Please see below on a game I lost today (i'm black). This game surprised me a bit since I thought i was ahead but ended up losing by over 20 points haha.

I should've counted territory more often

Again, you built superior positions, then let your advantage evaporate through dubious attacking technique.

wrong file attached

Thanks. Fixed it.

These days I'm careful giving reviews based on my 2d insights because much of what I would spontaneously exclamate might be simply wrong. There were a few surprises indeed when reviewing with Katago.

But let's first look at the 3 biggest mistakes as per AI

1) Responding to an overplay in a small way

As Joaz correctly points out, is too close to your strong stones and there's hardly any danger in the cut this move threatens, because both groups are strong on their own. Cutting 2 strong groups is just creating a weak group for yourself. so here is timid.

Click Here To Show Diagram Code

[go]$$W Moves 32 to 32
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . O . . . . . |
$$ | . . . O . . . . . . . . c d . O O . . |
$$ | . . . , . . . . . , . . . . X X X . . |
$$ | . . a b . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . X . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . x . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . 2 . . X . . |
$$ | . . . e . . . . . . . . 1 . . . . X . |
$$ | . . O f X X . . . , . . . . X X X O . |
$$ | . . . O O X . . . X . X X X O O O O . |
$$ | . . . . . . . . . . . . O O . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

KataGo likes x best but to me it's not intuitive. More intuitive moves it likes too, are all of a through f.

2) Surrounding a group which gets easy local life in return:

Click Here To Show Diagram Code

[go]$$W Move 51 to 50
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . O O . . . . |
$$ | . . . O . . . . . . . . . . X O O . . |
$$ | . . . , . . . . . , . . . . X X X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . 2 . . . b . . |
$$ | . . . . . . . . . . . . . . . 1 X . . |
$$ | . . . . . . . . . . . . . . O X a X . |
$$ | . . . , . . . . . , . . X X O O X . . |
$$ | . . . . . . . . . . . . X O O . O . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . X X . . X . . |
$$ | . . . . . . . . . . . . O . . . . X . |
$$ | . . O , X X . . . , . . . . X X X O . |
$$ | . . . O O X . . . X . X X X O O O O . |
$$ | . . . . . . . . . . . . O O . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

As Joaz says, protecting locally with a, or KataGo's preference b, is best here.

3) Wrong connectionm losing a big group

Click Here To Show Diagram Code

[go]$$W Moves 70 to 71
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O . O . . . . . . . . O O . . O . |
$$ | . . X O . . X 2 X . O . O . X O O X . |
$$ | . . X , . O O 1 a , . . X . X X X . . |
$$ | . . . X . . . . X . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . X . . . . . . |
$$ | . . X . . . . . . . . . . . . O X . . |
$$ | . . . . . . . . . . . . . . O . O X . |
$$ | . . . , . . . . . , . . X X O O . O . |
$$ | . . . . . . . . . . . . X O O . O . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . X X . . X . . |
$$ | . . . . . . . . . . . . O . . . . X . |
$$ | . . O , X X . . . , . . . . X X X O . |
$$ | . . . O O X . . . X . X X X O O O O . |
$$ | . . . . . . . . . . . . O O . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

Better connect at a, as you found out soon enough.

One surprising element was your move which Joaz criticized for being wrong direction and which was KataGo's favorite.

Click Here To Show Diagram Code

[go]$$B Moves 49 to 49
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . O O . . . . |
$$ | . . . O . . . . . . . . . . X O O . . |
$$ | . . . , . . . . . , . . . . X X X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . X . . |
$$ | . . . . . . . . . . . . . . O X . X . |
$$ | . . . , . . . . . , . . X X O O X . . |
$$ | . . . . . . . . . . . . 1 O O . O . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . X X . . X . . |
$$ | . . . . . . . . . . . . O . . . . X . |
$$ | . . O , X X . . . , . . . . X X X O . |
$$ | . . . O O X . . . X . X X X O O O O . |
$$ | . . . . . . . . . . . . O O . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

Silly computers still can't handle open ended tactical issues. They get to the leafy end of their search tree, can't find a winning play, and panic.

KataGo thinks Black is slightly ahead at this point. It actually also likes N12 a great deal, which is in Joaz's preferred direction although not as far, but doesn't like Q8 first because it doesn't think White will respond.

Not to spoil the general flow of the review I'm hiding my response to Joaz

Let's talk about 37. It might not be the biggest mistake highlighted by a bot, but it's a huge mistake in thinking and a very common one which can be readily corrected. Sidk, why did you play this move? To reduce White by 1 point? Because you thought you were playing a ko fight? Because playing sente moves make you feel like a boss? Because you planned to cut at q18, forgetting that at its on the second line you can be trivially captured? Because you felt ahead and wanted to make the game closer again? A timesuji?

This is a classic example of a "thank you move", a sente exchange which helps your opponent more than it helps you, your opponent says "thank you" afterwards. The black stone at p17 is of little use. But the White stone at p18 gives White a solid shape and o18 now had lots of liberties. For example, the o17 attachment black might have wanted to use later as a leaning attack against the middle would usually be sente because next you play p18 to destroy the corner, now White already has a stone there and is safe. Black also had a powerful local continuation with s17 and p18, though this a bit more advanced, and this push destroyed that. Or if you press at N17 then next blocking at n18 is a nice move for you because it reduces o18 to have only 3 liberties and had a nice follow-up at p18 clamp to capture it, but with this bad push o18 is already reinforced at p18 so has 5 liberties and is totally safe and can tenuki. All of these potentially useful future possibilities are destroyed by the push.

This push is a particularly bad instance of aji keshi (destroying your future possibilities). Usually with an aji keshi there are several reasonable local continuations, and you pick the wrong one too soon before it is clear which one is the best. But in this case the push is nearly always a terrible move and it's never on the menu of decent choices.

P.S I hope Joaz was joking.

.

thanks for all the advice.

@Joaz - for 3-3 invasion joseki (around move 17), i thought it was normal to end the joseki after building a wall of three and tenuki. i will follow your advice and try extending next time.
@Knotwilg - thanks for the variations. i'm not able to get Katago to work on my computer (too old!) so it's interesting to see what AI recommends
@Uberdude - haha, yes move 37 was bad (although i didn't know that bad). i can't remember what i was thinking, but i think i was taking a long time deciding my next move so i played something to not make my opponent too impatient. i will try to avoid aji keshi in the future.
@rhwocns - thanks for all the variations. i've looked at it once, but it's alot of info. i will review in more detail.

It is not normal. Pros play a press ( N17 or O17 ) most of the time.

But it is playable, and pros have done it. It depends on the rest of the board.
It is usually done when the opponent has strength along the side, and a stone must be added along that side to have a chance at space for eyes.

Click Here To Show Diagram Code

[go]$$c Example in which extending down the side is reasonable
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . 4 . . . . . |
$$ | . . . O . . . . . . . . . . . 2 O . . |
$$ | . . . , . . . . . , . . . . 3 X 1 . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . 5 . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . . . . . |
$$ | . . . . . . . . . . . . . . O O O . . |
$$ | . . . . . . . . . . . . . . . X X . . |
$$ | . . . . . . . . . . . . . X . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . X . . |
$$ | . . O , X X . . . , . . . . . X . . . |
$$ | . . . O O X . . . X . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

Black does not want to let white do this

Click Here To Show Diagram Code

[go]$$c Example in which extending down the side is reasonable
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . 4 . . . . . |
$$ | . . . O . . . . . . . . 5 . . 2 O . . |
$$ | . . . , . . . . . , . . . . 3 X 1 . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . 6 . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . . . . . |
$$ | . . . . . . . . . . . . . . O O O . . |
$$ | . . . . . . . . . . . . . . . X X . . |
$$ | . . . . . . . . . . . . . X . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . X . . |
$$ | . . O , X X . . . , . . . . . X . . . |
$$ | . . . O O X . . . X . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

In your game, black does not need to worry about a white stone attempting to steal eye space, for such a stone would be a target for attack:

Click Here To Show Diagram Code

[go]$$c Position at move 17 - Black has no fear
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . O . . . . . |
$$ | . . . O . . . . . . . . 1 . . O O . . |
$$ | . . . , . . . . . , . . . . X X X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . 2 . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . 3 . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O , X X . . . , . . . . . X . . . |
$$ | . . . O O X . . . X . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

It's one thing what pros do (and yet another what AIs do) but is it fundamentally wrong what the OP did here? He got influence from a corner exchange, White backed off to the second line and Black decided to play an extension. That's pretty sound thinking, is it not? - I see no reason to object.

If we model our game after what the pros do, we can also look into what AI would do. In this case Katago thinkgs the upper right has cooled down sufficiently to play the approach move in the upper left.

Click Here To Show Diagram Code

[go]$$ Approach - the blue move
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . O . . . . . |
$$ | . . . O . . . . . . . . . . . O O . . |
$$ | . . . , . . . . . , . . . . X X X . . |
$$ | . . 1 . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O , X X . . . , . . . . . X . . . |
$$ | . . . O O X . . . X . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

What are the differences with the OP's move and Joaz'move?

Well, Kata's preference makes for an equal game. It finds Joaz' move to lose approximately 0,5 point and likewise for the OP's move. All fall well into the margin of error that we shouldn't bother too much. All are approximately equally good.

Then which one to play?

In such cases, I think you can play what you like. If you want to follow the AI fad, then you go to the uncontested corner asap. If you want to "finish the joseki" that's what you do and if you like extending from influence to build a moyo, that's equally fine.

Click Here To Show Diagram Code

[go]$$ Continue to press down - fine as well
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . O . . . . . |
$$ | . . . O . . . . . . . . 1 . . O O . . |
$$ | . . . , . . . . . , . . . . X X X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O , X X . . . , . . . . . X . . . |
$$ | . . . O O X . . . X . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

Click Here To Show Diagram Code

[go]$$ Extending from influence - also fine
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . O . . . . . |
$$ | . . . O . . . . . . . . . . . O O . . |
$$ | . . . , . . . . . , . . . . X X X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . 1 . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O , X X . . . , . . . . . X . . . |
$$ | . . . O O X . . . X . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

It is normal. Joaz is wrong to say pros usually press at n17 or o17.

In modern pro play, which is pretty much the only place you see this shape as it's an AI joseki, here is how I would characterise the continuations. (Backed up with s quick Waltheri search).
- tenuki. Most common. Extend p16 is the move you choose when 3-3 invaded when you WANT to tenuki and reduce the temperature of the local situation. You can also get sente with the one point jump (if they bump and hane, if second line attach it can launch the Mi flying dagger joseki and then no one's tenukiing for ages), but it leaves a hotter local situation, i.e. one where your opponent following up is a relatively bigger move so you want to return to the area sooner.
- extend on side. Not so common for pros or AI, but not a bad move and I'd recommend it for kyu players to keep the game simple and wall safe. The idea of the p16 extend is the wall is not weak even if white pincers because of the aji against White's weak shape in corner, but pulling this off requires more advanced play. This is why push 37 is so bad, it helps White fix that weakness so actually makes the wall weaker.
- s17 hane combos with p18 to start a choice of local joseki to settle the shape. Probably most common continuation for black if other things aren't going on in the right side, but not immediately. The point of extend was to get sente for big opening points after all.
- o17 attach. This is a shape building move or probe (does white extend out, hane out, extend inside?). It is not so common to play when black is first in this quadrant, it is often played when black has tenukid and white pincered and black now wants to make shape. It is also used as a pressing move if there is an attack on the side, could have been nice in this game around move 50.
- n17 press. This is for building influence and growing a moyo. If you do it too early white happily tenukis and reduces the moyo.

As they say, the best way to get the right answer on the internet is to post the wrong answer.

(Joaz saying tenuki is bad because it's good for White to continue at N16 is also wrong and that's a common shape mistake as gives black the o17 s17 combo, n17 is the move there for white, but too early. Shimari top left is bigger per the LZ gospel).

I included that shape remark in my reply as well but then had the lucidity to check it with AI too. Although kosumi is the standard, the double keima with its aji is playable too - you can check that

Yeah, it's far from a terrible move, but it's the kind of move where a strong player's eyebrows will raise with the knowledge it's not the proper shape move, so although it might look like it's getting more as going further and a stronger threat against the wall, it's actually a weaker threat with the aji it leaves that can help black to settle. I saw this happen in one of Blackie 9p's teaching games on twitch (in fact the corner died later!). There are probably some special situations where it's better than the kosumi, but I'd bet they are few and far between. One of the most common mistakes from my reviews as a low dan with Guo Juan was where I played some non-standard move because I thought the situation was special so it was better than the normal move, but often I was wrong and the bad aji from my trying-harder move came back to bite me in the arse (or could have if my opponent played better, sometimes I got away with it). Of course there's a balance between trying to find the 100% move, stretching yourself and going too far and failing and just always sticking to the boring standard stuff (the former led to one of the most exciting British championship finals in recent years and a painful loss, the later a more mundane game and my first and only championship win). But when you do play these kind of slight overplays, it's best to do so in the knowledge that's what they are and why the standard move is the standard move and this one isn't.