It is currently Sun Mar 29, 2020 2:02 pm

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 10 posts ] 
Author Message
Offline
 Post subject: Shape problem - pro level
Post #1 Posted: Mon Nov 18, 2019 4:39 am 
Oza

Posts: 2474
Liked others: 16
Was liked: 3582
I chose the title of this thread with care. Shape was the first word. I'll bet that made quite a few eyes light up. After many decades of go I have learned that nothing gets amateur juices going better than the word 'shape.' Yet it is hardly ever mentioned in pro commentaries. So, is obsession with shape an amateurs' flaw (like obsession with joseki), or is it something pros got from their mother's milk and just do it, instead of talking about it?

I can across an example today of a Meijin (Honinbo Shuei) commenting on a move by Nozawa Chikucho 3-dan which, if nothing else, knocks the mother's milk idea on the head. But it also seems to call into question the whole amateur approach to shape (monkey see, monkey do).

In the position below Nozawa played the triangled move (24( and Shuei commented simply: "For White 24 it is better to jump to A." Commentaries in those days were for instruction, not therapy sessions. Players had to work out the reasons for their mistakes on their own.

There was no actual mention of shape, but even as an amateur I feel confident in saying that shape is behind this comment. So, a pro 3-dan good enough to sit at the feet of the master (and to improve very significantly later on) made a shape mistake. How do we explain the mistake?

It's easy enough to fall into the trap of saying 24 makes an empty triangle, tout court. But if Shuei had said 24 was better than A, we would all just as confidently quote another proverb and say it was because of five alive - White is filling in a liberty of a weaker enemy group. Those ways of looking at things are just like writing an essay at school by copying paragraphs from different books and calling it research.

But what is the proper way to look at it? Again it's easy to sound magisterial and say it depends on the aji or overconcentration left behind, or something like that. But then how do we evaluate the results?

I don't know. Remember my question was "How do we explain the mistake?" Not "what is the mistake?" - although that is stage we may have to go through.



(If you want to see how the game actually unfolded, it is game 1906-04-16a in the GoGoD database, and it is also in my e-book of Shuei's commentaries.)

Top
 Profile  
 
Online
 Post subject: Re: Shape problem - pro level
Post #2 Posted: Mon Nov 18, 2019 6:35 am 
Judan

Posts: 6356
Location: Cambridge, UK
Liked others: 366
Was liked: 3460
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
In some similar positions this empty triangle is actually better than the jump despite the superficial 'bad shape empty triangle', because not only does it take a liberty it creates 2 points more eyespace which is a big difference (it can change the white corner from being a dead L group to (practically) alive carpenter's square). This is based on the following assumptions which are obviously highly dependent on timing and nearby fighting circumstances:

If empty triangle then if black descends and white will block then following a-b exchange white has a carpenter's square which is hard to kill (particularly with the liberty) and generally a ko or a seki and even if dead has quite a few libs for counterattacking the outside groups whilst you are dying.
Click Here To Show Diagram Code
[go]$$W
$$ +------------------
$$ | . . . . . . . . .
$$ | . . . b a . . . .
$$ | . . . O . . . . .
$$ | . 3 1 O X X X . .
$$ | . 2 X X O O . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .[/go]


If jump then rather than just blocking black can push first and then block, then if white solid connects (not only choice, hanging connect for ko or tenuki also possible) then it's now a locally dead L group following a-b-c. Black might not like to push at 2 because it loses a valuable lib, and if block at 4 directly then there is the choice for white to maximize eyespace at 2 reverting to carpenter's square, or tenuki in which case black's followup could be clamp at 5 (often leading to corner ko) rather than push 2 and take a stone if he is in an attacking position rather than worrying about eyes for the 2 stones.
Click Here To Show Diagram Code
[go]$$W
$$ +------------------
$$ | . . . . . . . . .
$$ | . . . b a . . . .
$$ | . 5 3 O . . . . .
$$ | c 1 2 O X X X . .
$$ | . 4 X X O O . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .[/go]


Here given the closeness of the c12 stone the black group is in imminent danger so the b15 descent (in reply to empty triangle) or block (in reply to b16 jump) is far from a guaranteed sente against the corner and needs concrete reading of potential resistances and capturing races rather than just general shape principles.

In fact this jump and push then block and tenuki happened in the Iyama vs Shibano Oza title match 2 days ago: https://lifein19x19.com/viewtopic.php?p=250771#p250771.

Top
 Profile  
 
Offline
 Post subject: Re: Shape problem - pro level
Post #3 Posted: Mon Nov 18, 2019 7:12 am 
Gosei

Posts: 1536
Location: Earth
Liked others: 571
Was liked: 248
Quote:
Those ways of looking at things are just like writing an essay at school by copying paragraphs from different books and calling it research.

Not only at school, it is quite common at universities as well.


Quote:
But then how do we evaluate the results?

I do not know how you evalute, but I evaluate like this :)
(Dont take this too seriously. For a real evaluation one has to compare with different bots and enter/force some variations. I am well aware John had something different in mind. ;-) But I thought I just copy two pictures and call it an evalution.)
Attachment:
shape1.JPG
shape1.JPG [ 189.9 KiB | Viewed 699 times ]

Attachment:
shape2.jpg
shape2.jpg [ 189.79 KiB | Viewed 699 times ]

Top
 Profile  
 
Offline
 Post subject: Re: Shape problem - pro level
Post #4 Posted: Mon Nov 18, 2019 7:40 am 
Gosei

Posts: 1536
Location: Earth
Liked others: 571
Was liked: 248
Quote:
But it also seems to call into question the whole amateur approach to shape (monkey see, monkey do).


John, while I respect the difference between pros and amateurs, I do not think it is helpful to remind beginners and amateurs all the time of an unsurmountable gap even if it is real. I once overheard a talk between Cho Hye Yeon and a strong western amateur preparing for a tournament game the next day with a pro, where she explicitly warned the amateur against this thought. Even if there is such a gap, it is not advisable to for a teacher to tell his pupils they will never reach the golden fruit. I prefer to be the snake in the garden and offer the AIpple of sapientiae to bridge the enormous gap.

AI does not make us experts, we will stay humble amateurs and diletants for the rest of our lives. But with the AI experts at our side we can enjoy go even more. And for most amateurs and beginners this is a worthwile goal. John, I also enjoy go more because of your insightful books and posts, but I do not enjoy the view of the unsurmountable wall around eden you paint from time to time, allthough I am quite happy with me staying outside with the other mortals, it is quite fun on this side.

monkey see, monkey do is an underrated learning concept by the way ;-)


This post by Gomoto was liked by: dfan
Top
 Profile  
 
Offline
 Post subject: Re: Shape problem - pro level
Post #5 Posted: Mon Nov 18, 2019 9:51 am 
Honinbo

Posts: 9514
Liked others: 3001
Was liked: 3166
John Fairbairn wrote:
I chose the title of this thread with care. Shape was the first word. I'll bet that made quite a few eyes light up. After many decades of go I have learned that nothing gets amateur juices going better than the word 'shape.' Yet it is hardly ever mentioned in pro commentaries. So, is obsession with shape an amateurs' flaw (like obsession with joseki), or is it something pros got from their mother's milk and just do it, instead of talking about it?


A lot of players got strong through fighting, and it seems to me that they don't particularly care about shape. For an example see the commentary on :b25: by Kato Masao in How is your positional judgement? Game 16 at https://lifein19x19.com/viewtopic.php?p=250450#p250450 . Kato willingly takes bad shape (an empty triangle) to threaten to capture a couple of stones and save one of his own. Elf thinks that :b25: loses 30%! :shock:

Some years ago I did a search for some go terms in the online digital collection of go books in the Library of the Japanese National Diet, and found a book, Shape and Tesuji by Iwasa Kei (1925) at http://dl.ndl.go.jp/info:ndljp/pid/932263 . Iwasa starts off by saying that the most important thing in go is shape (囲棋で一番大切なのは何と云っても形である). An overstatement, I expect, but he seems to have a broader view of shape than other writers. Here is his first example of bad shape.

Click Here To Show Diagram Code
[go]$$Bc Bad shape
$$ --------------------
$$ | . . . . . . . . . .
$$ | . . . . . . . . . .
$$ | . . O 2 . . . . . .
$$ | . . X 1 W . . . . .
$$ | . . . . . . . . . .
$$ | . . . . . . . . . .[/go]


I hadn't looked at the book before, I just made a bookmark. But seeing this bolsters my view that shape is about local efficiency.

John Fairbairn wrote:
I can across an example today of a Meijin (Honinbo Shuei) commenting on a move by Nozawa Chikucho 3-dan which, if nothing else, knocks the mother's milk idea on the head. But it also seems to call into question the whole amateur approach to shape (monkey see, monkey do).

In the position below Nozawa played the triangled move (24( and Shuei commented simply: "For White 24 it is better to jump to A." Commentaries in those days were for instruction, not therapy sessions. Players had to work out the reasons for their mistakes on their own.

There was no actual mention of shape, but even as an amateur I feel confident in saying that shape is behind this comment. So, a pro 3-dan good enough to sit at the feet of the master (and to improve very significantly later on) made a shape mistake. How do we explain the mistake?

It's easy enough to fall into the trap of saying 24 makes an empty triangle, tout court. But if Shuei had said 24 was better than A, we would all just as confidently quote another proverb and say it was because of five alive - White is filling in a liberty of a weaker enemy group. Those ways of looking at things are just like writing an essay at school by copying paragraphs from different books and calling it research.

But what is the proper way to look at it? Again it's easy to sound magisterial and say it depends on the aji or overconcentration left behind, or something like that. But then how do we evaluate the results?

I don't know. Remember my question was "How do we explain the mistake?" Not "what is the mistake?" - although that is stage we may have to go through.



(If you want to see how the game actually unfolded, it is game 1906-04-16a in the GoGoD database, and it is also in my e-book of Shuei's commentaries.)


Iwasa's bad shape example is relevant to the present discussion, because :w24: in the game makes the same kind of mistake. Elf thinks that :w24: loses 13% by comparison with 25.

Click Here To Show Diagram Code
[go]$$Wcm24 White gives Black an extra dame
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 5 a . . . . . . . . . . . . . . . |
$$ | . 3 . O . . . . O . X . . . O . . . . |
$$ | . 2 1 O X X X . . , . . . . . , X . . |
$$ | . 4 X X O O . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . O . . . . |
$$ | . . . . . . . . . . . . . . . . X . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . X . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . X . . . |
$$ | . . X , . . . . . , . . . . . , . . . |
$$ | . . . . O . . . . O . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]


:w24: allows :b25: to hane at the head of White's stones, reducing them to 3 dame. After :b27: Black has 5 dame while White needs protecting. BTW, Elf thinks that :w28: is a minor mistake, which would be better at a. Once you see it it is kind of obvious that a is at least as good as :w28:.

If :w24: is at 25, Black does not get that fifth dame, and in this case that makes a big difference. In Elf's mainline variation Black sacrifices the corner stones. A full explanation would probably get into the complexities of the possible fight scenarios. I think the fact that :w24: gives Black an extra dame is the main point. :)

Edit: Corrected :b25: and :w26: to :w28:

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins
----
Many are cold, but few are frozen.


Last edited by Bill Spight on Mon Nov 18, 2019 11:44 am, edited 2 times in total.
Top
 Profile  
 
Online
 Post subject: Re: Shape problem - pro level
Post #6 Posted: Mon Nov 18, 2019 10:32 am 
Judan

Posts: 6356
Location: Cambridge, UK
Liked others: 366
Was liked: 3460
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
I am suspicious of white 28 in the game: it seems too obedient to allow black the hane rather than descent in sente, and to live so meekly that black can still force at e18. Surely white can fight back with 28 at e18 or f18? (f18 would be easier to play if white hadn't force black to play at g16 though, how about windmill at e14 instead of push?) Yes there is bad aji with the clamp in the corner. but I'd rather die in a ko that live so painfully whilst allowing black to get the 2nd line hane/turn than just descent. There was a similar issue in one of the bonus AlphaGo self play games for weiqi TV, see https://www.alphago-games.com/view/even ... /2/move/26, Chinese video review at https://www.youtube.com/watch?v=Ukt3mSBZKF0, Dhu's translation at: https://lifein19x19.com/viewtopic.php?p=233258#p233258.

Top
 Profile  
 
Offline
 Post subject: Re: Shape problem - pro level
Post #7 Posted: Mon Nov 18, 2019 11:59 am 
Honinbo

Posts: 9514
Liked others: 3001
Was liked: 3166
Uberdude wrote:
I am suspicious of white 28 in the game: it seems too obedient to allow black the hane rather than descent in sente, and to live so meekly that black can still force at e18. Surely white can fight back with 28 at e18 or f18?


Sorry for the typos. I wrote :w26: instead of :w28:. Elf agrees that :w28: is a mistake, but it prefers d18 instead.

Uberdude wrote:
(f18 would be easier to play if white hadn't force black to play at g16 though, how about windmill at e14 instead of push?)


Elf prefers the windmill over the push by 15% (Edited typo). After which Black will jump down to E-18, then White D-18, Black kosumi at B-16, White B-17, Black D-14, White G-15, Black G-18.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins
----
Many are cold, but few are frozen.


Last edited by Bill Spight on Mon Nov 18, 2019 2:47 pm, edited 2 times in total.
Top
 Profile  
 
Offline
 Post subject: Re: Shape problem - pro level
Post #8 Posted: Mon Nov 18, 2019 12:08 pm 
Oza

Posts: 2474
Liked others: 16
Was liked: 3582
Quote:
John, while I respect the difference between pros and amateurs, I do not think it is helpful to remind beginners and amateurs all the time of an unsurmountable gap even if it is real.


I do not take offence but neither do I agree with you nor feel chidden. I am not of the snowflake generation and I don't do kowtowing. I had thought of explaining more (cultural and geographical factors as well as generational) but decided against it. Why? Well you address me as John (thank you), but how do I address you? As I have mentioned many times here, anonymity is a bugbear of mine. It's not friendly.

Bill
Quote:
Iwasa starts off by saying that the most important thing in go is shape (囲棋で一番大切なのは何と云っても形である). An overstatement, I expect, but he seems to have a broader view of shape than other writers.


Yes: even the fuseki is a form of katachi, he claims. But he goes on to say that it is probably not an exaggeration!

The example of butting in at the head of two stones is especially interesting because it was very shortly after that that an amateur wrote in to Kido to ask why it was so bad. There were orientations where it seemed good, according to a group of fellow amateurs studying josekis. The then editor Hasegawa Akira was about to draft a dismissive "as any fule kno" reply, but something gave him pause, and eventually it occurred to him that it might not be so daft after all. And so the Avalanche Joseki was born, and Hasegawa introduced it to the world it in 1927. That, at least is his and the generally accepted account. However, I have discovered a game from January 1925 (two months before Iwasa's preface) in which Mukai Kazuo played the Large Avalanche (the actual name is from much later).

Either way, it does seem that a group of amateurs did make a major contribution in exposing a blind spot in the pros' thinking, much as AI has done in the fast couple of years.

Quote:
If :w24: is at 25, Black does not get that fifth dame, and in this case that makes a big difference. In Elf's mainline variation Black sacrifices the corner stones. A full explanation would probably get into the complexities of the possible fight scenarios. I think the fact that :w24: gives Black an extra dame is the main point. :)


The dame argument is essentially the one I came up with, too. But it seems Shuei took a quite different approach. According to him the error showed up on move 33, and the error was an order of moves one rather than a shape one. If Black had played 33 at A18 first, he could have ensured life on the side come what may (because of a subsequent forcing move), and that would have made his eventual move at 33 much stronger. That, I think, shows Shuei's skill (as with many top people in other fields) in extracting simplicity (half an eye) out of complexity.

But he did not ignore the complexity, and it may be even more dense than we all suspect. At any rate, it did not show up until White 60, which ought to have been at 61. White 61 would have created a double ko, which meant that he could not lose the capturing race. But White's mistake left an implicit triple ko!

uberdude (not really anonymous):
Quote:
I am suspicious of white 28 in the game
[/quote]

Shuei unfortunately does not comment on this move.

Top
 Profile  
 
Offline
 Post subject: Re: Shape problem - pro level
Post #9 Posted: Mon Nov 18, 2019 2:21 pm 
Gosei

Posts: 1536
Location: Earth
Liked others: 571
Was liked: 248
Thank you John, for your answer on my thought.

Top
 Profile  
 
Offline
 Post subject: Re: Shape problem - pro level
Post #10 Posted: Mon Nov 18, 2019 4:05 pm 
Honinbo

Posts: 9514
Liked others: 3001
Was liked: 3166
John Fairbairn wrote:
The example of butting in at the head of two stones is especially interesting because it was very shortly after that that an amateur wrote in to Kido to ask why it was so bad. There were orientations where it seemed good, according to a group of fellow amateurs studying josekis. The then editor Hasegawa Akira was about to draft a dismissive "as any fule kno" reply, but something gave him pause, and eventually it occurred to him that it might not be so daft after all. And so the Avalanche Joseki was born, and Hasegawa introduced it to the world it in 1927. That, at least is his and the generally accepted account. However, I have discovered a game from January 1925 (two months before Iwasa's preface) in which Mukai Kazuo played the Large Avalanche (the actual name is from much later).

Either way, it does seem that a group of amateurs did make a major contribution in exposing a blind spot in the pros' thinking, much as AI has done in the fast couple of years.

Quote:
If :w24: is at 25, Black does not get that fifth dame, and in this case that makes a big difference. In Elf's mainline variation Black sacrifices the corner stones. A full explanation would probably get into the complexities of the possible fight scenarios. I think the fact that :w24: gives Black an extra dame is the main point. :)


The dame argument is essentially the one I came up with, too. But it seems Shuei took a quite different approach. According to him the error showed up on move 33, and the error was an order of moves one rather than a shape one. If Black had played 33 at A18 first, he could have ensured life on the side come what may (because of a subsequent forcing move), and that would have made his eventual move at 33 much stronger. That, I think, shows Shuei's skill (as with many top people in other fields) in extracting simplicity (half an eye) out of complexity.


Interesting. :) Elf also regards :b33: as a blunder, losing 31½%! :shock: But Elf's top choice is H-14, which I consider a tesuji, but I don't know what to call it. Peep, maybe.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins
----
Many are cold, but few are frozen.

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 10 posts ] 

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: No registered users 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:  
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group