Unusual case of efficiency

John Fairbairn · #1

Even after a pro explanation I don't entirely understand what's going on in the following position. Enlightenment appreciated.

The sequence 1 to 8 was played by Osawa Ginjiro and Mizutani Nuiji. It is now the earliest instance of this position known to me, but it has been played by top pros such as Rin Kaiho and Segoe Kensaku. Despite that, it looks soooooo ugly and inefficient. I can't really see why it's played. My knee-jerk reaction is to replace Black 4 with 5, then we get White 4, Black 6.

That way Black has got an eye in (a sort of) sente, and White's position doesn't look materially different. In any event, in the extant games the focus of subsequent play is not really about attacking the White group on the left side. The focus is mainly on Black's access to the centre. I have no problems with the focusing but the alternative way to play on the side seems to me to give Black just as much access to the centre - and it has that extra eye.

But maybe I have been looking in the wrong place. In a commentary on the current game, Kato Shin does not criticise the tsukiatari empty triangle at 4 but does raise a query about White 1. He says the original commentary (which I have not seen) says it ought to have been White A. Kato then goes on to explain that White 1 gives Black a play on the side if he wants it, but then the correct Black response is not Black 2 but Black B. Now I can see that this is much nicer for Black on the left side, so in that sense I can concede that Black 5 in response to White 3 must imply a mistake. But the mistake must then surely be at Black 2. However, top pros do play Black 2 willingly which seems to mean they don't think it's a mistake. It may be ugly but it's not a mistake. I suppose I can just see that White's low and thin plays on the second line are not much to worry about, and I can easily accept that ugliness can be efficient (guzumi).

But the bit I can't get might head around is still why Black 4 and not Black 5.

Gomoto · #2

Without knowledge of the the whole board. I would say it is just a mistake.

To support this statement I checked with Leela Zero:
Black moves at
a 50% winrate white
b 65% winrate white

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

Bill Spight · #3

Well, here is an instance in a top level game.

Click Here To Show Diagram Code: [go]$$Bcm19 Iwamoto (W) - Takagawa (B), 1976 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . . 1 . . . . . . . . . . . . | $$ | . X X O O . . . X . . . . O . . . . . | $$ | . . . X X O O . . , . . . . . X . . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X . X . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . O . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

sorin · #4

John Fairbairn wrote:

Even after a pro explanation I don't entirely understand what's going on in the following position. Enlightenment appreciated.

The sequence 1 to 8 was played by Osawa Ginjiro and Mizutani Nuiji. It is now the earliest instance of this position known to me, but it has been played by top pros such as Rin Kaiho and Segoe Kensaku. Despite that, it looks soooooo ugly and inefficient. I can't really see why it's played. My knee-jerk reaction is to replace Black 4 with 5, then we get White 4, Black 6.

That way Black has got an eye in (a sort of) sente, and White's position doesn't look materially different. In any event, in the extant games the focus of subsequent play is not really about attacking the White group on the left side. The focus is mainly on Black's access to the centre. I have no problems with the focusing but the alternative way to play on the side seems to me to give Black just as much access to the centre - and it has that extra eye.

But maybe I have been looking in the wrong place. In a commentary on the current game, Kato Shin does not criticise the tsukiatari empty triangle at 4 but does raise a query about White 1. He says the original commentary (which I have not seen) says it ought to have been White A. Kato then goes on to explain that White 1 gives Black a play on the side if he wants it, but then the correct Black response is not Black 2 but Black B. Now I can see that this is much nicer for Black on the left side, so in that sense I can concede that Black 5 in response to White 3 must imply a mistake. But the mistake must then surely be at Black 2. However, top pros do play Black 2 willingly which seems to mean they don't think it's a mistake. It may be ugly but it's not a mistake. I suppose I can just see that White's low and thin plays on the second line are not much to worry about, and I can easily accept that ugliness can be efficient (guzumi).

But the bit I can't get might head around is still why Black 4 and not Black 5.

Thanks for posting this John, very interesting question!

It looks very strange indeed, the choice to not make an eye in sente. Maybe it has to do with the rest of the board?

Otherwise, locally speaking the only reason I can think of for the choice shown here is the fact that if black chooses the common-sense way to play and makes an eye, the 2x2 point in the upper-left (A18) becomes very hot, it is a crucial place for both groups to play to fight for eye-space - and white has sente.
So black's "eye in sente" is not free at all as it may seem at first, since after white A18 black will feel compelled to answer (to "defend" the choice to make an eye in the first place) which in turn gives white eye-space in sente for their group in the top-left.

Compared to the sequence shown in your diagram, where the 2x2 point in the upper-left is just "meh": white playing there would not make black feel obliged to respond at all.

This is obviously an "after the fact" analysis and it may be wrong/over-simplified.
Myself I would also just play the usual way and make an eye in an actual game.

jeromie · #5

Gomoto, is that kind is swing in win rate normal? That’s huge!

Whether it is “correct” or not, the pros obviously had a good reason for playing the move. Is it because white can get forcing moves on the outside that force black to make a group with fewer liberties and less accesss to the center?
I could see something like this happening:

The white group on the top is in danger, but it’s still annoying. Perhaps black would jump out one space above to prevent this, but that does give less access to the center. What kind of follow up do stronger players see if move 4 is different?

Kirby · #6

I don't know why 'b' vs. 'a', but my guess is to prevent White from getting good shape(?):

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

That being said, black doesn't have great shape either in the other variation.

Gomoto · #7

I checked with Bill's game position. Zen and Leela Zero agree that the disputed move is a mistake.

Why play pros moves like this?

For the same reason they did not play early 3-3 invasions.

Sometimes a beginners intuition (Do you remember your first thoughts about the 3-3 invasion) beats the established expert opinion. But the beginner will never be able to show he is right. :twisted:

(Until 20 years later and the advent of strong AI :lol:

)

Bill Spight · #8

Gomoto wrote:

I checked with Bill's game position. Zen and Leela Zero agree that the disputed move is a mistake.

Thanks.

Similar difference of 15% or so?

jeromie · #9

Gomoto wrote:

I checked with Bill's game position. Zen and Leela Zero agree that the disputed move is a mistake.

Why play pros moves like this?

For the same reason they did not play early 3-3 invasions.

Sometimes a beginners intuition (Do you remember your first thoughts about the 3-3 invasion) beats the established expert opinion. But the beginner will never be able to show he is right. :twisted:

I think there are two interesting questions here: was the move a mistake (Leela may help with this) and why did they play the move (Leela is useless here). We can learn from both answers, so even if we decide the move was a mistake it’s still worth trying to understand the rationale for the choice.

Bill Spight · **#10**

For reference, here is the Iwamoto vs. Takagawa game.

Gomoto · **#11**

@Bill

Zen: Move F18 white 53% vs. Move G17 43%
LeelaZ: Move F18 white 53% vs. Move G17 37%

(winrates for white)

dfan · **#12**

Is there any possibility that either of these cases was a recording error and the moves actually occurred in a different order that makes more sense?

ez4u · **#13**

Look back at the recommendation from Kato Shin that White should jump and Black should jump down.

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

Now compare this to the normal answer to the slide by White below. White has gotten in a forcing move, preventing Black's preferred reply.

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

In the game White does not want to jump out because of the weak point left behind below. Black 8 threatens to cut White's jump. Instead of jumping, White settled for the slower diagonal play.

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

What the programs are telling us is that perhaps the pros are over-thinking here and that being forced would be better than the bad shape.

John Fairbairn · **#14**

Quote:

Now compare this to the normal answer to the slide by White below. White has gotten in a forcing move, preventing Black's preferred reply.

I think your answer (and ditto Kirby's) is probably on the right lines, though the devil may be in detail, but first I need, apparently, to correct a misapprehension that I certainly also had until a day ago. The "normal" answer turn out not to be so. At least out of 18 cases in the GoGoD database the split between tsukiatari and kosumitsuke is 9:9. Furthermore, in his commentary on what was the first instance of this pattern until this new game (Shuei vs Sanei), Shusai made no comment on the tsukiatari, so there's another big name that wasn't troubled by it.

But I've finished transcribing the game now, and there was a significant clue in the later commentary. It seems that if Black plays the kosumitsuke he would allow White to jump to E11. Because he didn't want that he chose the tsukiatari because then if White jumps to E11 Black has an ideal peep at D10. This seems game-specific (and that is probably true of all the examples - we amateurs perhaps do tend to play josekis in a vacuum). In other words, the tsukiatari here is a way of making White play 42 instead of E11.

Anyway here is the full game now. It would be interesting to hear how Leela/Zen change their assessments in the game-specific position. I found it interesting to see how much Black 43 influences the rest of the play. On the one hand a foretaste of such jumps and L shapes in the centre that Shuei perfected, and on the other hand a hint at the centre style exemplified by the AIs. To put that in quite a different way, the Black group in the upper left doesn't really need a base - its main (sole?) function is to support Black 43 in the centre. We can perhaps see it as a very deep shoulder hit

jeromie · **#15**

That makes perfect sense, Dave.

This whole thread has been very instructive. “Being forced is better than bad shape” almost has the sound of a modern Go proverb, and Leela’s analysis emphasizes the true cost of a large group without eye shape.

mitsun · **#16**

ez4u wrote:

Now compare this to the normal answer to the slide by White below. White has gotten in a forcing move, preventing Black's preferred reply.

Don't forget that it costs W a full extra move to get this better result in the corner. In other words, for your second diagram to be a fair comparison, B should be allowed to place an additional stone anywhere on the board. This seems more than enough compensation for allowing W his forcing moves in the corner.

Kirby · **#17**

Btw, how strong is Leela, now?

Is it strong enough that we can use its analysis to call pro moves mistakes, simply from assessed evaluation percentages?

I didn't think so, but I don't know much about Leela.

Bill Spight · **#18**

John Fairbairn wrote:

I've finished transcribing the game now, and there was a significant clue in the later commentary. It seems that if Black plays the kosumitsuke he would allow White to jump to E11. Because he didn't want that he chose the tsukiatari because then if White jumps to E11 Black has an ideal peep at D10. This seems game-specific (and that is probably true of all the examples - we amateurs perhaps do tend to play josekis in a vacuum). In other words, the tsukiatari here is a way of making White play 42 instead of E11.

Anyway here is the full game now. It would be interesting to hear how Leela/Zen change their assessments in the game-specific position. I found it interesting to see how much Black 43 influences the rest of the play.

Yes, the D-08 stone is significant, isn't it? As well as the Black moyo on top, with thickness in the top right. :b43:

looks big to me.

It will be interesting to get the bots' assessments in this game. Even though I suppose that it is a no komi game, something else that gives Black leeway.

Bill Spight · **#19**

Kirby wrote:

Btw, how strong is Leela, now?

Is it strong enough that we can use its analysis to call pro moves mistakes, simply from assessed evaluation percentages?

I didn't think so, but I don't know much about Leela.

IMO, the Zero bots are probably strong enough to call pro moves mistakes, based upon self play, but not upon evaluations alone. Even the top chess engines are not that good.

Kirby · **#20**

When Leela gives a win rate for black (e.g. 59%), is that self play or just evaluation?

Unusual case of efficiency

Who is online