So, in the past half year there has been only one professional game of go I have been studying. I am interested in what your opinions are on the 'width' versus 'depth' of material you study.

I know personally that learning happens faster and more permanent when you can relate the material to familiar concepts. For example, a famous memory technique is relating all the material you want to remember to places in your own house. Because you can visualize your house very easily, this means the material will also be retained better.

Analogous, my approach to learning go from professional games is study, study and study the same game over and over. Of course this does not mean that I do not do random go problems or watch random go lectures on youtube. However, when I feel like going over a game on my board it has been the same game over and over again. I have noticed that my appreciation for some moves grew after I played through the game more and more and although I do not understand the 'why' of the timing of certain moves, I begin to appreciate it as well.

My theory is that you become so familiar with the game that you can visualize move after move in my head. Then the specific tesuji's and concepts will be easier to remember in your own games. Easier than if you study ten pro games and forget the moves of the first one when playing through the fifth.

Of course, I am only a kyu-player, so it takes me a long time to memorize a game fully and even more time to understand different variations and moves.

The main question I want to pose however is... Do you think that it is more beneficial to study a few pro games and study them again once you forget some things or would it be more beneficial to keep moving onto new material even though the old material might not be fresh anymore?

I believe that you will appreciate the game more and more, but rather because of the investment you've done than because of true increased understanding. I think you become attached to a game emotionally, just because you've spent so much time with it.

I won't deny you will get something from it but some moves may become too hard wired as well.

A pro game is a great source of inspiration. A pro commentary of it will help explain the moves. But then it's time to move on and implement what you've learnt into your own games and review afterwards if you implemented it well. And do that frequently, then move on to new ideas.

The technique of memorizing pro games, and many of those, is IMO only useful for high dan players because they already have the knowledge which is the fertile ground in which the ideas of those games can take root.

As a kyu/low dan player it's probably a waste of time, but then again, so is playing Go, so if you enjoy it, why not.

So to summarize your opinion is not to waste time on pro games at all.

If we talk about 'width' versus 'depth' however, would you prefer a smaller amount of problems over a larger amount? The smaller amount would allow you to 'recognize' situations from the problems and guide you a bit. I do not see any advantages over having a larger amount, unless you cheat with the small set and solve the problems by recognition instead of concentrating and reading out different variations.

My main point was not to discuss whether studying professional games is good. It was whether it is better to study things thoroughly and then again if it is not fresh again, or to keep moving onto new material.

Never underestimate this. At least half of a strong player's go comes from recognition - of shapes, tesujis and sequences. Reading fills in the gaps, but so much of good technique is familiarity and context.

My gut feeling is to aim for large numbers of problems and aim to spend no more than 30 seconds on each. With a big enough sample size (2000 or so as a starting point), you can then repeat them at the end, and again. Slowly your appreciation of shape just simply appears out of repetition. Just as it doesn't make sense for a 13k to spend 15 minutes on a move against a 3d because time isn't the problem - they're looking at the wrong things rather than not reading deeply enough - it's the same with go problems.

My feeling on such training I guess is: Always read, but never for too long on one position

DeFlow: the abstract matter of width vs depth is not the same for the concrete matters of pro games or problems.

For problems, as topazg points out, it's treacherous to study a few problems thoroughly. Most teachers seem to agree on doing many different problems and when in doubt on the difficulty level, err on the safe side. It is better to train yourself on things that fully or almost fully lie within your intellectual reach, slowly stretching your ability, rather than study a few complex problems, look at the solutions, then study the problem more, to end up with one memorized concrete situation.

There is some controversy about whether one should look at solutions at all, but I believe that when the problem turns out to be rather hard, then if you think you solved it, you can verify if you didn't overlook anything. If you didn't solve it, it may be best to leave it aside to pick up later, but some say that glancing at the solution will bring new ideas.

In a larger sense this is also true for pro games. The fact is: pro games are always outside our reach. Studying many of those makes no sense. Studying a few can be a good source of inspiration. As we don't really understand them, we need pro commentary. A pro game can be seen as a very complex problem, with many new ideas or food for thought.

Personally I prefer going through more pro games, preferably commented, and thinking about specific interesting positions. Just studying a few games may lead you to think a certain way, but there are a lot of ways to play various positions that you can only see with a breadth of games.

Why pick just one approach?

I like the idea of finding a single game that really speaks to you and returning to it again and again. Trying to play it out in your mind before bed sounds like good visualization practice. Going back to it as you improve and seeing what you can see that's new sounds like a good way to appreciate your growing strength.

But just one game sounds very limited to me. You mention feeling that the tesuji's in that game will be so ingrained you'll spot them easily, but you're only studying a few tesuji in a few situations. What works in one game fails in another for thousands of reasons: a stone one space left, a different score differential, etc. No game you ever play will look quite like this singular game. The shapes, the goals, the followups, they all vary so much. So why not return every month to your favorite game, but also browse many others?

Same with tsumego. Some people swear by racing through simple problems. Do that. Some people think you should labor over a handful, exhaustively confirming you really have the answer. Do that too. Some people recommend looking at the answers from hard problems to get inspiration for novel approaches. Sure! One thing I like trying is to take a problem with a solution far too deep for me to read, memorize the answer, and then practice visualizing the whole solution out. Under the stones, or solutions that bounce between different areas are hard for me to visualize, so I keep at it until I can mentally add one stone at a time and recall all the stones I added. Then I start mentally exploring around the solution, trying to reach all the dead ends I would have reached solving the problem myself. It's my way of practicing the mechanics of visualizing paths.

I personally had good results studying many pro games lightly as a mid-kyu player. I certainly didn't understand the games, but I gained a sense of the type of moves professionals play. I'd find myself pausing before playing something very slow and saying "no, wait, a professional would never play here." and then finding a better move. A lot of groups died along the way, but it also helped immensly to have the conviction that my move _was_ reasonable and that there _was_ a followup here that I could find to prove it.

The trouble with solving problems by recognition is that you only need a few features to recognize a problem, and they may not be the crucial ones. Except for easy problems, the main lines of success and failure are generally insufficient to reveal which stones are necessary and why.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

It's an interesting question that has surfaced in my mind from time to time.

I don't know the right answer, not least because I haven't studied go for some 40 years. However, when producing the Go Seigen ten-game match books, I went over every one of those games many times, because my way of working was to gather as many commentaries as possible (sometimes as many as twenty per game). I had to read each of those, which of course meant going through the game in an effortful way. I then had to produce my own composite text, which was even more effortful, and I also went through each game several more times as part of the proofreading cycle. I suppose this counts as a kind of study.

I am certain I learned about go in a much deeper way, both as regards go concepts and game psychology, which may not make me stronger but does add to understanding and enjoyment. I am sure it did me no harm, but whether it's better than having a superficial familiarity with many modern games, I can't really say. However, my impression from the chess world is that the top players go over noteworthy games many times (they are shorter, of course) early in their careers, to the extent that they can still remember the games. Only once they have this foundation in place do they shift the focus on keeping up to date. It appears that weaker players either skip the first stage (trying to go too fast in competitive play) or the second stage if they are studying mainly to enjoy chess more. I see no reason why this paradigm wouldn't transfer over to go, and it is my sense that it is indeed an important way in the go pro world.

One thing I have never done is make any attempt to memorise a game (or a joseki, or a problem). But because of repetition many things do stick. What sticks (for me) in the case of a game is not the moves so much as key positions, or signature positions - not the whole position but the main shapes (chess researchers call this chunking).

To be specific, I think you need to be able to know a game well enough to compose a geographical narrative about it. Even something simple such as "White started an early fight on the upper left with a new joseki, and got an early lead, but failed to appreciate Black's strength on the entire right and so carried on the fight too long. The result was a ko fight in the centre and Black therefore had the advantage when it came to ko threats. This reversed the lead. But then White's hanetsugi in the lower right, a move highlighted by all commentators, suddenly made White thick. Black tried to sit on his lead but was too passive at several points (e.g. the centre right), and so White was able to rely on his new-found thickness to claw back points in the endgame and so won by half a point."

That will show you understand a game much better than being able to replay all the moves.

Ah, thanks for your interesting comments, John. Finding the main narrative of a go game seems to be a very nice way to study them. I will surely start looking for this in my games.

A game I have been studying for a while now (more because it's the one that I first started with than any other reason), is the first one from 'Appreciating Famous Games'. Honinbo Dosaku (W) vs Yasui Sunchi.



Mmm... Let's try to uncover the narrative of this game. Black makes a few early mistakes and allows white to take the initiative in fuseki. White then attempts to pull out his very weak stone in the top left. This gives black a wall facing the top right, but white chooses not to walk ahead on the left and instead invades on the right. White loses a ko on the top right and thus stays weak, but gains compensation with his ko threat killing blacks top left corner. In an exquisite sequence white manages to get to safety, even making tenuki once for points. In the end, black's thickness in the center gains him enough to win the game by one point. In my personal opinion, white manages the pace of the game beautifully. Black plays very well in endgame, but it is hard to find black moves that inspire awe in me.

Moves that I have really started to appreciate after playing through the game + variations a couple of times are 21 and 23, 33, 57, 84, 93, 97. Of course I do not have to look up the move numbers, because I have memorized it into late midgame. The moves one person appreciates can be totally different from the next. For me, W21 and W23 punish B14 beautifully and induces the rest of the fuseki sequence. I would have never seen W33, but it surrounds black's corner and makes all white stones work together. 84 is a move that I would possibly omit which would end really bad. Lastly 97, the nose tesuji, is a move that I rarely see myself and I should watch out for.

I also agree that memorization is not the goal of studying games. It's just that when you play through a game and it's variations many times, seeing future board positions makes you see how decisions affect the board 100 moves later.

I'm starting to believe in the value of studying to the point of recognition many problems that are closely related in shape. For instance, you can study the L group, the L+1 group with legs on either side, the L+2 groups with outside first-line descents on either side by the surrounding player, etc. Or as another example, there's the tripod group, the tripod group with a leg on one side, the group with a leg on the other side, etc. Or, just the huge number of related shapes in the Carpenter's Square. When you've seen lots of very similar shapes that all have their own answers, you start to get a better understanding of the value and implications of each stone in that shape.

_________________
We don't know who we are; we don't know where we are.
Each of us woke up one moment and here we were in the darkness.
We're nameless things with no memory; no knowledge of what went before,
No understanding of what is now, no knowledge of what will be.

I'm not quite sure how old these noteworthy games are you had in mind, but I immediately thought of old Japanese players like Shusaku as I was under the impression that Japanese pros still study his games (I seem to recall an anecdote, quite possibly from one of your books, about Kobayashi Koichi having gone through the whole collection 4 times or so). However, when I asked Ko Juyeon, a young female Korean pro if she had studied Shusaku she said no, they'd mostly studied more modern (as in last few decades) games but had done some mid 20th century players like Go Seigen or Sakata. Similarly when I was in Korea On Sojin 7p was studying Go Seigen's games. I hope he has a nice 100th birthday as he's given us so much

The anecdote is not from me. It is my impression that the only very old player widely studied by those on the pro path has been Shuei. Shusaku has always been the poster boy and pros pay lip service to his memory, but in practice don't spend much time on him. Dosaku even less, although there are adherents of the Castle Games. Go Seigen, however, is a perennial favourite everywhere.

But do note that there is a simple question of availability. Old games - singles or collections - are actually quite hard to come by. Until very recently, the games of GSG, and Shusaku were also the only inexpensive collections available (Shusai, Takagawa, Hashimoto and Fujisawa existed but cost an arm and a leg, and the likes of Shuei, Dosaku and Castle Games also awaited reprints). The GSG set in particular was pushed by the Nihon Ki-in, and was reprinted several times in different sizes, even during his career. China and Korea initially relied on Japanese sources. GSG and (a part collection of) Sakata have been available in China for a while, but I think only GSG in Korea. It might be worth adding that commentaries on the games of these players were also more widely available than most, which made them more attractive to study.

It's a completely different scene now, of course. We are spoilt for choice.

Incidentally, I'm sure GoGoD has the biggest collection of GSG games: 881. I know a few more against amateurs exist, but I don't think he'll ever become one of our milligods.

If you replay a game over and over again, there is a point, where it isn't really fun anymore. You know the moves, thought about what you are able to think about and well this won't change by repeating it five more times.

I am really not fond of the usual comments, they never explain the moves, where I would mess up in a game.

I think it all comes down to what you enjoy. Without expert guidance, we aren't able to fine-tune our study efficiency anyways. So doing what is fun to you is more than likely the best way to improve(largely by avoiding burn-out).

Personally, I suggest mixing it up. Have one batch of games where you intently study and another where you just replay a few times and consider whatever peaks your intrigue.

Concerning the anecdote about Kobayashi, with permission I pass this on from John Power:

"About 30 years ago, Koichi told me that he had played through the complete games of Shusaku ten times. There was also a high-level book of L&D problems often studied by inseis (perhaps compiled by Segoe) that he said he had gone through 20 times. At one time I remembered what it was, but foolishly didn't make a note.

"On another occasion, Koichi said his favourite old player was Shuei."