I see reading problems and direction problems…
The AI doesn’t like this move, presumably because it leads to the following sequence which is too comfortable for Black:
Variation 1z: White’s shape is pretty cramped. Black takes the corner and the side. The White group is not completely alive (although in no immediate danger) and might be a burden later in the game.
Instead I would prefer one of these three moves for White:
A is the old recommended variation, which leads to the kite joseki. The AI doesn’t like A and professional players no longer play it, because it is a second-line move and as such, they consider that it is gote. In a pro game, if White played A, Black would probably tenuki instead of playing the kite joseki.
B is the new move that replaces A for the AI and professional players. It’s kinda the same idea as A, but more sente since it’s not second-line and it’s a contact move. C is an extension to the side, which is a wider extension than the one you played in the game; it looks a bit too wide but the idea is that it would become very good if Black played the “kick” / kosumi-tsuke of variation 1z.
Variation 1a: The kite joseki. With this joseki, the White group has a little more space than in variation 1z, which means it will be easier to live, which means your group will be under less pressure during the game, which means Black will probably get less profit by attacking your group.
In pro games, this sequence is no longer played, because Q2 is considered gote, but at your level you can still play it if you want, and as long as the opponent answers with the kite joseki, it’s good. When you start playing against stronger opponents and they start to tenuki instead of playing the kite joseki, you should probably stop playing Q2.
Variation 1b: The new variant that replaces the kite joseki. Move 9 in this sequence could be at A, B or C instead. Also Black 6 could be at 7 instead to go for a different direction.
And finally my third suggested move was the wide extension on the side. The idea is to discourage Black from playing the kosumi-tsuke:
Variation 1c: Kosumi-tsuke after a wide extension. Good for White.
So you should expect Black not to play the kosumi-tsuke if you played the wide extension, and instead Black will probably go for one of these two variations:
Variation 1d: White takes a larger extension on the side.
Variation 1e: Black tries to fight back, which leads to a weirder game.
In the game, you played the small extension of move 8, but Black played a slack answer at move 9, so it’s good for White; and you can still play A or B to get a wider living space:
I strongly dislike move 22. There is a cut at the marked point, so I think you should either protect that cut, for instance with a tiger mouth at A, or tenuki. But adding a move at 22 is just begging for Black to cut, in which case stone 22 is just going to become an extra prisoner for Black.
Also, this stone at 22 looks like it might be removing your own liberties from your stones. Compare these two sequences where Black cuts, with and without the White stone at 22:
Variation 2a: Black cuts; White has 4 liberties.
Variation 2b: Black cuts; White’s move 3 obviously becomes bad, because move 1 is removing one of the 4 liberties compared to variation 2a.
Of course, you might want to sacrifice the White stones rather than to play move 3, but regardless of whether you are going to sacrifice this stone or not, my point is that the only effect of move 22 is that you lost a liberty in a potential fight.
Move 22 is like a reverse “hane at the head of two stones”, where you’re bumping your own head against a hane that’s already there, and that’s usually not good.
In the game, Black first reinforced you by playing the exchanges 23 to 26, and then cut at 27, and then went back to the side at 29, so it looks like Black couldn’t decide what they wanted, to cut or to take points on the side, and this indecision makes this sequence good for you:
Also Black now has lots of cutting points in a row, at the four marked points, so it’s vital for Black to play at A next in order to get a better shape with more liberties.
In the game Black played the hane at 32 instead of playing at A. This is a mistake.
Variation 3. White cuts.
There is no risk of Black escaping at the marked point to save the Black cutstone, because you can capture it in a net / geta. So you could cut immediately. Actually any of the cuts would be good, I think. The exact cutting-point you choose for move 32 is not too important in this case.
But in the game you forced Black to connect! This is what we call a thank-you move. Black had a very bad shape with lots of annoying cutting points, but now it’s all fixed. With Black 33, you can no longer exploit the cutting points. It’s a big thorn off of Black’s side. In general, do not help your opponent fix their cutting points!!
Also, two moves later you have this shape, a torn keima, which is a good hint that White 32 was a not very good move:
This is the game. White 50 hurts my eyes. The empty triangle is not a very efficient move, it doesn’t give a lot of liberties to your stones. The Black cutstones now have a lot of wiggle room and you are going to have to be very careful with your White empty triangle that doesn’t have a lot of liberties, plus the two cutting points at H5 and L4 which can quickly become very annoying.
Instead there was a much simpler solution:
Variation: White captures the cutting-stone in a net.
If you can capture a cutting stone in a net or a ladder, do it. It’s almost always the best option. There is a trick to getting used to read ladders and nets: when the opponent has a cutting stone, always look for a ladder or net to capture it. Even when you first think that there is no ladder or net to capture it, still take some time to try to find a ladder or net to capture it. Cutting stones are hyper important and capturing them in a ladder or net is huge. And there is very often a net. Sometimes it’s a simple geta, sometimes it’s a wider net, sometimes it’s a knight’s move. My advice is: everytime there is a cutstone, try to find a net. A simple net, a wider net, or a net with a knight’s move. Sometimes there is no net, but it doesn’t hurt to look for it anyway. Capturing stones in a net if you can is really super important.
This is the game. If you look at the AI’s winrate curve for these moves you’ll notice very sharp sawteeth, with several White and Black moves at -50 winrate, which is huge. The reason for this is that move 83 is so good that any other move is considered awful. White 80 should be at 83. Black 81 should be at 83. White 82 should be at 83. Finally, Black 83 is at 83, so this sequence ends up good for Black.
Black 83 is good for two reasons: it connects the two weak Black groups, and it is a torn keima for White.
Variation: White cleanly separates the two Black groups and avoids the torn keima. In this variation, the four Black stones marked B in the centre are pretty much dead already, and Black will need to add a move to save the Black group A on the North side. Then Black will be gote to save A, and you can play at C to close your centre territory and make lots of points.
In the game, Black connects the two weak groups A and B, and your territory in the centre is pretty small. Remember that the main strategy in go is always separating the opponent’s weak groups and connecting your own groups.
This is the game. This is a very good sequence for White, well done! Just imagine if there had been a White stone at the marked point, if you had separated the two weak Black groups instead of letting Black play move 83. Then this sequence would not only make lots of territory on the North side, but it would also kill the Black group. It would look like this:
Variation: The same sequence which you played in the game, but now Black is all dead. Of course I’m fantasizing a little here, but the point is that separating the opponent’s weak groups is super huge.
Well I’m going to assume you were distracted This kind of mistake often happens to me in correspondence games.
Thank you, I might come up with some follow-up questions later
In full honesty, I didn’t miss the atari. I vaguely remember a corner thingie where W goes for the corner and gives B the centre, but I botched it beyond recognition. So, I thought I had to give up that stone. I’m embarrassed.
You can give up the other stone, the White stone at O18:
But P17 is the central White stones that separates all the Black groups. You cannot give up that stone.
In the game, your move 122 is a cutstone too, but this cutstone only makes sense if you still have at stone at P17 to cut. Without P17, there is no cut, so 122 doesn’t work:
Now White has three groups (O18, Q16 and Q18). The three groups are separated; the two White stones in the corner at Q18 are already dead, the two White stones on the side have so few liberties that they might die too, and the White stone at O18 on the second line is also one move away from death. By contrast, all the Black stones are connected.
In the game Black played N18 instead of 125, so you got some territory back in the corner, which is a little better, but still not very good.
There are several joseki variations where White sacrifices a stone at O18 or at Q18, and there is even a variation where White sacrifices Q16, but never P17. There is no variation in which White sacrifices P17. You can never sacrifice P17, otherwise all of the Black stones become connected, the White stones become cut, and so all the other White stones lose their usefulness.
This is a very detailed explanation, thank you.
Even if I don’t exactly remember the joseki, knowing the logic behind it can guide me .
Move 22 is an example of self-smothering. If anyone finds moves like these hard to understand, the explanations at Play Go at online-go.com! | OGS and the preceding two puzzles might help.
Wow. This post should be stuck somewhere as “mandatory reading” or similar.
What a fantastic, comprehensible and extensive explanation of the scenario there!