Life In 19x19

dust wrote:It's probably a subject more suited to another forum thread, but I think language has its limits. It's easy to think that if only you have the vocabulary nailed down correctly, and the associated concepts adequately described and understood, that 'you've got the right tools for the job' to play Go - as John put it above.

But playing Go is a different experience for me though. When I'm playing intensely it's mainly a strange kind of non-verbal thinking, some with instinctive and some deliberate considerations. As far as I can tell, when I am playing it's very rarely that I'm translating what I'm doing to a 'verbal level' or language. This might happen more formally after the game when I'm reviewing it.

And when I'm learning from handicap games from AI (and from stronger human players as well), I sense that I'm learning the 'feel' of the game and how to play better without necessarily being able to put it into words.

For me, the connection between language and Go is a lot less obvious than it is for some others.

You are right about the level to which we articulate concepts while playing: we don't, we rather act according to understanding, which is mostly visual-intuitive but which is fueled by articulated understanding through analysis.

Let me compare this debate to a hot debate in tennis: the pronation on the serve. There are those, like me, who claim that pronation is a consequence of leading the racket on edge and then going into square on contact with the ball. There are others who claim that pronation is a purposeful action to generate extra power from the "wrist snap" and therefore you have to lead on edge. The debate is not conclusive, at least at the strong amateur level where the youtube debates happen. However, no one is having a different understanding what pronation is, what leading on edge means and what square contact means. No one is debating what split step means or lagging the racket on the forehand or what it means to have the dominant shoulder in front at contact. All those terms are equally well understood and once you understand them you can apply them. Then when you have mastered the technique you don't really articulate it any longer.

This is very different from the debates on thickness or sente, which always get stuck at the conceptual level. We can't apply them, some state, because we're applying the terms wrongly. That makes it very hard to learn.

A similar, perhaps more relevant analogy is table tennis, which is dominated by the Chinese. The Chinese have a much more refined understanding and they apply different techniques than the Europeans. They topspin on the forehand with a straight arm at the start, so that the elbow snaps into contact. European teaching is to keep the elbow stable and only use hip rotation. The debate is there but no one is arguing that hip rotation or straight arm are misunderstood. On the backhand, the Chinese apply leg thrust. My teacher says I should not use leg thrust, just the arm. Again, leg thrust is a well understood concept, the debate is only whether you should do it or not. I believe the CHinese have it right, so my teacher is wrong, but there is no debate on the terms.

I'm not saying my understanding of sente is right and John's is wrong, only that the fact that we are debating the term is preventing us from making progress because we're stuck at the conceptual level, not the executional level. I suspect John agrees on that.

I'm not saying my understanding of sente is right and John's is wrong, only that the fact that we are debating the term is preventing us from making progress because we're stuck at the conceptual level, not the executional level. I suspect John agrees on that.

I'm not sure that I do agree, actually:).

When learning things we tend to start with from words, from which we grasp implications. Once we have the implications, we turn that into an image (or a technique, or application, or whatever), and once that happens we forget both the words and the implications. Of course, even when we all start that journey with the same words, we can end up at a different destination because we choose different implications and images. But I would argue that we most of us end up in the same ball park watching the same game, just in different seats.

But the basis of my argument is that if we start with a different set of words, we are likely to follow up with different implications and different images/techniques/applications. The result is that we end up in a completely different ball park watching a different game. Same rules, different game.

But is one game better than another?

It is my belief that have been fed a diet of technical words (representing concepts) by Japanese pros over several decades. I also believe that if we had all been Japanese speakers, we would have been likely to have come up with the same sort of implications and images or concepts that the pros themselves did. Differences would be largely a matter of degree, not substance. In particular, we would have had no real problems with the basic concepts, as we would all be on the same trajectory. And I project that tot he belief that, on the whole, we would have been stronger at go. Or, at least, would have become stronger much faster. We would be playing, or at least understanding, Major League Go.*

But our word bank in reality is a Japlish mixture which doesn't mean quite what the Japanese pros intended. So we end up in a different ball park, or on a different go board, playing ball at college level, perhaps, but still not at Major League level. We have achieved some sort of conceptual understanding, but just not quite the one that the pros have. But actually we are not even in the Minor Leagues. We are in the European League. So it needs a little more than fine tuning.

So, our don't see that we are stuck on concepts. At least in the sense that I believe we are all perfectly capable of coping with concepts. I believe we are stuck on the starting points (the words for concepts). And the reason we are stuck there, I believe, is not because we don't understand what the right starting points are, but because it is truly difficult for amateurs (i.e. hobbyists) to summon up the energy to start all over again. It is a psychological block, not a mental or intellectual one.

I think the best we older players can do is to just select one element at a time. But for study time only. I certainly wouldn't recommend trying to play an actual game by applying concepts consciously. But I think too many of us do, simply because we have little free time so we try to get twice as much for our money by combining study and play in a single actual game. Turning that from words to implications to images, that to me just conjures up getting blood out of a stone!

*One thing that influences me greatly in this view is that I come across so many Japanese (or Chinese or Korean) dignitaries (writers, poets, politicians, military men, etc) who play go to a rather advanced level, say 5-dan amateur upwards, when it is clear that they have never had time to study the game properly or even play all that often. I speculate that the reason they are strong, apart from being of high intellectual calibre is that they (a) started as children and (b) learned go concepts via the words understood in exactly the same sense as the respective pros intended. We see the same sort of thing in chess over here.

Can I invite you to watch a portion of this video? Ma Long, arguably the strongest player of all time, shows the various techniques in table tennis as he applies them and also explains what exactly is going on. After that, his coach at the time adds to the explanation, reasserting what amateurs should take away from this.

https://youtu.be/ddVkXRh1UeY?t=72

You can stop watching after the serve part, or continue to look for all other techniques.

What interests me is that the terminology he uses, although the emphasis is very different from what I'm getting in my local club from experienced but amateur coaches, is not mystifying. It is translated from Chinese into an English that is rough around the edges but the terms are very clear. "Center of gravity", "reverting from the serve grip to the standard grip", "forward thrust", "contact point on the racket and on the ball" ...

Perhaps the terms are clearer because it's a biomechanical activity which is demonstrated along. Or maybe they do have a counterpart in simple terms like "hane", "atari", "seki" ... on which there's also no confusion. More elaborate concepts like "power from the ground" or "minor adjustments" may equate with the more contentious go terms "thickness" and "sente".

Table tennis players who go to China for a bootcamp always return stronger, because of the high quality and intensity of the training, and the tougher competition they meet. They get their instructions in (poor) English or by mere demonstration. Chinese players are stronger because they have been immersed in this environment from childhood and because of the survival of the fittest, out of a big pool to start with. I strongly doubt that language has a lot to do with it, but I can't prove my point.

Professional content like this is rare but there are channels by professional coaches from other countries, who have trained and competed in China the Vietnamese Ti Long being the prime example. Even though his English is poor, his lessons are not to be mistaken. They do cause confusion in the sense that European club coaching is lagging behind, or differing from some of the technical advice.

For professional Go content, we have the formidable Michael Redmond, whose lectures can't be doubted and who is a native speaker on top of that. Baduk Doctor has stopped his instructional videos, unfortunately. Western amateurs like Dwyrin, Nick Sibicky or Golasses are fun to watch for a while but are obviously not at the level of Redmond, both technically and pedagogically. They're anyhow in it more as youtube personalities than real teachers.

I should look more closely into Redmond's content to see if the terminology he's using seems to mean something very different than what I'm used to. He's raised and bred in Japan, speaks the language fluently, knows all the concepts and their articulation in both Japanese and English, so if there's a strong difference in quality, even meaning of the terms, we should see that gap being closed in his videos.

Dieter:

I'm very conversant with the type of high-level coaching you talk about for ping-pong, through taiji, which also has a huge input from the Chinese side. But I don't think that's relevant here. Things like martial arts, sports and music can be taught without words. Instead of a paradigm "words > understanding > mastered technique", the paradigm becomes "monkey see > monkey do > mastered technique". Furthermore, such activities can be easily broken down into individual and highly repeatable elements, where a coach can watch what you do and give instant feedback, and repeat until the end technique is mastsred well enough.

Go (like chess) is very different. The elements cannot be broken down and repeated so easily. Many of them can only be seen to work by seeing how they work in the context of a whole game (not via a single stroke on a ping-pong table, or a throw in judo, etc). Therefore, the normal form of coaching is heavily biased towards words, and in particular towards commentaries.

The words do not necessarily have to be written down. If you live in Japan, say, the likely way you learn go is from a parent or sibling, or maybe a teacher in an after-hours school class (all exactly as per chess for us). They will use spoken words and you can have a conversation (which you can't do with books, of course), and that may well be a superior way of learning, but it still follows the words > understanding > mastered technique paradigm.

I too would love to see Redmond take a bigger part in western go, but at the moment he is a commentator not a coach and as regards terminology just seems to follow the western crowd. In videos, he seems either to respond to questions in the form put to him, or to use Japanese words without delving into their nuances.

My various posts are not directed at you specifically. It's just that you are kind (and brave) enough to act as an Aunt Sally to get discussion going. But I will take you up on one specific point.

As for initiative meaning control, for me, the player in the lead controls the game. Since we already have a word for that (lead), I don't need initiative.

In your five-games, the game begins with you having a big lead, and even with several mistakes you will usually maintain a noticeable lead well into the game. But who has the initiative? Who has control? As I understand you, you believe Black has the lead but White has the initiative.

But I don't. My working definition of initiative/control in such situations is that the person in control is the one who decides where play should continue next. That person may not necessarily be the first to play in that next area. As in a handicap game. White has sente at the start but the places he can play and the types of move he can use are very strongly constrained by Black's grip on the game. I am not advocating abolition of the term sente when it means first to move or have the prerogative. But I consider it useful to abolish it for situations where we are talking about control of the game. If you don't do that, and retain the term sente for that, you (the collective readership, note) are going to end up always thinking you have to have the next move. That leads to testosterone-fuelled invasions and overlooking honte-type moves. Two of the biggest faults in western go, I'd say.

John Fairbairn wrote:Dieter:

I'm very conversant with the type of high-level coaching you talk about for ping-pong, through taiji, which also has a huge input from the Chinese side. But I don't think that's relevant here. Things like martial arts, sports and music can be taught without words. Instead of a paradigm "words > understanding > mastered technique", the paradigm becomes "monkey see > monkey do > mastered technique". Furthermore, such activities can be easily broken down into individual and highly repeatable elements, where a coach can watch what you do and give instant feedback, and repeat until the end technique is mastsred well enough.

Go (like chess) is very different. The elements cannot be broken down and repeated so easily. Many of them can only be seen to work by seeing how they work in the context of a whole game (not via a single stroke on a ping-pong table, or a throw in judo, etc). Therefore, the normal form of coaching is heavily biased towards words, and in particular towards commentaries.

The words do not necessarily have to be written down. If you live in Japan, say, the likely way you learn go is from a parent or sibling, or maybe a teacher in an after-hours school class (all exactly as per chess for us). They will use spoken words and you can have a conversation (which you can't do with books, of course), and that may well be a superior way of learning, but it still follows the words > understanding > mastered technique paradigm.

I too would love to see Redmond take a bigger part in western go, but at the moment he is a commentator not a coach and as regards terminology just seems to follow the western crowd. In videos, he seems either to respond to questions in the form put to him, or to use Japanese words without delving into their nuances.

My various posts are not directed at you specifically. It's just that you are kind (and brave) enough to act as an Aunt Sally to get discussion going. But I will take you up on one specific point.

As for initiative meaning control, for me, the player in the lead controls the game. Since we already have a word for that (lead), I don't need initiative.

In your five-games, the game begins with you having a big lead, and even with several mistakes you will usually maintain a noticeable lead well into the game. But who has the initiative? Who has control? As I understand you, you believe Black has the lead but White has the initiative.

But I don't. My working definition of initiative/control in such situations is that the person in control is the one who decides where play should continue next. That person may not necessarily be the first to play in that next area. As in a handicap game. White has sente at the start but the places he can play and the types of move he can use are very strongly constrained by Black's grip on the game. I am not advocating abolition of the term sente when it means first to move or have the prerogative. But I consider it useful to abolish it for situations where we are talking about control of the game. If you don't do that, and retain the term sente for that, you (the collective readership, note) are going to end up always thinking you have to have the next move. That leads to testosterone-fuelled invasions and overlooking honte-type moves. Two of the biggest faults in western go, I'd say.

Thanks John, both for clarifying my role as Aunt Sally and for making the distinction in the second part. It makes sense. However ... AI Sensei has the initiative, by your definition, in these games because it is stronger. Sometimes I will play well and have or keep the initiative but overall I will mostly succumb to its strength, by virtue of either taking sente where I shouldn't or not taking sente where I should. It's these actions that are useful for me to think about, not the accumulative effect of AI Sensei's superior strength.

In the 5d game I won, you might say I had the initiative despite the 5d being stronger, because I crushed that player, mostly because he decided to leave his lower left corner undefended (he took sente) and I punished him for that by killing the corner. The concept sente is useful to me here in the sense that he claimed to have sente but he didn't. It is not useful in the sense that I won the game because he "didn't have the initiative". Sente describes a situation: sometimes you have it, sometimes you don't, sometimes it is on the edge. Over the course of a complete game, a player who has the better understanding of sente wins, but it doesn't mean he had sente all the time. I would call the more comprehensive, accumulated concept "control" while I prefer using "initiative" (or sente) more situationally. In your boxing analogy, I could agree that Muhammed Ali controlled the match, even if he left the initiative to his opponent most of the time. The same thing can happen in Go: the player in control makes non-aggressive but efficient moves, strengthening and expanding his position, leaving sente to the opponent, who then uses his sente to make ineffective attacks against the stronger player's positions, which only increase the gap (thank you moves). Then at some point, if necessary, the stronger player executes his sente once to administer a lethal blow to the weaker player's overstretched positions. The stronger player was in control all the time but didn't have the initiative all the time. That's how I would use the terms to describe the situation.

John, I can't help wondering if there's a touch of "typical mind fallacy" in your comments. It's clear that precise use of words, and awareness of their rich connotations, has helped you get a better understanding of go, and could help a fair few other people. But that doesn't make it universal, or the only way to go about this. I think you underestimate what can be achieved through non-verbal learning.

Nowadays, there's a bunch of kids who've got online, played a few games of go, played some more games, played a lot more games, and ended up as 5 dan without ever opening a book or learning any of this terminology. (Of course, this mode of learning doesn't work for everyone. There are people who've played thousands of games online and are still stuck at the kyu level. But then again, there are also kyu players who've read lots of books without significant improvement, even in Japan.)

John Fairbairn wrote:Things like martial arts, sports and music can be taught without words. Instead of a paradigm "words > understanding > mastered technique", the paradigm becomes "monkey see > monkey do > mastered technique". Furthermore, such activities can be easily broken down into individual and highly repeatable elements, where a coach can watch what you do and give instant feedback, and repeat until the end technique is mastered well enough.

This is a little below the belt. "Monkey see, monkey do" as you put it is how we learn to speak our native language and walk on two legs, both of which involve a blend of calculation, intuition and subtle context-dependent adjustments which are more sophisticated than professional-level go strategy (as evidenced by the relative progress of AI in all three domains). And there's a little more to high level sport and music.

I realise I'm probably the weakest go player in this conversation, so let me talk about another domain, classical music, where I've both performed and taught at a professional level.

Music has its mechanical aspects: you have to master the physical techniques of making sound on your instrument, just as you have to master capturing, cutting, connecting and so on in go. But a large part of musical instruction is interpretation: how to shape your sounds in such a way that all the details add up to a big picture that means something. If you use words to teach this, you can outline some general principles ("musical phrases should have an arch shape"/"play way from thickness") -- but then you'll start running into the exceptions (here's an interplay between the melodic line and the harmony which means you should shape it differently; here the form of the piece as a whole means that your local sense of direction should change, ...)

Some teachers do indeed have a rich vocabulary to describe interpretation, and to philosophise about music in general. But others are much more pragmatic and detail-oriented: they might dictate exactly how to shape a specific phrase of music, leaving the student to generalise for themself across multiple examples ("in this position, you should play at this point"...repeat several times per lesson), or they might direct the student towards roles model ("play through lots of pro games"/"listen to Martha Argerich's recordings"). My experience is that teachers who fall on the extreme "philosopher" end of the spectrum tend to be careless about details; they're popular with students because they're fascinating to listen to (and because they don't make the students work too hard), but the students don't improve.

Heinich Neuhaus famously went into "philosopher mode" when he taught Sviatoslav Richter. Some of the lessons didn't even involve a piano. "Today we're going to walk around the art gallery". But this is the exception that proves the rule. Richter came to Neuhaus already in possession of a formidable technique, and Neuhaus set out to lift him from "world-class" to "legendary". He tailored his instruction to Richter's needs, and didn't teach his other students in the same way.

If you praise a great musician, you generally praise their artistry, the spiritual or emotional qualities of their music-making, their sense of form and balance -- not their ability to play a C major scale evenly -- just as you praise a go master's fighting spirit, style and creativity, not the way they hane at the head of two stones.

Another thing: in the first half of the 20th century, a disproportionate number of world-class pianists came from Russia. England and America were way behind. Was this due to some unique feature of the Russian language, a set of words that describe musical interpretation perfectly and that are commonly mistranslated? I think the real factors were the decades of investment in musical culture in Moscow and St Petersburg. They planted the seeds in the 19th century, and harvested the fruit in the 20th. If you study piano in those places and want to play professionally, or even at a "serious" amateur level, then you have a formidable peer group, a lot of excellent role models, and the necessary infrastructure. You need to measure up, and there's a lot of motivation and a lot of support to help you do so. Just as for go players in some parts of China, Korea and Japan. For pianists, a lot of the English-speaking world caught up last century, Asia has pretty much caught up during my lifetime, and I can tell you it wasn't achieved by better translations of technical or conceptual terms.

(At the risk of starting yet another tangent in Knotwilg's thread -- I think the failure of western go to catch up to Asia is a mix of the perceived status of the game here, plus the way we keep modelling our tournaments and rankings on Asian models, without either the cultural norms or the funding to support them, instead of learning from what has succeeded for the chess community. I think our fetish for the pro-amateur divide is doing a lot of harm. That boundary is a lot fuzzier in chess. But that's a whole other debate.)

To be clear John, I love your work, I've bought a few of your books, and I always look forward to your posts on this forum. You're making a big contribution to the go community, and I hope you keep on doing the things that you do better than anyone else. I'm just saying that language is but one part of thought, and for many of us it doesn't have the primacy that it does for you.

Intitiative = Meta-sente=∞Sente

The point is a higher meta-level sente always takes precedence over a lower level sente. The easiest demonstration for this that amateurs would be familiar with would be in the endgame. A kyu player might see multiple sente moves on the board and think that means they're all miai. But if you play a sente move where the opponents response is sente, then that sente move had already lost the meta-sente. Even if the kyu player maintains sente in the endgame they might lose points by choosing an order of moves that loses the meta-sente. If language wasn't so important, no one would read books. Even before AlphaGo people would have been getting stronger only through playing games!

Started a new topic viewtopic.php?f=10&t=19043

I'm just saying that language is but one part of thought, and for many of us it doesn't have the primacy that it does for you.

xela: I often say here that I love to read about how other people think, and your remarks about music were therefore an enjoyable interlude for me. But I still feel I am being misinterpreted.

First, I did already mention music as one field where it is possible to achieve good technique without too much use of words. There are also examples closer to home in go that I have mentioned often. To take one, I often quote my GoGoD partner T Mark Hall. He became a British Open champion mostly by ignoring books in the normal sense, and of course he was stronger than me. He despised most English books and just collected books with lots of game records he could transcribe. That was what made him strong, he said. He would spend usually about 5 hours a day on that, and several more hours playing on kgs. Similarly, I have written about Sumire. I would imagine she was stronger than me when she was six and probably couldn't even read and wrote properly.

In fact, in several activities I do, I ignore the word route myself. I learn Scottish country dances, which is fairly challenging both physically and mentally. Dances emerge in written form, with devisers writing "cribs" which consist of phrases such as "1s cast off round 2s and turn 1 of 3 to face 1st corners, turn 1¾ to face 1st corners for two half diagonal reels of three, finishing 3 2 1." All those numbers (popular with many) make me scunner. Instead, I prefer to be a seeing-and-doing monkey by watching word-free videos of dancing being performed. I think I may even be in the majority there. Since I spend more time on dancing than on go, I can perhaps claim to be giving primacy to words-free thinking.

But I can use dancing words if I have to, and I know Mark could, and Sumire can, use go terms. Huang Longshi was nicknamed the Prodigy and was cited as poorly educated, but he went on to write the classic work on go theory. I'm not sure I believe that scores of kids online reach their level just by playing and never using go words. At least I'm sure they can use them if they want to. Sorry.

People who study via videos are using words. People who study with AI use words, as in this thread, to try to encapsulate what they have learned. In fact, pros seem to do the same. I took special notice of the recent Game 3 of the Oza because it involved a rare example of AI being mentioned in a Japanese commentary. Iyama was talking about a large-scale centre move by Yu Zhengqi (move 23 if you want to look it up) which he answered with a splitting move on the side. But Yu seemed to have expected that and followed up by pressing Iyama down on the side, so as to create a lot of influence in the centre. It was one of those AI-type positions where it was hard to say if Black's influence constituted just a knobbly wall, thickness or a moyo. But an extra AI-ish dimension came in from the fact that White had incipient thickness and an incipient moyo on the opposite side, so that Black's knobbly wall was impacting on that. Iyama said specifically that he had no experience of such AI-inspired positions, and the colour writer also added that Iyama had shown surprise on his face when he saw Yu's moves. But when Iyama tried to explain what was going on and how he had been thinking of how to respond, guess what, he used words like thick, thin, moyo and open skirt. (For the record, the way he decided to play was to choose moves he was comfortable with - the "human" route. And he won.)

But actually, I think all that's irrelevant in a way. The plain fact is that I have simply been addressing myself to someone else who chose to put the focus on go words by starting this thread. There was no encouragement from me - no control, no influence - and the initiative came from him. He also chose to respond to me (interestingly) with more about go words. And, to me, he seems typical of the overwhelming number of western go amateurs.

Even when I did respond, I did not specifically say you have to use go words or more of them. All I have been saying, in essence, is that if you do use go terms, I believe you will get more out of them if you use them properly. C'est tout. Furthermore, the number or words in question is tiny: thickness, yose, sabaki, sente. And in similar, ma non troppo mode, I have limited myself to ways for amateurs to absorb techniques into their intuition (i.e. words are a starting point, not the end result). I have eschewed high-level "interpretation" level matters such as fighting spirit or becoming a pro.

So, I don't feel guilty putting of primacy on go words (even though I confess to being a words guy personally), nor do I feel I have been guilty of assuming everyone else is made in my image. Perish the thought!

As someone who has been interested in how we internally interpret language recently since my literary ability is not what it used to be when I was a teenager, I'd be interested in an experiment where we get 64 beginners from a wide range of backgrounds to play 8 games each every day on Fox, Tygem, Oro, Wbaduk, Yuugen no Ma, Pandanet, KGS, OGS, so 64 19x19 games a day of intense training, for 64 days, and after each game they must annotate their thoughts. They're not allowed to read anything to do with Go to avoid any possible linguistic bias in their thinking. It would be very interesting to see how they play, how good they get and the language they use when reviewing each game.

And we can do the same for Table Tennis and Piano. I deliberately want to avoid learning music theory to see if I can play the 2nd ending of Hikaru no Go by ear. Something similar would be getting Table Tennis and Tennis pros to play each others games, Or even the Xjianshochess I made, haha!

These scenarios push people outside the normal lexicon while having a similar situation while perhaps getting to the truth of the matter, and I find it very interesting

Thank you John for your considered reply. Yes, you did mention music. The reference to monkeys is unfortunate, as people often use that phrasing as a put-down. And the comparison with learning the steps of a dance isn't all that helpful. My point is that in music lessons, there's a lot more going on than literal repetition of patterns. (To be fair, I've done just enough Scottish country dancing to know that there are points of style to be learned, and more to it than just memorising the steps.) Perhaps you actually have a much higher opinion of monkeys than most people. (Ambiguity intended.)

In any case, you've gracefully corrected my broader misinterpretation, and even rewarded me with the vivid image of a "knobbly wall", some words that even I will remember. I had fun looking up the game. Thank you for this gift, and please accept my apologies for the faulty interpretation.

I appreciate the debate and the apparent interest in my progress as Aunt Sally for the Western go amateurs

But perhaps some the tangent on the value of language in the development of expertise deserves its own thread. It's not that I don't want to host the current discussion, so feel free to continue.

In the meantime, the promising results have not yet been confirmed. Last 5H game I lost by blundering in the late endgame while 13 points ahead but the good thing is that at some point I outplayed the 50 playout AI as confirmed by the 500 playout analysis AI.

https://neuralnetgoproblems.com/ seems to be down, so no training today there.

Knotwilg wrote:https://neuralnetgoproblems.com/ seems to be down, so no training today there.

Strange, it's working fine for me right now.

Knotwilg wrote: But perhaps some the tangent on the value of language in the development of expertise deserves its own thread. It's not that I don't want to host the current discussion, so feel free to continue.

I tried to create a new thread here haha.

Another win against 5H 50 playout AI Sensei. Not too many missed chances, still it reduced the 55 point advantage to below 15.
The danger of playing this for too long is that one becomes submissive. Almost any attempt to dictate the pace of the game (sente) runs into a counter. Most gote are either proper moves or not too bad.

On the other hand, it becomes very apparent how shape matters. White's stones are superefficient. Mine are often a little overconcentrated.


[/sgf]

Knotwilg wrote: The danger of playing this for too long is that one becomes submissive. Almost any attempt to dictate the pace of the game (sente) runs into a counter. Most gote are either proper moves or not too bad.

Maybe move on to 4 stones....

My own experiences ("You" = me):

- At some point, you reach a handicap where you can't win by simply defending. It's worth exploring how KataGo would itself play as Black - which is surprisingly active.

- I think you learn a lot about "how not to lose points". A lot of 'active play' is an actually a form of gambling unless you can read it out at some basic level (obviously AI reading level is way above that). Overall, active play is a very good thing though.

- You develop a 'sixth sense' about vulnerabilities in groups/territories and where a future attack may come from further down the line.

- You develop a better endgame through repeated practice

- Completely beyond my understanding and abilities, there's unexpected levels of skill in emphasising the centre and applying pressure on groups, and through this being able conjure up moyos seemingly out of just a few moves. [Edit for clarity: obviously I'm already aware of this possibility, but this is with Black having 3 or 4 stones handicap already on the board ]