It is currently Tue Mar 19, 2024 4:15 am

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 25 posts ]  Go to page 1, 2  Next
Author Message
Offline
 Post subject: Applying endgame principles in other phases of the game ?
Post #1 Posted: Mon Oct 19, 2020 7:50 am 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
In the "How evaluate double sente moves ?" I try to explain (see https://lifein19x19.com/viewtopic.php?p=260680#p260680) that AI (sorry to have use the french acronym IA) was probably able to use its knowledged of endgame principles through all phases in the game.
Let me try to explain this with an exemple. For strong players it will look quite obvious but I suspect it could be interesting for weaker amateurs (like me). BTW don't try to correct bad moves but only to take the idea.


Click Here To Show Diagram Code
[go]$$Bc
$$ ----------------------------------------
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . 1 . . . . . 5 . . . . . 2 . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . 3 . . . . . . . . . . . . . 4 . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|[/go]


Years ago I was explained that after the sequence :b1: :w2: :b3: :w4: in this part of the game (I did not show the moves made in the bottom part of the board because it is not the point) the move :b5: becomes locally a very big point and the best one and, to prove this, my teacher showed me the two following diagrams:

Click Here To Show Diagram Code
[go]$$Bc
$$ ----------------------------------------
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . 1 . . . . 5 . 6 . . . . 2 . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . 3 . . . . . . . . . . . . . 4 . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|[/go]

black :b5: is to passive because after white :w6: black has gained nothing

Click Here To Show Diagram Code
[go]$$Bc
$$ ----------------------------------------
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . 1 . 8 . . 6 . 5 . . 7 . 2 . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . 3 . . . . . . . . . . . . . 4 . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|[/go]

black :b5: is here to greedy because after white :w6: :b7: :w8: again black has gained nothing.

Fine isn'it?

Now let's try to analyse these moves with endgame theory tools.

When the board is empty, the temperature is t=14 => when black plays first the resulting score of the game will be t/2 = 7 (komi?).
When black plays :b1: she gains 14 points by this move but it is now white to play and, assuming the temperature remains t=14 then the expected result of the game will be 14 - t/2 = 7. Everything is OK; all moves are correct.
Considering white is a good move worth 14 points we reach again a position at temperature near from 14 and it is the same for black 3 with an expected result of 7.

Now what about the situation after white :w4: ?

Click Here To Show Diagram Code
[go]$$Bc
$$ ----------------------------------------
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . 1 . . . . . a . . . . . 2 . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . 3 . . . . . . . . . . . . . 4 . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|[/go]


The situation above is quite different because now the point "a" appears bigger than usual. Let's assume that the temperature is now t = 17 (just an example for an easier discussion OC)
By taking this point black gains 17 points but the big point has disappeared and the temperature drops to return to t=14. The expected result for black is now 17 - t/2 = 10 points and eventually black wins the game by 10 poinnt instead of 7.

You have to conclude that the white move :w4: should be a mistake isn'it?.

Let's go further. Let's suppose white :w4: but consider white has enough compensation because another big point "b" (say 17 points) has been built somewhere in the bottom of the game.

What is the expected result of the game: black plays "a", white plays the miai point "b", the temperature drop to t=14 and the expected result for the game is 17 - 17 + t/2 = 7.

Here it becomes very interesting

Look at at following endgame

Click Here To Show Diagram Code
[go]$$Bc black to play
$$ -----------------
$$ | . . c a X . . X |
$$ | X X X O O . O O |
$$ | O X . X O . O X |
$$ | O X X X X O O X |
$$ | b X X . X X X X |
$$ | O O X X X X X X |
$$ | . O O O O O O O |
$$ | O . O O O O O O |
$$ -------------------[/go]


The points "a" and "b" are worth 2 points gote and look miai but ... black may begin by playing the move "c" (which is locally a loss) which is sente but change the miai situation in a tedomri one with only the point "b" as remaining point => eventualyy black gains one point.

What about such strategy in our fuseki?

Click Here To Show Diagram Code
[go]$$Bc
$$ ----------------------------------------
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . 1 . . . . 5 . 6 . . . . 2 . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . 3 . . . . . . . . . . . . . 4 . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|[/go]


Assume black is able to force the exchange :b5: :w6:
Locally it is a loss but here the point: the initial position was a miai situation (a big 17 points gote point "a" in the upper part of the board against a big 17 points gote point "a" in the bottom part of the board) but after sequence :b5: :w6: black keeps sente and the situation has become a tedomari one with only a big point "b" in the bottom part of the board.

As a consequence after the exchange :b5: :w6: followed by black "b" the expected score for the game becomes 0 + 17 - t/2 = 10 and blacks gains the game by 10 points instead of 7.

The conclusion is obvious but can lead to very strange moves:
1) in a tedomari situation you play simply the best gote point (OC after some urgent kikashi moves)
2) in a miai situation try to be creative: look for a sente sequence aiming at lowering the temperature of one big point (may be at the cost of a small local loss) and change the situation to a tedomari one.

That what tell us endgame principles!

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #2 Posted: Mon Oct 19, 2020 8:17 am 
Honinbo

Posts: 10905
Liked others: 3651
Was liked: 3374
Just a note on this position.

Click Here To Show Diagram Code
[go]$$Bc 3d line attachment
$$ ----------------------------------------
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . 8 . . .|
$$ | . . . 1 . . . . . . . . . 7 5 2 . . .|
$$ | . . . , . . . . . 9 . . . . 6 , . . .|
$$ | . . 3 . . . . . . . . . . . . . 4 . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|[/go]


I think that more than one of today's top bots will prefer the 3d line side attachment to the simple extension.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #3 Posted: Mon Oct 19, 2020 8:33 am 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
Bill Spight wrote:
Just a note on this position.

Click Here To Show Diagram Code
[go]$$Bc 3d line attachment
$$ ----------------------------------------
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . 8 . . .|
$$ | . . . 1 . . . . . . . . . 7 5 2 . . .|
$$ | . . . , . . . . . 9 . . . . 6 , . . .|
$$ | . . 3 . . . . . . . . . . . . . 4 . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|[/go]


I think that more than one of today's top bots will prefer the 3d line side attachment to the simple extension.


Surely you are right Bill.
Because it is a gote sequence it seems good for a tedomari situation. In my example this sequence may happen after a sente sequence in the the bottom of the board, intending to lower the temperature in this other big area.

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #4 Posted: Mon Oct 19, 2020 9:12 am 
Judan

Posts: 6082
Liked others: 0
Was liked: 786
Gérard TAILLE wrote:
Click Here To Show Diagram Code
[go]$$Bc
$$ ----------------------------------------
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . 1 . . . . 5 . 6 . . . . 2 . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . 3 . . . . . . . . . . . . . 4 . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|[/go]

black :b5: is to passive because after white :w6: black has gained nothing

Click Here To Show Diagram Code
[go]$$Bc
$$ ----------------------------------------
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . 1 . 8 . . 6 . 5 . . 7 . 2 . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . 3 . . . . . . . . . . . . . 4 . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|[/go]

black :b5: is here to greedy because after white :w6: :b7: :w8: again black has gained nothing.


I don't want to discourage your investigation but you also must consider the possibility that either position is not quiet yet. Maybe Black continuing a bit can actually get a sequence with which he gains?

Quote:
Click Here To Show Diagram Code
[go]$$Bc
$$ ----------------------------------------
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . 1 . . . . 5 . 6 . . . . 2 . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . 3 . . . . . . . . . . . . . 4 . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|[/go]


Assume black is able to force the exchange :b5: :w6:
Locally it is a loss but here the point: the initial position was a miai situation (a big 17 points gote point "a" in the upper part of the board against a big 17 points gote point "a" in the bottom part of the board) but after sequence :b5: :w6: black keeps sente and the situation has become a tedomari one with only a big point "b" in the bottom part of the board.


If there is that bottom tedomari, you must also study either player's earlier tenuki there.

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #5 Posted: Mon Oct 19, 2020 9:19 am 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
[quote="RobertJasiek"]

I don't want to discourage your investigation but you also must consider the possibility that either position is not quiet yet. Maybe Black continuing a bit can actually get a sequence with which he gains?

[quote]

Of course she will gain by continuing but by 7 points instead of 10. That is the point.

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #6 Posted: Mon Oct 19, 2020 9:36 am 
Judan

Posts: 6082
Liked others: 0
Was liked: 786
That is what you hope but is it correct?

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #7 Posted: Mon Oct 19, 2020 9:52 am 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
RobertJasiek wrote:
That is what you hope but is it correct?


It is the point Robert and it is up to you, as go player, to decide.

Click Here To Show Diagram Code
[go]$$Bc
$$ ----------------------------------------
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . 1 . . . . 5 . 6 . . . . 2 . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . 3 . . . . . . . . . . . . . 4 . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|[/go]

If you consider the temperature after :w4: is higher than after :w6: it is true otherwise it is false.
My feeling is that is true but really it doesn't matter for this specific position. My message was only to shwo how endgame principles can be used in other phases of the game.
If my example is not convincing I am sure you can build another one can't you?

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #8 Posted: Mon Oct 19, 2020 10:01 am 
Lives with ko

Posts: 248
Liked others: 23
Was liked: 148
Rank: DGS 2 kyu
Universal go server handle: Polama
Here's an example of this sort of thing from a Michael Redmond Alphago commentary I managed to hunt down. He can do it more justice (https://www.youtube.com/watch?v=-NT0PJ3dTcA)
but here's my amateur retelling. White to play:

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


Michael commented that two areas look clearly bigger than the rest (a and b):

Click Here To Show Diagram Code
[go]$$Wc
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . d . . . a . . . . . . . |
$$ | . . . , . O . . . , . . . X . , X . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . X . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . O . . . |
$$ | . . b . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . c . . . . . . . . . . . . O O . . |
$$ | . . . . . . . . . . . . . . . O X . . |
$$ | . . . O . . . . . , . . . . . O X . . |
$$ | . . . . . . . . . . . . . X . X X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]


We can see that they both extend from a white group, and undercut black's position. But neither is clearly sente: whichever white takes, black can extend away from the other group at c or d. Michael's opinion was that a is a bit bigger than b, so white can take that move and be happy, black will end up taking c and the game continues.


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


Instead, white plays at 1, which is smaller than either of them. And kind of reminds me of your example: b is now not just an extension from a 4-4 stone, but an extension from a shimari (and black can't approach at 1 to sort of over-concentrate white). So if a was bigger than b, white has now made b bigger. It was white's initiative to choose which side he wanted, he lets black choose. But he still gets one of them, and he got the smaller, but still big move at 1.

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #9 Posted: Tue Oct 20, 2020 7:35 am 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
Polama wrote:
Here's an example of this sort of thing from a Michael Redmond Alphago commentary I managed to hunt down. He can do it more justice (https://www.youtube.com/watch?v=-NT0PJ3dTcA)
but here's my amateur retelling. White to play:

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


Michael commented that two areas look clearly bigger than the rest (a and b):

Click Here To Show Diagram Code
[go]$$Wc
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . d . . . a . . . . . . . |
$$ | . . . , . O . . . , . . . X . , X . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . X . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . O . . . |
$$ | . . b . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . c . . . . . . . . . . . . O O . . |
$$ | . . . . . . . . . . . . . . . O X . . |
$$ | . . . O . . . . . , . . . . . O X . . |
$$ | . . . . . . . . . . . . . X . X X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]


We can see that they both extend from a white group, and undercut black's position. But neither is clearly sente: whichever white takes, black can extend away from the other group at c or d. Michael's opinion was that a is a bit bigger than b, so white can take that move and be happy, black will end up taking c and the game continues.


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


Instead, white plays at 1, which is smaller than either of them. And kind of reminds me of your example: b is now not just an extension from a 4-4 stone, but an extension from a shimari (and black can't approach at 1 to sort of over-concentrate white). So if a was bigger than b, white has now made b bigger. It was white's initiative to choose which side he wanted, he lets black choose. But he still gets one of them, and he got the smaller, but still big move at 1.


It's a very good example Paloma.
Michael Redmond commentaries are as usual very interesting but you have to remember he is 9 dans and as a consequence when he looks at a move like white "a" or white "b" he is able to visualize very quickly a local answer sequence for black as well as a local follow-up by white in case of a black choose tenuki. The point is that it is really impossible for him to explain all what he is able to take into account before chosing a move. The choice made by Michael Redmond to draw our attention to the value of an area which can change by each move near this area is judicious for amateurs like us who may easily miss such very important point.
In other words Michael Redmond commentary are correct OC but I am quite sure other important point has to be taken also into account.

Let me try an other view:

Click Here To Show Diagram Code
[go]$$Wc
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . g . . . a . . . . . . . |
$$ | . . . , . O . . . , . . . X . , X . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . d . . |
$$ | . . . . . . . . . . . . . . . . f . . |
$$ | . . . X . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . O . . . |
$$ | . . b . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . e . . . . . . . . . . . . O O . . |
$$ | . . . . . . . . . . . . . . . O X . . |
$$ | . . . O . . . . . , . . . . . O X . . |
$$ | . . . . . c . . . . . . . X . X X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

For white point of view I can easily see the four possilities a, b, c and d

White "a" looks like a very good move being a good extension from the from two white stone in the upper left corner and aiming at some weakness in the upper right black corner. Locally a white move at "a" greatly increase the value of point f and we can almost say that white "a" black "f" is a sente sequence for white.

White "b" is also a good move but the extension from the white stone on 4-4 point in the bottom left corner is not ideal because this 4-4 white stone being alone may be easily be a target for black. As you mentionned Paloma, if you add a white stone at "c" the white move at "b" becomes now a very good move.

What about a shimari by white in the bottom left corner. You have the choice between white "b" or white "e". What is best? No hesitation if you play white "e" you put the influence on the bottom side of the board where the black stones in the bottom right corner are very strong. White "e" is just a losing move isn't it? On contrary white "c" put the influence on the left side where you can clearly see a good field for a future fight. My conclusion is that white"c" can be consider almost sente move implying a black answer at "e".

The fourth move white "d" will increase the value of black "g" and my feeling is that black "g" works better than white "g" (just a feeling of course).

How to choose between the two almost sente moves white "a" and white "c".
If black answers white move as expected the order of move doesn't matter OC. Here is the point when white choose to play "a" or "c" black, instead of answering passively may choose an answer like a reverse sente in the other area.

Let's look at this idea:

Click Here To Show Diagram Code
[go]$$Wc
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . 2 . . . a . . . . . . . |
$$ | . . . , . O . . . , . . . X . , X . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . X . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . O . . . |
$$ | . . 3 . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . O O . . |
$$ | . . . . . . . . . . . . . . . O X . . |
$$ | . . . O . . . . . , . . . . . O X . . |
$$ | . . . . . 1 . . . . . . . X . X X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]


Against white :w1: you imagine black :b2: white :w3:
The result seems good for white because the extension :w3: from the white shimari in the bottom left corner is more efficient than the extension :b2: from the black shimari in the upper right because the influence of this shimari is rather towards the right side than the upper side.

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


On the other after :w1: :b2: black kakari threat greatly increase the value of white answer at :w3: or "h" here and black will then be able to answer :w1: by :b4:. The result looks a success for black because now the influence of the two white stone in the bottom left corner is towards the wrong direction.

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


To be a little provocative white :w1: looks like a double sente move which have to be played immediately!
Here the white shimari in bottom left corner is oriented towards a potentiel fight area while the black shimari in the upper right corner is in the wrong direction to fight in the upper area.

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #10 Posted: Tue Oct 20, 2020 9:08 am 
Lives in gote

Posts: 476
Location: Netherlands
Liked others: 270
Was liked: 147
Rank: EGF 3d
Universal go server handle: gennan
Are black F3 and C6 really sente? With white I would be tempted to follow up from M17.

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


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


Just to be sure, I quickly checked with KataGo and all of these options are very close in its evaluation (differing by only about 1% or 0.1 point), though I did not let it run to very high numbers of playouts.

So all of the suggested moves should be fine opening moves for humans (and so is wR13/bR12), but none of them is clearly sente. It's mostly a matter of taste/style if white responds or ignores a black approach in the lower left corner.

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #11 Posted: Tue Oct 20, 2020 9:24 am 
Honinbo

Posts: 10905
Liked others: 3651
Was liked: 3374
gennan wrote:
So all of the suggested moves should be fine opening moves for humans (and so is wR13/bR12), but none of them is clearly sente.


IMO there are few sente in the opening. That's one reason why the bots tenuki so often. :)

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #12 Posted: Tue Oct 20, 2020 9:28 am 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
gennan wrote:
Are black F3 and C6 really sente? With white I would be tempted to follow up from M17.

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


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


Just to be sure, I quickly checked with KataGo and all of these options are very close in its evaluation (differing by only about 1% or 0.1 point), though I did not let it run to very high numbers of playouts.

So all of the suggested moves should be fine opening moves for humans (and so is wR13/bR12), but none of them is clearly sente. It's mostly a matter of taste/style if white responds or ignores a black approach in the lower left corner.


I agree with you gennan. Surely It's mostly a matter of taste/style. My intention was to show basic endgame principles may be used and help any player to imporve their feeling of the game.

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


Seeing a 9 dan player saying a move at "a" is really a good move may be better than a move at "b" I always try to find some justification hoping in the process to improve my game. Maybe in this specific situation I am completly wrong (I am not a strong player) and a discussion between us can be useful.
How do you understand that a white move at "a" may be better (just slightly better) than a move at "b" as suggested by Michael Redmond?

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #13 Posted: Tue Oct 20, 2020 9:40 am 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
Bill Spight wrote:
gennan wrote:
So all of the suggested moves should be fine opening moves for humans (and so is wR13/bR12), but none of them is clearly sente.


IMO there are few sente in the opening. That's one reason why the bots tenuki so often. :)


Sure Bill.
The definition of a sente move (or a double sente) shall not be a move we have to answer immediatly. You can say that it is a move which increase the temperature of the local area or whatever you want but don't take sente word in its stricly sense simply beause it is not the common understanding of go players who know perfectly that, before answering the so called "sente" move, it might be better to play first tenuki or even some yose miru move in order to find the best answer to the so called sente move.

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #14 Posted: Tue Oct 20, 2020 10:42 am 
Lives in gote

Posts: 476
Location: Netherlands
Liked others: 270
Was liked: 147
Rank: EGF 3d
Universal go server handle: gennan
I didn't mean that white will play some forcing sequence to respond to black's approach later. I meant that white can truely ignore black's 1st approach, accepting that black will play a double approach.

For example, this is a perfectly playable continuation for both:

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

So I agree with Bill. AI taught us that opening moves tend to be less sente than we thought. Reponding or not is an even choice in many cases.

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #15 Posted: Tue Oct 20, 2020 11:02 am 
Judan

Posts: 6082
Liked others: 0
Was liked: 786
Gérard TAILLE wrote:
The definition of a sente move (or a double sente) shall not be a move we have to answer immediatly. You can say that it is a move which increase the temperature of the local area or whatever you want but don't take sente word in its stricly sense simply beause it is not the common understanding of go players who know perfectly that, before answering the so called "sente" move, it might be better to play first tenuki or even some yose miru move in order to find the best answer to the so called sente move.


1) There need not only be one definition of a sente move, but there can be, e.g., different definitions for "considered only locally" versus "considered globally".

2) For what purpose do you suggest "shall not be a move we have to answer immediately"? When considering a global context, fine (e.g., because the first local move might be played as a ko threat for a ko elsewhere). However, your suggestion does not specify whether you consider locally or globally. For a local consideration, you would need to explain why "shall not" because "shall be a move we have to answer immediately" is the supposed nature of sente.

3) There is hardly "the common understanding (or: sense) of go players". Weak justifications like this do not bring us anywhere.

4) "before answering the so called "sente" move, it might be better to play first tenuki": Uhm, no. There are study purposes for which it makes sense to study only locally. E.g., first study each local endgame locally, characterise it by move value, gains and count, and only then use such local information for easier global decision-making. In the local study stage, there is no such thing as a tenuki or test move elsewhere.

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #16 Posted: Tue Oct 20, 2020 12:08 pm 
Lives in gote

Posts: 476
Location: Netherlands
Liked others: 270
Was liked: 147
Rank: EGF 3d
Universal go server handle: gennan
Gérard TAILLE wrote:
How do you understand that a white move at "a" may be better (just slightly better) than a move at "b" as suggested by Michael Redmond?

Click Here To Show Diagram Code
[go]$$Wc
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . e . . . b . . . . . . . |
$$ | . . . , . O . . . , . . . X . . X . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . c . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . X . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . O . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . d . . . . . . . . . . . . O O . . |
$$ | . . . . . . . . . . . . . . . O X . . |
$$ | . . . O . . . . . , . . . . . O X . . |
$$ | . . . . . a . . . . . . . X . X X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]


KataGo shows only tiny differences between a to d. Overall it slightly prefers b and c (those seem to be almost identical in value: "blue" is switching between those all the time).
If I let it run for a longer time, this seems to be the order of its preference: c, b, a, d.

With Michael's proposal of white a, KataGo would continue with black e, white b. So KataGo doesn't really agree with Michael's reasoning. I think Michael's idea is interesting, but I'm not sure if white a is really better than b.

I am ofcourse much weaker than Michael (and probably not much stronger than you). I would probably not come up with his idea in my own game. And even if I did, I don't know if I could really convince myself that a is definitely better than b. I feel like a is a bit fancy.
If this were my own game, I would just play b without much thought, because it looks normal and good to me.
Even now that I've seen Michael's idea, I would probably only play a when facing a calm 5d+ player (playing something out of the ordinary to hopefully increase my chances).

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #17 Posted: Tue Oct 20, 2020 11:36 pm 
Lives in gote

Posts: 476
Location: Netherlands
Liked others: 270
Was liked: 147
Rank: EGF 3d
Universal go server handle: gennan
BTW, if white plays b(M17), KataGo foresees this continuation on my machine, which sort of summarizes everything said above with an optimized move order (and Michael's idea is even in there, but with a slightly different timing):
Click Here To Show Diagram Code
[go]$$Wc
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . 9 . . . . |
$$ | . . . O . . . . . . . 1 . . . 7 8 . . |
$$ | . . . , . O . . . , . . . X . 5 X . . |
$$ | . . X . . . . . . . . . . . . 6 . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . 2 . . |
$$ | . . . X . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . O . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 4 . . . . . . . . . . . . O O . . |
$$ | . . . . . . . . . . . . . . . O X . . |
$$ | . . . O . . . . . , . . . . . O X . . |
$$ | . . . . . 3 . . . . . . . X . X X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

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

KataGo likes the efficiency of black R12 compared to white F3, so it feels that it's not quite good enough for white (but this sharp perception might be a bit beyond unaided human sensitivity). For that reason, it slightly prefers white c(R13) over b(M17) in the original position.

KataGo thinks this should probably happen instead:

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

For both players, this variation seems to follow the principle of preferring mutual damage over responding to your opponent's "sente" moves (puppy go).


This post by gennan was liked by: zermelo
Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #18 Posted: Wed Oct 21, 2020 3:33 am 
Oza
User avatar

Posts: 2404
Location: Ghent, Belgium
Liked others: 359
Was liked: 1019
Rank: KGS 2d OGS 1d Fox 4d
KGS: Artevelde
OGS: Knotwilg
Online playing schedule: UTC 18:00 - 22:00
For me the alphago move was an easy guess thanks to Uberdude's opening gospel based on LeelaZero.

https://senseis.xmp.net/?LeelaZerosOpeningGospel

Before having acquired that kind of thinking, I would definitely have chosen one of the two extensions/undercuts suggested by Redmond and very likely have chosen the one on the left side. I may also have been seduced into playing a defensive extension on the right side.

Also, I would have treated the small knight corner reinforcements (to avoid "enclosure") as equivalent so I would not have thought the lower left corner was urgent.

Correcting (bad/old) habits is not easy but in the opening I seem to have managed changing my thinking and playing during an actual game. I won't go as far as repeating the old amateur adage "I'm strong at the opening" but I have found it way easier to learn about the opening from bots (and players like Ubderdude reviewing bot behavior) than about the middle game, where bots' ability to find life or break through enemy lines is very case specific.

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #19 Posted: Wed Oct 21, 2020 4:54 am 
Oza

Posts: 3644
Liked others: 20
Was liked: 4620
Quote:
For me the alphago move was an easy guess thanks to Uberdude's opening gospel based on LeelaZero.


I think it may be time for St Dude to take up his hammer and chisel again, if he is willing to engrave more tablets. While I'm sure following the old advice will do no harm and will probably produce better openings for amateurs, my strong impression is that the pro world has moved on. We are now seeing lots of ordinary one-space high and low shimaris, lots of 3-3 plays in empty corners, lots of ignoring low approaches and accepting being pressed, more pincers, and more emphasis on the side (Ke Jie's daidaigeima response to the low approach by Mi Yuting in the recent Ticai Cup struck me).

In fact, apart from the still ubiquitous Direct 3-3 I'd be tempted to say pro go is going back to "normal". If that's the case, is it because they are going back to their comfort zones or because their research with AI has progressed? I'd guess the latter. Earlier this month (Oct 2020) the Nihon Ki-in invited Dr Yoshizoe Kazuki (a 4-dan amateur) from the Search and Parallel Processing Unit of the Rikagaku Kenkyusho (a major scientific research institute in Japan) to advise a committee set up specifically to help pros understand AI. Details were obviously not revealed, though I infer that at this stage it was pitched more at the introductory level of what's out there, but they did focus on Leela and katago, and there was a reference to statistics having already been compiled.

Top
 Profile  
 
Offline
 Post subject: Re: Applying endgame principles in other phases of the game
Post #20 Posted: Wed Oct 21, 2020 5:27 am 
Oza
User avatar

Posts: 2404
Location: Ghent, Belgium
Liked others: 359
Was liked: 1019
Rank: KGS 2d OGS 1d Fox 4d
KGS: Artevelde
OGS: Knotwilg
Online playing schedule: UTC 18:00 - 22:00
John Fairbairn wrote:
Quote:
For me the alphago move was an easy guess thanks to Uberdude's opening gospel based on LeelaZero.


I think it may be time for St Dude to take up his hammer and chisel again, if he is willing to engrave more tablets.


I'm not 100% sure but I think Dude used the word gospel tongue in cheek, or at least as a catch phrase, which is the way I have adopted it. Unfortunately I don't have the time and energy to follow the pro evolution post-AlphaGo and it's very comforting to see that human experts keep trying to be creative and keep studying the opening for themselves, to try and have an edge over those simply obeying any "gospel".

One thing that has definitely not been satisfactorily explained or understood, is why a 3-3 invasion would be so urgent that it's played almost without hesitation while a 3-3 opening would be so slow that it's hardly ever played. Might be a perversion (self confirming loopholes) by the algorithms. If the pros now move to more 3-3 openings, that seems to indicate AI is indeed not fully onto the truth there.

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 25 posts ]  Go to page 1, 2  Next

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