cut, :w2:, more potent than if the bottom left corner were empty.

_________________
There is one human race.
----------------------------------------------------

The Adkins Principle:

At some point, doesn't thinking have to go on?

— Winona Adkins


Last edited by Bill Spight on Mon Feb 12, 2018 3:27 pm, edited 1 time in total.
Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 64 posts ]  Go to page Previous  1, 2, 3, 4  Next

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: Google [Bot] and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group