Attack & defense by Ishida and Davies reviewed with KataGo 3

Knotwilg · #1

We still need to investigate more examples but here we see that the typical "attack to make profit" by capping a floating group, backed up by influence to make central territory, is conceivable but not clearly better than other options.

We also continue seeing moves recommended which maximize efficiency and prevent being forced into an inefficient configuration. The concept of "efficiency" still needs some more development (at least for me) as it can easily become tautologic.

Uberdude · #2

My interpretation of this is it's another example of bots hating the opponent getting to make profitable exchanges in sente; not that humans like it, but bots 1) hate it more 2) see and evaluate the sente exchanges more accurately (higher priority than the not-so-weak group on at lower right). Here it's the e14 d14 exchange being nice for black, so white's move is a valuable reverse sente to prevent that. I would imagine bots would think both white and black have been making mistakes for earlier moves in not playing in that area. If white had managed to exchange f15 for g16 (obedient black) earlier, then perhaps that top left area wouldn't be so hot? Also interesting that KG is happy to play f14 gote, rather than f15 trying to grow that boundary in sente which would be my idea, trying to overconcentrate black on the top side (but probably a dream).

P.S. I suppose the exchanges KG wants on the lower left are ensuring black b11 won't be a good move, which once you play f14 you are committing to making the left territory so could pose an annoying hane inside or outside question later.

Bill Spight · #3

FWIW, my eye was drawn to C, first because of White's thinness in that area, and second because of the thickness of White's wall in the bottom left, which C can work with. It's too late in the game to worry about using thickness to make territory.

As for attacking Black's group in the bottom right, I would much rather have done so before Black played at O-07. After topping the tree with O-09, White's sphere of influence is still thin.

B follows Znosko-Borovsky's advice to generate weaknesses in the opponent's position.

And I guess C is in line with John Fairbairn's idea of controlling a major diagonal.

Knotwilg · #4

Bill Spight wrote:

As for attacking Black's group in the bottom right, I would much rather have done so before Black played at O-07. After topping the tree with O-09, White's sphere of influence is still thin.

Apologies for misleading you if you thought O-7 was the last move before the position. Maybe it was but I only have the end position. I should manipulate it into White's sente differently. Like just posting a diagram instead of an sgf (but I like the variation capability of it)

Otherwise, all aboard.

Uberdude · #5

If you accept KG's judgement that:
1) o9 cap can be tenukid because r7 life is not white profiting much in sente
2) black e14 is sente, as no worry of interposed attack of 1)
then it comes down to would white rather have:

o9, and black got e14 exchange for free

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

or f14, seeing as black takes j2 in both lines and it's then white's move.

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

My guess is the authors would not have thought black would j2 but jump around o9. This doesn't make any points for black, but reinforces his group (but KG thinks it doesn't need it) and makes white's plans of using the marked stones to make some centre territory evaporate. Also I'd start to wonder about black shenanigans against the double 2 space extension on the right, so white o9 has some prophylaxis against that too. Would KG then continue taking real profit around the sides with 3?

Click Here To Show Diagram Code: [go]$$Wc Authors' fear? $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X . X . . . . . . . . . . . . . . | $$ | . . O X . . . . . . . . X . X . . . . | $$ | . . O , . X . . . X X . . . . , X . . | $$ | . . . O . . . . . . W . W . W . . . . | $$ | . . . . . 1 . . . . . . . . . . O . . | $$ | . . . . . . . . . . . . . S S . . . . | $$ | . . . . . . . . . . . . . S S . b . . | $$ | . . O . . . . . . . . . . S S . O . . | $$ | . . . O . . . . . , . . . . . , a . . | $$ | . . X O . . . . . . . . . 2 . . . . . | $$ | . . X O . . . . . . . . . . . . O . . | $$ | . . X O . . . . . . . . . X . . . . . | $$ | . . . O . . . . . . . . . . . . . . . | $$ | . . X X . . . X X . . . . X . X X . . | $$ | . . . , . . . X O O . . O O X , O . . | $$ | . . . X . . . 3 . . . . . . O O . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Bill Spight · #6

Knotwilg wrote:

Bill Spight wrote:

As for attacking Black's group in the bottom right, I would much rather have done so before Black played at O-07. After topping the tree with O-09, White's sphere of influence is still thin.

Apologies for misleading you if you thought O-7 was the last move before the position. Maybe it was but I only have the end position.

No, I didn't think that.

I still don't know how that Black shape arose, however. :scratch:

Quote:

I should manipulate it into White's sente differently. Like just posting a diagram instead of an sgf (but I like the variation capability of it)

You have found quite a nice way of presenting a problem without showing possible plays or comments. :salute:

Attack & defense by Ishida and Davies reviewed with KataGo 3

Who is online