Life In 19x19

Continuing on the underlying debate in other threads, I'm posting a game review as an example of how I review with KataGo
This review report is more formalized for the occasion, so perhaps the inspection has influenced the inspected. Still, the main idea is

1) to review the game (immediately) without KataGo first,
1a) to record the thinking process and the major decision points during the game
1b) and to already figure out some potential mistakes on either side but mostly on mine
2) then review the game with KataGo
2a) to get answers to the questions asked in 1a 1b
2b) to discover "unknown unknowns" via KG's pointing out "major mistakes" (= biggest point differences) which do not occur in 1)

This is also more or less what BadukDoctor, a Hong Kong based Korean amateur of pro strength (at least in blitz), does.
-----

In this review of an easy victory in a no komi game as White against a 1k, I'm asking 10 questions to KataGo.
The answers to these questions can be evaluated in their potential for learning:

10: applicable follow-up of cornter pattern
22: applicable confirmation of shape suspicion
32: probably reinforce sente heuristic
38: indifferent, sharp play not confirmed
56: confirm sharper play
68: indifferent, doubt not confirmed
84: indifferent, safe play not confirmed
92 & 95: tactical verifications, mostly confirmed
116: large scale capture confirmed

KataGo also told me 2 things I didn't ask

12 which side to split - potential lesson,
60 incidental tenuki/shape/threat - hard to learn from
However, this key point comes up a few times afterwards, so it is something to think about.

What would happen if I only looked at "major point differences"?
I would then focus on 12 and 60 but also on 116. I would have missed the "easier to apply" but smaller in impact lessons of 10, 22 and 32

It's an interesting idea for a thread to pick up on earlier Kata Go discussions in various other threads,

I think the problem with this example though is that the game is so one-sided. I think Black made 3 or 4 major strategic (rather than tactical) errors at:

9 - A bit strange to invite you to split his stones and make them weak, instead of the usual L3

11 - I don't really understand the thinking behind this move - maybe it was purely or mainly to create tactical confusion?

39 - Why not just defend at D16? Why let White have the option of taking the left side? 41 could also be at C16 to try to prevent this

85 - this is cutting with no preparation. If B gets a move around J7, then maybe B can aim at weakness in this area such as 05 and keep the cut in reserve?

and this overshadows the tactical considerations.

Maybe I'm just saying there's not enough of your aim 1(b) in your review

I decided to join in the experiment. My process was:

1. Download the uncommented game record from KGS, make diagrams, lay out the stones on a board. While doing so, I wrote down some impressions:

• Moves that looked like clear mistakes
• Moves where I would have played differently but I don't think it makes a big difference
• Moves where I struggled to choose between two alternatives and thought it would be worth asking an expert such as KataGo for an opinion

2. Review with Lizzie+KataGo, check the things I noted from step 1, and also pay attention to moments where the recommended move was in a completely different place from where I was looking
3. Look at Knotwilg's commentary to compare impressions.

There was the usual crop of surprises:

• Things I rated as big mistakes that KataGo said were actually OK
• Ideas of my own that turned out to be awful
• Shapes that were just not on my radar
• Suggestions pointing to positional judgements that seem to go against received wisdom

Some specific examples:

Black 9: My first impression was that it's an awful move! But KataGo rates it equally with the joseki move. Looking at some variations, and thinking about it, I can rationalise this like so: black can look after the groups on each side independently, so being split costs nothing. On the left, of course there's the 3-3 invasion. On the right, I'd forgotten about black's tesuji with the P4 attachment to help settle the group. I've seen this move before, but it didn't come to mind today. So it's good to be reminded about that shape.

Black 19: Again this looked very bad to me, but it's actually KataGo's first choice move! The problem is the followups: after move 25, black has thrown away any hope of invading at C5, and the evaluation drops. Psychologically though, the faulty thinking probably happened on move 19 -- the AI evaluation on move 25 is flagging an error of judgement from six moves earlier.

White 21: Knotwilg asked about N5 versus N4. Again N4 just wasn't on my radar, and I was surprised to see how much KataGo likes it.

White 28: I wondered if white could get away with extending one point further. KataGo said yes, it very slightly prefers the game move but either is OK. But the surprise: KataGo also had a slight preference for tenuki (approach at top right) rather than playing either extension. This seems to go against the proverb "urgent before big". I explored some variations, and indeed white could end up defending a weak and baseless group at the bottom left, but it seems to be worth it considering the points gained at top right.

This might be an example of "an AI move" that humans, or at least amateurs, shouldn't imitate? Getting a good result is conditional on being able to look after the weak group, and it's often easier to find attacking moves, harder to defend well, so the fight might favour black in a game between two people who aren't as strong as KataGo? Hard to say. In a casual game, I might try out the tenuki as a fun learning experience. In a tournament, I'd be more inclined to play the extension, and stick to patterns that I [think I] understand.

White 38: Knotwilg's attachment surprised me, but KataGo says it's not much worse than the usual extension. The other surprise here: white can extend four spaces instead of the usual three (i.e. J17 instead of K17): either is OK according to KataGo. Wait, if you extend too far, don't you just get invaded? Yes, but you make a base at K17+G17, hurt black's top left corner, take a bit of damage on the right, and gain as much as you lost.

That's another "AI move" that I'd be hesitant to try in my own games. But looking at the variations, it's helping me to think more flexibly, and to practice judging the position in terms of trades, not just static territory.

White 46: This is one that I hadn't noted for investigation, but "the blue move" popped up in a place where I wasn't looking for it. The game move is fine -- white has the initiative either way -- but KataGo suggested attaching on top of black's one space jump, playing F14 instead of F16.

Is this violating another proverb, "don't attach to weak stones"? The variations get interesting: if black looks after the two stones below, then the top group can end up weak, since white has the attachment at G17 to play with. This is an idea I actually would feel comfortable trying out in a serious game.

Black 51: I correctly listed this as a mistake, but my alternative was nearly as bad. I didn't see KataGo's recommendation, the shoulder hit at D11. So here's another not-entirely-new-but-forgotten shape for me to relearn.

Black 95: On my first pass, I thought black was clearly lost by this point. But here, KataGo says it's still close, and shows me some very nice defensive techniques (am I allowed to use the word "shinogi"?) starting with an attachment at O5. Well, really, it's as simple as crosscutting to make complications in an area where you're weak. But it just didn't occur to me in this particular context.

All up, I've listed eight examples here. Moves 21, 51 and 95 are the ones where a big drop in the score graph pointed to something interesting. The other five examples are not that significant in terms of the raw numbers, but I'm learning things that I think are helpful. The big swings are a useful starting point, helping me find my bearings so that I don't get lost exploring thousands of variations, but they're not the full story. Note also that some of the lessons are around judgement and direction of play: we can learn more than just tactics from the AI.

Ignoring opening and some strategic elements on the assumption that this is how you wish to play

What stands out the most to me is that white's play from 60 to 80 didn't make much sense to me. What happens is that white is trying to live, that requires many moves and the white groups start fighting each other.

The best way to handle the situation, with insight from katago, is to hit the vital points and fight back, see the diagram.

Click Here To Show Diagram Code

[go]$$Wc Hitting the vital point and flighting back.
$$ +--------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . X X . . . . . . . . . . . . . .|
$$ | . . . X O X X X . . . . . O X X . . .|
$$ | . . O O O O O O X X . . . O O , X . .|
$$ | . . . . . . . . O . . . . . . . . . .|
$$ | . . X . X . 2 O . . . . . . . . . . .|
$$ | . . X . . X . . . . . . . . . . X . .|
$$ | . . O O 1 . X . 3 . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . , . . . . . , . . . . . , . . .|
$$ | . . O . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . O . . . . . . . . . . . . O . .|
$$ | . . . . . X . . . . . . O . . . . . .|
$$ | . O O O . . . . O O O . . . . O . . .|
$$ | X X O X . X . . O X X X . X . . O . .|
$$ | . . X X . . . . . . . . . X . X O . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ +--------------------------------------+[/go]

That was in the thick of the fighting and we shouldn't just assume white should have found the best way here. Instead we can try to understand why white played as in the game.

What I think happened is the following:

1. It wasn't spotted how vital and in the variation are for setting up a good flow for the game.

2. White imagined that black was fairly strong on the left side, that there weren't effective moves like and .

3. White was too concerned with finding permanent safety than he should be in this situation.

4. Since black and white both made bad moves it was hard to realize something went wrong and many moves were played based on these wrong impressions.

Hopefully this doesn't sound too harsh.

Click Here To Show Diagram Code

[go]$$Wcm60 The game
$$ +--------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . X X . . . 4 3 . . . . . . . . .|
$$ | . . . X O X X X . . 1 . . O X X . 8 .|
$$ | . . O O O O O O X X 2 . . O O 7 X . .|
$$ | . . . . . . . . O . . 0 . . . . 5 6 .|
$$ | . . X . X . . O . . 9 . . . . . . . .|
$$ | . . X . . X . . . . . . . . . . X . .|
$$ | . . O O . . X . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . , . . . . . , . . . . . , . . .|
$$ | . . O . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . O . . . . . . . . . . . . O . .|
$$ | . . . . . X . . . . . . O . . . . . .|
$$ | . O O O . . . . O O O . . . . O . . .|
$$ | X X O X . X . . O X X X . X . . O . .|
$$ | . . X X . . . . . . . . . X . X O . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ +--------------------------------------+[/go]

Click Here To Show Diagram Code

[go]$$Wcm70 The game
$$ +--------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . X X . . . X O . . . . . . . . .|
$$ | . . . X O X X X . . O . . O X X . X .|
$$ | . . O O O O O O X X X . . O O O X . .|
$$ | . . . . . . . . O . . X . . . . O X .|
$$ | . . X . X . . O . . O 4 . . . . 5 6 .|
$$ | . . X . . X . . . . . 2 . . 1 7 X . .|
$$ | . . O O . . X . . 3 . . . . . . 8 . .|
$$ | . . . . . . . . . . . 0 . . 9 . . . .|
$$ | . . . , . . . . . , . . . . . , . . .|
$$ | . . O . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . O . . . . . . . . . . . . O . .|
$$ | . . . . . X . . . . . . O . . . . . .|
$$ | . O O O . . . . O O O . . . . O . . .|
$$ | X X O X . X . . O X X X . X . . O . .|
$$ | . . X X . . . . . . . . . X . X O . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ +--------------------------------------+[/go]

Now, how do we analyze this position based on this understanding? I think we must assume that the and combo in the variation doesn't exist. In that case there are many possible moves but I feel the main once are a to c in the diagram. Other moves that exists are d and e.

This understanding can be checked and explored, even with KataGo, by just adding a stone on the squared spot. That is not perfect but if we take it completely literally that the variation just doesn't exist it is an approximation. It is not only I that am more approving of the effort when the extra stone is added, KataGo also finds less fault in the moves.

Click Here To Show Diagram Code

[go]$$Bc Some moves that exist if we assume the marked weakness doesn't exist.
$$ +--------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . X X . . . . . . . . . . . . . .|
$$ | . . . X O X X X . . . . . O X X . . .|
$$ | . . O O O O O O X X . . . O O d X . .|
$$ | . . . . . . . . O c e . . . . . . . .|
$$ | . . X S X . . O . . . . . . . . . . .|
$$ | . . X . . X . . . . . . . . b . X . .|
$$ | . . O O . . X . a . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . , . . . . . , . . . . . , . . .|
$$ | . . O . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . O . . . . . . . . . . . . O . .|
$$ | . . . . . X . . . . . . O . . . . . .|
$$ | . O O O . . . . O O O . . . . O . . .|
$$ | X X O X . X . . O X X X . X . . O . .|
$$ | . . X X . . . . . . . . . X . X O . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ +--------------------------------------+[/go]

I'm, of course, not the player and likely misinterpret to some extent why the player made these choices (something that can also happen when you are the player) but I think this kind of analysis is more useful than saying this was a big points loss and that wasn't. Often comparing with KataGo is casting pearls for swine. When we try to interpret the game from the players standpoint we can sometimes better appreciate how well the players were really playing.

kvasir wrote:
What stands out the most to me is that white's play from 60 to 80 didn't make much sense to me. What happens is that white is trying to live, that requires many moves and the white groups start fighting each other.

I was playing defensively in a situation where I thought I should have had the initiative - and I had, I just failed to convert the mismatch into action.

The best way to handle the situation, with insight from katago, is to hit the vital points and fight back, see the diagram.

Absolutely!

Click Here To Show Diagram Code

[go]$$Wc Hitting the vital point and flighting back.
$$ +--------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . X X . . . . . . . . . . . . . .|
$$ | . . . X O X X X . . . . . O X X . . .|
$$ | . . O O O O O O X X . . . O O , X . .|
$$ | . . . . . . . . O . . . . . . . . . .|
$$ | . . X . X . 2 O . . . . . . . . . . .|
$$ | . . X . . X . . . . . . . . . . X . .|
$$ | . . O O 1 . X . 3 . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . , . . . . . , . . . . . , . . .|
$$ | . . O . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ | . . . O . . . . . . . . . . . . O . .|
$$ | . . . . . X . . . . . . O . . . . . .|
$$ | . O O O . . . . O O O . . . . O . . .|
$$ | X X O X . X . . O X X X . X . . O . .|
$$ | . . X X . . . . . . . . . X . X O . .|
$$ | . . . . . . . . . . . . . . . . . . .|
$$ +--------------------------------------+[/go]

That was in the thick of the fighting and we shouldn't just assume white should have found the best way here. Instead we can try to understand why white played as in the game.

What I think happened is the following:

1. It wasn't spotted how vital and in the variation are for setting up a good flow for the game.

2. White imagined that black was fairly strong on the left side, that there weren't effective moves like and .

3. White was too concerned with finding permanent safety than he should be in this situation.

4. Since black and white both made bad moves it was hard to realize something went wrong and many moves were played based on these wrong impressions.

Hopefully this doesn't sound too harsh.

Not at all! All good points. I think (3) was most at work here. I was too aware of my group being separated and without a base and insufficiently aware of Black's weaknesses in the shape. Had I been able to transition back into an assertive mind set, I think the shape point would have occurred to me.

I'm, of course, not the player and likely misinterpret to some extent why the player made these choices (something that can also happen when you are the player) but I think this kind of analysis is more useful than saying this was a big points loss and that wasn't. Often comparing with KataGo is casting pearls for swine. When we try to interpret the game from the players standpoint we can sometimes better appreciate how well the players were really playing.

I agree that we should first review from the player's thinking process, especially when we are the player. That's 1a and 1b, where I'm checking my own thought process during the game and my thinking in the "non-AI" review. Xela's addition shows how you can do that with someone else's game, factoring in the review. In this case my own review. In my AI review of Shusaku, I'm using pro comments as an angle.

Where we disagree is (2): the usefulness of KG' big point differences. I think they uncover the bigger gaps in our understanding although they may go beyond our current grasp. You seem to insist a player should still be able to find that without KG or at least the point difference argument is a moot one.

But you are a stronger player so you will notice some things in my game without needing KataGo. You made a better assessment of the position in the upper left, hence found the vital shape point and the subsequent capping move to attack rather than defend.

I can't do that. I can't find the things I can't find. Yes, I can find some things in the self review which I didn't find in the game. This is (1b). But I can't find (2) without KataGo, by definition, not on my own. Admittedly, only when you articulated your analysis on the upper left,the whole picture became clear for me. Even with KG pointing out the shape move to me, I didn't fully realize WHY I had not found it. However, it DID point to the major gap in my understanding of this game, being the relative strength of the upper left groups. So - in absence of a stronger player to tell me the full story - KG's big point differences are the best I have to find major gaps in my understanding of the game.

Knotwilg wrote:Even with KG pointing out the shape move to me, I didn't fully realize WHY I had not found it. However, it DID point to the major gap in my understanding of this game, being the relative strength of the upper left groups. So - in absence of a stronger player to tell me the full story - KG's big point differences are the best I have to find major gaps in my understanding of the game.

I wonder if there is a bit more to the "full story" of this area of the board?

The relative strengths of the upper left groups changed after Black played the bad moves of 51 and 53 - creating a shape defect . If Black had simply made one point jumps (for example if he had been reading Kvasir's recent thread on these!), then the strengths of the groups would have been more evenly balanced, and Black wouldn't have left a shape defect behind to be potentially exploited by White to defend the group.

In which case, White would perhaps be a bit 'busy' at the top, taking care of 2 groups. So I wonder if White 38 (the attachment against the black shimari), which let to this position, was actually a good idea - or whether White may have been better to play the standard extension of K16/K17, and keep the game simple?

[/sgf]

Same principle applied: self review first without, then validate with, then check "unknown unknowns"

Self review + validation
10: I was hesitating between a regular extension or my more ambitious approach; I reckoned AI would favor a 3-3 invasion but I didn't want to go there; KG confirmed my thoughts and validated my choice

36: I thought I was ahead here and KG confirms +8,5

38: in my review I thought I should answer rather than attack again; my attack is large but not really urgent; KG confirms (-2)

52: in the game I thought I should cut at the bottom but I still wanted to follow-up on my attack at the top and use the weakness at the bottom as a way of settling after tenuki-ing twice. KG confirmed that the alternative was better (-7) although I'm still ahead (+7)

64: in the review it was clear for me I should have blocked; KG confirms (-7)

74-80: I have squandered my lead; I realized that during the game but KG is showing me many useful variations to help me play more assertively in these situations

85-89 is returning the favor by killing his own group; probably Black thinks he's alive locally; KG and I know better

100: which side to cut? KG: doesn't matter
126: simply descend is better? KG: yes, but the lead is huge anyway
154: D17 is better? KG: yes


Unknown unknowns?
If I had only looked at major point differences, then 52, 64 and 74-80 would stand out. It would basically emphasize some of my questions more than others. Most of all it would be a more negative review, whereas this kind of self review + post review with KataGo includes positive feedback.

In this game I had the following questions:

24 The table shape is natural but I know AI is not always a fan. Is it good here? - KG: it's ok
26 I thought hard about this. It has a scent of overplay. But I thought it was playable -KG slight overplay
30 this can't be a bad result. I'm ahead now, I think - KG: correct
34 I didn't want to play a heavy cut. Was this a good move instead? - KG: playing on the right side would have been best

KG: 44 was anti-suji

50 should I connect instead? - KG: no but the game move was not very good either
74 I come out of this ko with an advantage. - KG: correct but the whole idea of fighting the ko was misguided. The group is not important, the other group is.
98 I think Black's capture was too small - KG: correct
144 After playing this, I thought it was an overplay - KG: the relevant issue was on the left side
166 I shouldn't be on the defence like this but how to play? - KG: it was pretty ok and I could cut any time I wanted

The instructive takeaway is to think of mutual cuts as a kind of soothing. If you can be cut but so can the opponent, a connection is not necessarily better than the mutual cut. If the opponent has to invest a lot into a capture, it is small. AI thinks a lot sharper about such forcing to capture than I do.

Click Here To Show Diagram Code

[go]$$B Black to play
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X X O . . . . . . . . . . . . . . |
$$ | . . X O . O . . . . . . . . . . . . . |
$$ | . . X O . . . . . , . . . . . X . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

I recently analysed with katago a game in which the configuration in the upper left corner appeared and I was quite surprised with the move chosen by katago.
In order to understand the point I simplified the position as far as possible to reach the position above.

Click Here To Show Diagram Code

[go]$$B Black to play
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X X O . . . . . . . . . . . . . . |
$$ | . . X O . O . . . . . . . . . . . . . |
$$ | . 3 X O . . . . . , . . . . . X . . . |
$$ | . 1 O . . . . . . . . . . . . . . . . |
$$ | . 2 O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

I expected to see a black hane but instead katago suggests here to apply the 1-2-3 principle by playing:

Click Here To Show Diagram Code

[go]$$B Black to play
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X X O . . . . . . . . . . . . . . |
$$ | . . X O . O . . . . . . . . . . . . . |
$$ | . 1 X O . . . . . , . . . . . X . . . |
$$ | . a O . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

I just don't understand the point. With a yose point of view I assume white will be able to play later at "a" in sente and the result is a loss of about 4 points for black comparing to the hane in the first diagram. 4 points is rather big isn't it? Where is the compensation for black when chosing the second diagram?
Can you help me?

Black shall not strengthen White on the outside more than necessary.

If white plays A now, katago is not gonna live in gote inside, probably a move on the top side. The corner is still small.
The interesting part is the timing ^^

After White A, the next local move value is 13 while a tenuki is almost 14.

Click Here To Show Diagram Code

[go]$$B Black to play
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X X O . . . . . . . . . . . . . . |
$$ | . . X O . O . . . . . . . . . . . . . |
$$ | . 1 X O . . . . . , . . . . . X . . . |
$$ | . 2 O . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

Oops it is not so obvious.
It seems to me that the exchange strengthen White on the outside. In this context if black plays tenuki after why black plays instead of the immediat tenuki?

White 2 is a mistake, of course.

Compare Black 1 tenuki, White 2 kills still having a larger move value.

Click Here To Show Diagram Code

[go]$$B Black to play
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X X O . . . . . . . . . . . . . . |
$$ | . . X O . O . . . . . . . . . . . . . |
$$ | . 1 X O . . . . . , . . . . . X . . . |
$$ | . a O . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

Thanks to Robert's commentaries and deeper analysis with katago I begin to undersand why this move is better than he hane at "a".
Maybe I am wrong but if after black white has he opportunity to play at "a" in sente early in the game (I mean in the first part of the middle game) then this black move is bad because black loses 4 points comparing to the hane at "a".
As a consequence the crucial point is to evaluated the value of a move after the exchange black and white "a".
Robert's evaluation for such move seems to be 13 points. I agree that in the fuseki the value of a move is about 14 points and thus a white move in the upper left corner is not sente. But in the middle game the situation is quite different: most of the time the temperature is higher than 14 but I guess (not often but regularly) this temperature may be smaller than 14 to reach let's say only 12 points. When this happens white can play at "a" in sente and the result is bad for black.

Why then black can be a good move?
My answer is the following : after the exchange black and white "a", a white move in order to kill black has to be evaluated far less than 13 points.

Click Here To Show Diagram Code

[go]$$B Black to play
$$ ---------------------------------------
$$ | . . . 4 . . . . . . . . . . . . . . . |
$$ | . . X X O . . . . . . . . . . . . . . |
$$ | . . X O . O . . . . . . . . . . . . . |
$$ | . 1 X O . . . . . , . . . . . X . . . |
$$ | . 2 O . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

Let's suppose as an exemple that white kills black by choosing the white above.
Black will be able later to follow by:

Click Here To Show Diagram Code

[go]$$B Black to play
$$ ---------------------------------------
$$ | 6 . 5 O . . . . . . . . . . . . . . . |
$$ | 4 1 X X O . a . . . . . . . . . . . . |
$$ | 2 . X O . O . . . . . . . . . . . . . |
$$ | 3 X X O . . . . . , . . . . . X . . . |
$$ | . O O . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . b . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

and you can see that black has created two sente moves at "a" and "b", decreasing the value of the white killing move.