If I had time, I would turn this into a series. Now I'm not so sure ...

Very interesting game!

I don't understand move 132. Why does White think it is important to save his two stones at this stage of the game? How big is that move?

Elf thinks that it loses 18½%, by comparison with Q-02.

I think that the main point of White 132 is the kikashi against Black's long center string. White can partially seal off prospective territory in the bottom right.

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

My two main guides in life:
My mother and my wife.

Everything with love. Stay safe.

Probably because it keeps m9 sente, which makes m8 a gote instead of sente cut.

Not saying that's a good enough reason to play it now, but it is a plus that needs to be considered as well as the yose value.

Thanks Knotwilg, this was a fascinating game, and I liked the comments. Yes please, this would be a great start to a new series (-:

Was there a reason for looking at this particular game, or is it a more or less random choice?

I have a bunch of other questions about the moves.

At , to me it seems more natural to probe at O17 first, then decide whether or not to invade. KataGo isn't interested in either move, and evalutes at L16 as a small mistake (drops 1% winrate). But when I put O17 on the board, KataGo thinks it's almost exactly as good as its preferred move at P10 (-0.2%). And in response to O17 on move 24, it wants black to hane on top. But when O17 is played in the game two moves later, KataGo initially prefers P16 in response, but then analyses the game move as +1%.

At , in this type of position I have trouble choosing exactly where to leap out. So does KataGo! It looks at a bunch of moves and finds them pretty much equal, including anything on the straight line between L13 and Q13.

I'm puzzled by the timing of the various exchanges between moves 67 and 80. KataGo doesn't seem to shed any light on this. The moves , , , , -- seems to be a logical sequence, but they get interrupted a few times. Why, after exchanging S12 for S13, does white decide this is the right time to settle the top left? Why does the R3-R4 exchange inspire black to fix up the L17 area at that moment? Why do both players seem to agree that H3 is a forcing move if played at move 80? Throughout this patch, KataGo seems to advocate finishing what you started instead of jumping around so much!

Moves 101-105, it seems odd to me to drop black 101 into the middle of white's area and then abandon it. KataGo kind of agrees, suggesting black 105 at M3, and then black has various options for trading off those two stones against threats in the corner (the evaluations keep changing as I let it run).

At move 131, black having come away from the left side pushing battle with sente, I thought black was going to come back to the bottom and help out the lonely K3 stone at this point. Here, KataGo suggests black 131 at N2 (or the same point for black 133).

I thought black 143 looked like a lovely tesuji, wringing as much as possible out of the situation, postponing the connection at E16 for as long as he can get away with. What a shame KataGo doesn't agree!

I was lucky. Like our kyulearner, I have been using Otake as a role model and now wanted to combine studying his games with usage of Katago.



What I learnt from it is that both players attached more importance to the fuzzy top while Kata found the lower side to be bigger.



This was the third invasion of that kind in this game, a one space check if you wish. It may have been "en vogue". In this case there's a relationship with the corner, with N2 as key point. I assume Ishida knew that.



I'm still struggling with the way the left side ko came to life and how little Kata's estimations changed when it did.

After Black 143, White pushes two times into Black's stones, activating the ko potential because of shortage of liberties. White 148 is "honte", Black must connect at 149 and now WHite can start the ko sequence. That being so, Black's earlier moves in this area must have been overplays, surely? And why doesn't Kata jump a major percentage here?

The potted pro commentary at the time was rather different. But note first that the komi given here is strictly wrong. It was "5 points but White wins jigo", so is effectively 5.5. And White is male.

"Rather than White 12, the counter-hane 13[B17] is now more usual.

If, instead of 16, White had played [L17] 73 on the upper side, the game would have been quieter. He was preceded by Black to 17 and then White inevitably had to ward off invasion at A [C9] by defending at 18 and 20.

When Black checks at 21, the fuseki is painful for White.

The invasion at White 24 is natural. 26 and so on is then a common tesuji sequence.

White 36 was a mistake. Nobi at 37 [N17] was correct. If then Black B [R7], White can settle himself with C [P13], Black D [Q16], White E [P15].

Black's hane at 37 has good momentum.

Black 49 is severe and starts a confused fight.

After Back 55 to 65, Black has thicker shape. [i.e. better endgame prospects]

White 78 at 79 was also big but White's strategy was to stress the centre by checking at 80.

Black 101 was a bold invasion. After White 102 and 104 the fighting becomes confused, though White ended up with the better prospects.

But White 194 was a mistake. The atari at 195 was correct."

FWIW, is Elf's top choice, with 111.3k rollouts. No other play gets as many as 1500 rollouts.



Elf's top choice for is the AI/Go Seigen shoulder blow at O-16, which has several plausible replies. If Black blocks at P-16, White makes the two space high extension from the top left corner at J-16 ( ).



Elf thinks that loses 5% to par (a minor error). It's top choice is the 3d line attachment against the top right corner at R-15, and after Black stands at Q-16, White plays the enclosure of the bottom right corner at O-03. Then Black does not invade at C-09, but attacks the top left corner with the kosumi at H-17.



Elf thinks that the fuseki has been painful for White since .



For Elf slightly prefers xela's play, the 3d line attachment at O-17.



Elf thinks that loses 6% to par, another minor error. But it does not care about the top right. Instead, it chooses to crawl twice on the 3d line, and after Black extends to Q-07, jump to R-06.



Elf agrees with KataGo that is a big mistake, losing 21% to par. White should tenuki.



Elf agrees with KataGo on preferring the atari at L-15, and says that loses 21% to par, and is thus another blunder.



John, do you mean thicker overall shape?

is the atari at L-15. Every Black play from to is Elf's top choice, and Elf gains 10½%. This gain is almost entirely due to at K-13, which pushes Black into the center and loses 10% to par. Instead Elf says to simply snag one Black stone with J-15. Then if Black plays at K-13, White jumps out to M-11.



This explanation fits Elf's variations as well.



According to Elf, Black 101 ( ) loses 5% to par, another minor error. Then loses 5½% and loses 5% again.



Elf agrees that was a mistake, losing 17% to par. At that stage of the game, we may perhaps consider it a blunder. Elf also agrees that the atari at would preserve White's advantage, although it has a minimal preference for the throw-in atari at G-19.

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

My two main guides in life:
My mother and my wife.

Everything with love. Stay safe.

Thanks a lot John for the pro comments at the time. I was curious if they considered the same move to be the losing move.

Even the (abbreviated?) pro comment doesn't discuss the ko at the left side, which came about only due to Black's shortage of liberties there, due to reducing in the center. A mystery to me.

And sorry for the inadvertent usage of "she" - an old SL habit which I don't consistently omit when discussing actual (male) players.

Game 2

Highlights

1) Ishida, he, plays parallel 3-3 and later adds a 10-3 to form a strange 3rd line "framework"
2) Otake plays a standard side pattern that loses the advantage, says Kata, then alukewarm defensive boshi, allowing White to turn what seemed to be a splitting attack by Black, in a surrounding tactic
3) Comes a flabbergasting decision by Ishida to live locally and let Black connect ... because he can take profit on the outside !!! Kata prefers maintaining the cut, but appreciates Ishida's originality
4) Then it's Ishida's turn to make a soft move in the middle, giving the initiative to Otake, who shoulder hits the remaining lone 3-3, which isolates the formation Ishida so astutely created
5) on the right side there's a ko, which Otake quickly resolves; Ishida then attacks the whole lower side, which Otake reduces, in order to gain influence and attack the left side. Kata agrees and puts Black's chances to win the game higher than ever
6) eventually, Otake seems to make bad choices in the whole board fighting; when Ishida lives on the left side and resolves the lower side, Otake resigns


This game boils down to a fight. Again, ko plays a big role but it's one that Otake quickly resolves.

In this third game, Otake takes white against Kunihisa Honda. The game’s theme is Black’s left side moyo, which Honda develops early on, leaving in particular the bottom right corner for grabs. When White gets a chance to firmly take that corner, in exchange for a complex battle in the centre and right side, which would put him ahead according to Kata, Otake prefers a strong position in the centre, giving the corner back to Black. One surprising moment in the game is when White is allowed to make a ponnuki in the centre, which Kata seems to agree is not so problematic for Black, as he uses the time to solidify the right side.

Kata prefers Black’s territories throughout the remainder of the game until Honda apparently blunders in a fight involving his group that made up his original moyo, and resigns.

This “Kata exploration of a pro game” shows again the flexibility of bots and pros alike. While the bots like secure territory and central complexity Otake tends to prefer central strength and leave some corner complexity. Bots are less likely to blunder in such fights than humans, in this case Honda.

Incidentally, I have had the great honour to play Kunihisa Honda, when he was on a visit in our town with citizens of our jumelage Kanazawa.

Regarding the last comment "Black now has several options to make two eyes", I don't think Honda would have any problems making 2 eyes if he wanted to, but that with 139 kosumi he wanted to kill white (or at least make it a grovelling 2nd line life) but Otake and Kata both know that's an overplay.

Very interesting game, hard to believe that such a huge black moyo became no territory at all.

I was puzzled by the sequence 92-97. Why does Otake play in the corner, then makes the exchange on the side before coming back to the corner with ?

92 can be seen as a probe, asking black how much he wants to kill the p17 stone vs the r18 stone and if he's willing to trade the corner for killing p17. There are several answers:
- s18: this offers white a trade: you can hane at s17 and live in the corner in gote but that means you sacrifice p17 (s17 r17 t18 s19 r19 q18 t19 p18 s16 s15 t16). Or white can r17 and black will block s17 and then white is trying to save r18 and p17 together, often with coming out at p15 and then getting q16 to make some eyeshape
- s17: this denies white the easy gote life in the corner above and says "I want to attack the whole group", but compared to s18 r17 s17 variation white has a little more eyespace as he can push at s18. So this choice is probably used by black when he is fairly sure he CAN kill white, in which case r18 was probably a mistake (or s17 was).
- o16: seal white in and connect black, locally soft so black probably chooses this when he doesn't mind letting white live comfortabl in the corner and just wants to connect his groups (e.g. because n17 was weak) and build some influence.
- other things around the p18 shape point, e.g. if o17 p16 then black p18 hane in sente is key shape point.

94 is then another probe with some choices for black:
- o16: I don't care about l17 (e.g you can jump out at l15) and just want to make life harder for p17. In this case white wouldn't r17 because o16 cut off the escape route and white would do the trade for corner life with s17 (but maybe no something with l17 first). But black's top left group isn't totally safe so he probably doesn't want to go so easy on l17.
- l16 or m16 seek to enclose l17 and keep top left safe, but give p17 more room and maybe white can use l17 to help them.

So when black played s18 he is saying he will let white take the corner in gote. White doesn't want to do that yet, seeing as it would make the top side strong and j16 group safe, so before potentially doing that he sees if he can get some benefit there. Black attaches on top so then white makes r17 exchange because he thinks he can use l17 to settle p17, probably hoping for black to answer m17 at m16 and let him connect up. But black resists that with m18, which actually makes r17 become a bad exchange, because with the way the game went it would have been better for white to not have exchanged that but retain the possibility of corner life with s17. So maybe r17 should have been after m17.

Not to take away from Uberdude's excellent explanation, is a standard play when Black has played a pincer, 4th line kosumi, and extension from the kosumi. It goes back centuries. Elf has a different idea for .

Click Here To Show Diagram Code

[go]$$Wcm92 Elf's mainline variation for
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O . O . . X . . . . . . . . . . . |
$$ | . . X O . O O O X . . . X . O . 3 4 . |
$$ | . . X , . . . . X , . . . . . 7 X . . |
$$ | . . . X X X . X . . . . . 9 5 X 8 . . |
$$ | . . . . O . . . . . . . . . 6 . . . . |
$$ | . . . . . . . X O . . . . . . . . . . |
$$ | . . . O . . O O . O . . . . . . X . . |
$$ | . . . . . . . . O . . . . . . . . . . |
$$ | . . . X . . . X . , . . . O . , 2 . . |
$$ | . . . . . . . X O . O . . . 1 . . . . |
$$ | . . . . . . . . O . O X X O X X X X . |
$$ | . . . . . . X X O . X O O O O X O O . |
$$ | . . . . X X O O X X . X . . X O . . . |
$$ | . X X . X O . . O . . . . X X O O O . |
$$ | . . O , O . O X O , X . . . O X . . . |
$$ | . . . O . . . O . . . . . . O X . O . |
$$ | . . . . . . . . . . . . . X O X . X . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

White takes kikashi on the right side and then plays the jump attachment on the 3-3. After Black replies with a hane, White makes another jump attachment on the 5-5. Rather skillful sabaki. Elf regards this result as 14% better for White than the game sequence through W100. Edit: In the game sequence both sides made mistakes, according to Elf. The 5-5 jump attachment was often a good play for White.

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

My two main guides in life:
My mother and my wife.

Everything with love. Stay safe.

I just want to say thanks Knotwilg for continuing this excellent series. I haven't responded in detail to games 2 and 3 because I've been obsessed with something else recently. But be assured that I'm following this thread with great interest.

Game 4

Otake takes Black here, against Kudo Norio White.

While the opening evolves fairly evenly, both sides taking a small lead at certain points, in the middle game Black takes the advantage, peaking at move 85

At 103, Otake seems to misjudge the strength of his stones at the top and in the middle. He attacks for profit on the right side, but White counterattacks, involving Black's groups at the top, on the left and on the right.

Through 119, Black clings to his stones at the top, which become heavy, while White's stones are light and can be exchanged for profit on the left side.

At 133 Otake decides to live locally on the right side, giving White the momentum to strengthen the centre, so that Black's top or left side will suffer. From that point onwards, there's no chance to come back

While Otake is known to play solidly and perhaps slowly in the early parts of the game, usually the strength he builds pays off in the end. This game is unusual in that respect: at 103 he could have strengthened the top and maintain the advantage.

And game 5, where Fujisawa Shuko takes control of the game fairly linearly until he blunders and resigns (while the game is still close).

Game 2: move 135

Actually, KataGo (and LZ network 258) has to do quite a lot of reading to find that J4 is best: it takes about 6,000 playouts when I try it. If you click "show policy" in Lizzie, you'll see why. The bots think that first, pulling out of atari at E5 is the most instinctive move, so they have to read that for a while before trying other things. (It's actually quite neat to see how white's F6 stones can connect out to safety.) Then H4 and J3 also get moderate policy values (over 10%), and KataGo (but not LZ-258) thinks the game move also looks better than J4 according to the policy network.

White 36 also looks incredibly weird to me. I would play the atari at N15 almost without thinking. It seems that so would KataGo.

Patterns found in these 5 games

It's a small sample for patterns but it gives me a hypothesis to further confirm or disprove.

1. Ko fights

On two occasions, Otake and his opponents at the time did not take an opportunity for ko which Kata recommends. Kata seems generally more adventurous or willing to make the exchange.

game 1 - move 33
game 4 - move 82

2. being forced (to capture what's yours)

When the dust of a fight is about to settle, forcing sequences occur, i.e. forcing the opponent to take what's already his but then at the smallest possible scale. The idea that being forced this way is bad, is not novel, but Kata regularly points it out in these games
game 1 - move 89, 154
game 5 - move 105

3. forcing (to stabilize with territory)

On the other hand, we often see Kata boosting the probabilities of a side which has just been "forced" to live with territory, where the status was previously not clearly settled. It appears to be better not to settle the shape this way and leave the aji.

game 1 - move 91
game 3 - move 66

4. boshi reductions/attacks for influence

while shoulder hits are bots' favorites, boshi reductions or even attacks clearly aren't; exception: when the influence cooperates with influence at the other hand, creating potential territory

game 2 - move 31, 60
game 3 - move 57, 61

5. simplifying the centre vs simplifying the corner

when there's a big fight in which there's a choice to simplify by taking a big corner, leaving the centre insecure, or simplifying the centre, leaving the corner at stake, bots prefer central complexity more than pros do

game 3 - move 54, 69

6. tactical "blunders"

there are situations where one player has a clear lead but then makes a tactical mistake that loses the game

game 3 - 139
game 4 - 133
game 5 - 149