Reviewing Lee Sedol's commented games with KataGo

[jlt]

John has read and written so many go books, if he thinks he doesn't study, then what does it mean to study?

[John Fairbairn]

John has read and written so many go books, if he thinks he doesn't study, then what does it mean to study?

What I mean is that I don't play and so don't study to become stronger in competitive play. Most of my books are about go personalities and so much of my time is to do with researching their biographies and backgrounds. I accept that in the course of going through commentaries, say, I may become better at understanding aspects of go (in fact I'd be a bit worried if I didn't ). But that's just a spin-off, and I don't think it translates into strength on the board. because non-go factors such as desire to win, concentration and general match fitness are all lacking. I am a go fan rather than a go player. A bee rather than a workhorse. I absorb go's nectar in the hope of being able to produce honey.

[Kirby]

It's helpful to think of KataGo's opinions as being similar to that of a strong pro. Maybe prior to 2016, you had a favorite pro to study - maybe you liked their games or commentary. If that particular pro wasn't the number one player in the world, does it make their commentary invalid? Of course not.

That's because the "blue move" is not really what's important. What's important is the thought process used to try to get to a blue move. Even pre-AI there were out-of-date joseki that were no longer considered even for both sides. However, it's totally educational and valid to study those sequences and understand the reasoning behind the moves. Furthermore, those sequences just might be the correct sequences in certain circumstances.

One thing I've found interesting in listening to Inseong's lectures on various patterns is that he'll give the recommended sequence, but also show variations that aren't recommended for whatever reason. But he makes the comment that, "If you like this result as black, then you can consider playing it." So it's not really like, "Don't play this way because pros have deemed it inferior", but rather, "here are some different options you have and what kind of board position you'll likely get if you play it out".

This way of thinking is more useful. If you play a game and are in a tricky situation, understanding what options you have and making the game flow according to your plan is more important than a "blue" move. After the game, if you played out your non-blue sequence and lost because of it... then you can think about what alternatives you may have wanted to play.

tl/dr: the thought process and sequences are more valuable than the answer itself

[RobertJasiek]

the exercise of reviewing it with KG is valuable to me.

If I may, a question for both you and Robert, one I ask because I haven't a clue what the answer is. I don't study, but you both do, intensely, and so may have some insight I've missed.

Neither of you seem to have reported any significant increase in grade. [...]
Have you got any sense of what [Sumire 2p is] doing that you are not doing, or advantages that she has that you don't?

(Earlier, I have mentioned to have much improved due to replacing subconscious thinking by reasoning. This refers to my improvement from 5k to 5d, which was very fast to 1k, fast to 3d and reasonably progressive to reaching 5d in 1998. My 5d of 2022 is stronger, maybe 1 rank in terms of 1998. Therefore, the remaining question I am trying to answer below is: why has my improvement been slower since 1998 and not surpassed 5d?)

When I started go seriously, I was almost 21. Therefore, I missed the advantage of somebody starting as a child. If I had the chance, I could have started seriously at the age of 4.

I have not had the playing environment of an east Asian prodigy, who might have the chance to play equally strong human opponents every day. I had to travel throughout Europe to tournaments to have more interesting games.

While I could spend an average of 14h per day on go from 1993 to 1995 and a bit less to 1998 (obviously while neglecting university study), I studied go rules more than go from ca. 1996 to ca. 2006 because, for me as an advanced amateur mathematician, playing a game with ambiguous rules was unacceptable. We are aware of the importance of fundamentals and I take them more seriously than most: they start with the rules. Hence, my rules study time has been unavailable for go theory study time.

Eventually, the necessities of life caught up. Although my jobs are go-related, study during the job does not even remotely compare with separate study for improving. IOW, I have had too little time for the latter. If I had started go 17 years earlier and not studied the rules, I would have had very much more time on studying go for improving.

I have not seen any flawless IQ test so my IQ of 155 is just an approximative guess. The level is not bad but there are always more intelligent people with IQs around, say 170 - 190. Or you might say: greater talent. Go playing strength can be the result of time, effort (aka hard work), environment and talent. I could have invested more hard work but cannot compete with greater talent, except by even more hard work that I should have done. You know, instead of TV during the evening, solve another 100 problems.

When I compare the go skills of pros and myself, I notice two major differences: 1) their faster life and death reading, 2) their faster endgame value calculations. Both I might achieve by more hard work. When I compare stronger amateurs, they also make fewer blunders on average (where we speak of high dan blunders rather than fast server game fun) and have a bit better understanding of some other go aspects (example: they can use thickness better on average than I can). It does not matter enough that my understanding of some yet other aspects is better.

More talent would help but what really makes the difference is willingness and time / money to spend day and night, year after year on hard work of studying.

Like the athletes is international events. They train many hours every day for many years.

It is like I did for go theory research on rules, endgame, ko etc. There, I have achieved what might belong to world-leading standard. However, the time spent for it I have not spent on separately studying go directly for improving.

If my opponents have better advice of improving than more time, more focus and more directed hard work, please let us know! :)

[CDavis7M]

I have looked, because my next book (currently being proof-read) is entitled "Ogawa Doteki, go prodigy."

Doteki! I'm one step ahead since I was just appreciating Doteki this past weekend. Wait... am I one step ahead or 45 years behind?

DSC_1419.JPG (54.35 KiB) Viewed 6519 times

Since we are on the topic of study, the only study I've done lately seems to go to waste since my opponent (B) never cooperates with my joeski when I (W) attach.

$$Wc W11 at D8 | Other variation
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . . O . . . . . , . . . . . X . . . - . . . O . . . . . , .|
$$ | . . . . . . . . . . . . . . . . . . . - . . 0 . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . 9 8 . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . 7 6 . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . 5 2 4 . . . . . . .|
$$ | . . 9 7 . . . . . . . . . . . . . . . - . . 1 0 5 . . . . . .|
$$ | . . 8 X 5 . . . . , . . . . . , . . . - . . 9 X 7 . . . . , .|
$$ | . . 0 1 2 6 . . . . . . . . . . . . . - . . 5 1 2 8 . . . . .|
$$ | . . 2 1 3 4 . . . . . . . . . . . . . - . . 6 4 3 . . . . . .|
$$ | . 4 3 7 . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . 8 5 . 6 . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . 9 . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . 0 X . . . . . , . . . . . O . . . - . . . X . . . . . , .|
$$ | . . . . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ ---------------------------------------

Click Here To Show Diagram Code

[go]$$Wc W11 at D8 | Other variation
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . . O . . . . . , . . . . . X . . . - . . . O . . . . . , .|
$$ | . . . . . . . . . . . . . . . . . . . - . . 0 . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . 9 8 . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . 7 6 . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . 5 2 4 . . . . . . .|
$$ | . . 9 7 . . . . . . . . . . . . . . . - . . 1 0 5 . . . . . .|
$$ | . . 8 X 5 . . . . , . . . . . , . . . - . . 9 X 7 . . . . , .|
$$ | . . 0 1 2 6 . . . . . . . . . . . . . - . . 5 1 2 8 . . . . .|
$$ | . . 2 1 3 4 . . . . . . . . . . . . . - . . 6 4 3 . . . . . .|
$$ | . 4 3 7 . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . 8 5 . 6 . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . 9 . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . 0 X . . . . . , . . . . . O . . . - . . . X . . . . . , .|
$$ | . . . . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ | . . . . . . . . . . . . . . . . . . . - . . . . . . . . . . .|
$$ ---------------------------------------[/go]

[kvasir]

It's one of the most personal books by a pro I have ever seen. There will surely have been some help in editing but I genuinely believe the commentaries are his. The commentaries are also spiced up with stories by himself and his sister.

OK. I'll consider it as coming from Lee Sedol. My reason for raising the question is that some Go books obviously have minimal input from the credited author or have otherwise become remote from this input in the English translation.

I tried to continue the discussion by sharing what I see in the first game with the help of KataGo but this thread is wandering off to other topics quickly.

It is hard to say much about the second game. It is almost error free in the opening and early middle game but almost every move from 150 - 181 is a bad point loss. Lee Changho wins this game despite making a -12.1 point move at 181.

I read your observations about the comments on the second game:

I don't understand what KataGo is telling me about move 22 but there doesn't seem to be any benefit in making the inside cut instead of the hane.

The 37 - 47 does seem to squander blacks advantage but I think mostly because 43 is an overplay.

Black's 83 - 101 action effectively gives white the lead, I am concluding that players didn't have close to the same assessment as KataGo because the game is B+0.3 at move 82 but white has a chance to be W+9.6 at move 100. You mention 99 - 100 as appearing in the comments but what I see in KataGo is Lee Changho almost throwing the game and Lee Sedol not completely taking advantage.

Then there is what I said above about moves 150 - 181. I wonder if there is any word on that in the commentary.

[jeromie]

I think the most interesting comment that you posted from the second game, Knotwilg, is that Lee Sedol felt “white’s flow was unfavorable.” You can see from KataGo’s analysis that the game was pretty even at the point where he wrote this, but it sounds like he didn’t get the game he wanted. This bring us back to some of the earlier discussion on the first game: the overall result wasn’t bad, but perhaps it was bad for Lee.

This is, overall, an interesting way to look at professional commentary and how to play the game of go. Each of us is making choices, and we can try to direct the game toward one that is favorable to our strengths and preferences. I wonder how many of my clear blunders (which are many at my level!) come because I let myself be lead into a situation that doesn’t play to my current strengths?

[CDavis7M]

Just to go back a bit... The opening post is about how Lee Sedol's judgment of a move or position may have been biased and how the judgment is sometimes different from AI judgment. Having looked again at Vol. 2 (I don't have Vol.1) I just wanted to point out that even if Lee Sedol's judgment is wrong (I don't think so), his judgment is not the teaching point of the commentary. Lee Sedol teaches techniques for making your own judgment (summarized):

• Taking profit at the beginning makes the game easier to play, but giving away too much thickness may cause you to lose initiate. Maintain balance between profit and thickness.
• If preferring thickness, do not fail to take the initiative using thickness.
• Do not obsess over an area where the opponent did not respond. Punishing tenuki may not be big.
• While being attacked, continuously aim at the opponent's weaknesses.
• Escaping with poor shape may be better than settling quickly.
• If you focus on playing too safely in a game you're winning, it might keep you from seeing good variations. Protect against weaknesses that might reverse the game without being passive.
• Read and be confident if you attack in the beginning of the game.
• You may need to play a bad move to defend a weakness in sente, but that might be worse than just defending in gote. Be certain about the benefits.
• If you always play proper book moves, you may miss a practical move for your situation.
• Judging a move to be absolute sente and missing the chance is a weak point for pros. It is not easy to determine the right timing for exchanges. It might be beneficial or might lead to a loss.

It should be no surprise that KataGo has different judgment from Lee Sedol. But KataGo cannot provide better judgment-techniques or even assess whether these techniques are good or bad.

[John Fairbairn]

Dieter/Robert

Thank you for sharing so much. Both what you mentioned and didn't mention was interesting. I had hoped for some discussion of what an adult focuses on as opposed to a child (as I said in my question post, I suspect adults focus far too much on strategy), but at the same time I admit I did stress that my choice of Sumire as a comparison point was more to do with her having far less time to study go than post-school pros, and so seemed a better fit for us amateurs.

However, I am planning to post something soon about Sumire which I think will feed back into this discussion (though I will post that in the Sumire thread for the sake of tidiness).

[Elom0]

I don't believe go is any different from chess, shogi or xjianggqi in that you should study closed positions, the endgame and tactics, first. Because the point of the opening is to get to the endgame so studying tactics may help one's opening, but not the other way around, probably .

When I was a more wee lad, my view of any expertise was such that of course, I am completely incapable of understandiing anything to do with the subjective opening and of course must rely on the prophetic words of the pros. I just didn't even see how it was possible for humans to have developed an intuition for the opening, so the fact that pros have one to a significant degree must mean they have magic powers I must pray to absorb , and I was specifically interested in reviews where different pros commented on the same game to see any differences in opinion, especially in the opening. Now it seems to me that those intuitions were based first on tactical consideration.

In chess and chess-like games, I had the opposite problem to Taranu sensei, which is that in go it was easy for me to have an intuition of where I wanted to play, whereas in chess I have no idea how any opening relates to somehow ambiguously helping you to capture the opponent's king/general. I mean what does pawn to e4 have anything to do with how I can eventually capture the other side's lead piece? Or protect my own?. So, in the case of chess for me I guess it's more obvious to do what the well-known chess saying says; study tactics and second half of the game first, and the first half of the game should take secondary priority or be studied afterwards, I guess.

A is the tendency. So this means playing moves in the opening that are not necessarily the best, but wanting to win through the pure willpower of intention, and this usually leads to fighting. I think this also makes both of us like reviewing games; at least An Youngill said specifically he likes to do so, and. This intention means However, I happen to be familiar with someone who shares an exact gene with him that I don't, and I know what it does. It means him nd I both complain about the structure of whatever pro organisation we may be in, but he is more likely to get very frustrated and resign, and get stuck on a solution that may not actually work, and be absolutely frustrated with bot's abilities to swerve him, in the opening haha . . .

Actually, I've studied people down to the likely precise set of genes in terms of behaviour, so of course I've seen other amatuers where I think, 'wow, their "style" is just or very similar to so-and-so pro'. Most strikingly a sister pair who's styles are basically Yu Zhiying and Choie Jeoung, and the proof was that their personality types were also similar, as I mentioned many years ago.

Choie Jeoung has a gene that probabilistic situations--exaclty as would be useful in the endgame and chaotics fighting, as An Youngill said she was good at. Yu Zhiyng has a realist personality element that overpowers any tendency of women towards fighting games. Does it mean that we interpret mindsports in the social category of our brains, for that would make sense for women to then be more aggressive in it while men are generally more catious? Interestingly, my tendency to start fights is regardless is apparently the hallmark of feminine go, and we all know what happened with Korea, and it's strange that I haven't seen this point belaboured more. Did the Koreans get stronger by playing more feminine go?

It's probably more important to find a pro with a similar personality type to you, and study them. This is a current limitation of AI, I guess. In that case Tasmin Jones was right, and I've now moved more towards her opinion; we should imagine as amateurs we have a style, even if it's one that doesn't exist yet. It's strange that I'm saying we should study pro games more like aesthetic art pieces, but I am. In fact, studying a pro game hoping to glean objective facts to improve is likely exactly what would guarantee you won't improve. his is the mistake I made because of my view of expertise as god-like, haha. Studying pro games subjectively, tapping into the emotional center of the brain, is likely better, and children probably do this better than adults, so that might exlpain Robert Jaisek's findings did do better relying on non-intuitive based study later on, but that's likely because his initial intuitions learned as an adult where only as good as most adult brains can make them, so an 'adult' brain, one that's tired of growing already, likely is more dependant on that kind of thing .

So pro opinion in go can have as much to do with genetics as actual ability. And of course, which teacher you had, which go dojo you had, whether you learned go in the honinbo house or some other house that has their own fashion and dogma for the opening, in which case the point is very obvious. I think we amateurs think pros have moved on from those days, after shin-fuseki, but maybe it's a delusion, maybe not as much as we think!

[Knotwilg]

Just to go back a bit... The opening post is about how Lee Sedol's judgment of a move or position may have been biased and how the judgment is sometimes different from AI judgment. Having looked again at Vol. 2 (I don't have Vol.1) I just wanted to point out that even if Lee Sedol's judgment is wrong (I don't think so), his judgment is not the teaching point of the commentary. Lee Sedol teaches techniques for making your own judgment (summarized):
(...)
It should be no surprise that KataGo has different judgment from Lee Sedol. But KataGo cannot provide better judgment-techniques or even assess whether these techniques are good or bad.

Vol 1 provides similar lessons and you make a good case of these being valuable and valid advice, regardless of the tactics and evaluations presented.

Still I disagree with your last statement. KataGo has not just different judgments from Lee Sedol but must on average have better judgment, since it's likely on par with AlphaZero who beat the leading pros 64-0 and pushed them on an estimated 2-3 point handicap. Kvasir has made an interesting case that this 20 point difference on average may come rather from the compound effect of smaller "mistakes" which don't even add up arithmetically to the difference in the end, than from the "bigger mistakes". However, if KataGo's precision is off in any way, the bigger mistakes might even be bigger than quantified.

KataGo doesn't articulate its judgment except for showing sequences underneath its percentage and point wise assessments. It needs interpretation and articulation by a human. At the pro level this is beyond my capacity. At my own level I have been able to discover heuristics from KataGo's evaluations and sequences.

There have been records of professionals analyzing their game right away by showing sequences to each other, because they didn't speak each other's language or even in case they did. Their "hand talk" is often a better communcation tool than speech. I'm sure professionals are learning a lot from AI, despite the lack of articulation.

I'm going to rest my case here and can agree to disagree on the usefulness of AI analysis of pro games, commentaries, articulations and claims.

[Knotwilg]

(...) This bring us back to some of the earlier discussion on the first game: the overall result wasn’t bad, but perhaps it was bad for Lee.

In some parts of the analysis he indeed personalized the evaluation in that way: whether the result fitted his style of play. In other parts he made the claims more absolute, like no one should play that way, while KataGo thought it was playable or even best. (and the other way round). I don't blame Lee for that (anymore): AI hindsight is too easy to dismiss a thoughtful commentary by arguably the best player of his time.

This is, overall, an interesting way to look at professional commentary and how to play the game of go. Each of us is making choices, and we can try to direct the game toward one that is favorable to our strengths and preferences. I wonder how many of my clear blunders (which are many at my level!) come because I let myself be lead into a situation that doesn’t play to my current strengths?

I can give that to a professional. At my level I consider myself to have mostly weaknesses and hardly any strengths. The overall degree of idiosyncracy of AI is IMO overestimated. The only "advice" I'm reluctant to take from KataGo is its choice of invasion points and therefore its leniency to opponent moyos forming. I think I will play a little more influential go than what KataGo advises because I find its invasion tactics too complicated. Still that leaves plenty of room for conceptual learning: slow connections, heavy cuts, raw peeps ...

[jeromie]

AI hindsight is too easy to dismiss a thoughtful commentary by arguably the best player of his time.

Yes, I agree. I have a lot to learn from a variety of sources. Disagreement at a high level can help me to understand that professional advice is not absolute, but I can certainly glean something from their thought process!

I can give that to a professional. At my level I consider myself to have mostly weaknesses and hardly any strengths.

At an absolute level, this is certainly true. And I’m considerably weaker than you. I am sometimes amazed in review (or even in mid game) at the ridiculously poor level of play I can demonstrate. But I do think it can be helpful, at some level, to think of the type of game I want to play and try to move the game in that direction. Not because I’m truly strong in that style, but because it gives me some strategic direction that might help my moves make sense together.

[Elom0]

It would be very interesting to see how much KataGo agrees with pro opinion in this game., although I may be hijacking your thread a bit . . .

[CDavis7M]

KataGo has not just different judgments from Lee Sedol but must on average have better judgment

I think you're missing my point. I don't disagree that KataGo has "better judgment" than Lee Sedol but it is better for itself, not for Lee Sedol, and certainly not for much lower ranked players. A lot of AI suggests have miniscule benefits if playing as an AI and many AI suggestions are too difficult for even pros to pull off.

I'm going to rest my case here and can agree to disagree on the usefulness of AI analysis of pro games, commentaries, articulations and claims.

I don't disagree. It is useful. My only disagreement is that the original post implied that Lee Sedol's judgment wasn't on-point because KataGo said otherwise, but I think Lee Sedol's judgement was correct for him.