Question for Dan players about Tsumego

[gowan]

I agree somewhat with what Bantari said about learning from playing and not needing the weird ones. However, some standard shapes, like variations on "Carpenter's Square", will be difficult to learn just from playing. I see many kyu players and even dan players automatically play the key defense point for the carpenter's square in gote even when it isn't necessary and, vice versa, leave that defense move out and get killed. There are a lot of other standard real-game shapes that are difficult to read out from scratch. So I would recommend studying a basic text like Davies's Life and Death to learn the basic shapes.

As for the "weird" problems, there is a lot to learn from them. You won't find those shapes in your games but you can get a good sense of your weak points in reading. I find that those weird-shape problems really work on reading precisely because they are not recognizable shapes; you have to read it in detail.

I have a Korean life-and-death book which is great study material because I can't read any Korean. I don't even know which color moves first. This is a real-game practice, determining the status for each situation when Black moves first and when White moves first. The book is organized usefully for study because a basic shape type problem is given and the next removes or adds a stone of one color. Now the shape is changed a little but what happens with the status becomes the problem to solve.

[RobertJasiek]

Tsumego varies in difficulty from easy (1 sequence, 1 tesuji or known basic L+D shape) via almost easy (a few sequences) and intermediate (ordinary L+D without known shapes or intermediate tactics) to arbitrarily difficult (complicated L+D or other tactics). I might have solved ca. 6,000 problems and invented hundreds. Although this has been important for learning how to solve easy to intermediate problems incl. tesuji, endgame etc., the by far most important thing is acquiring a general understanding and knowledge of how to solve problems when the solution is not obvious. Without such knowledge, reading can easily remain too complex. For the intermediate or arbitrarily difficult problems, I have acquired too little knowledge because too little knowledge is available for them; therefore, I research to invent such knowledge but it takes many years. So what I am left with is sheer reading effort when the applicable simplification techniques still leave a complex problem. I think that everybody needs the ability for doing sheer reading because everybody meets the too complex problems. Whatever else you learn about tsumego, eventually you must also improve on the speed and correctness of your sheer reading.

[Bantari]

I agree somewhat with what Bantari said about learning from playing and not needing the weird ones. However, some standard shapes, like variations on "Carpenter's Square", will be difficult to learn just from playing. I see many kyu players and even dan players automatically play the key defense point for the carpenter's square in gote even when it isn't necessary and, vice versa, leave that defense move out and get killed. There are a lot of other standard real-game shapes that are difficult to read out from scratch. So I would recommend studying a basic text like Davies's Life and Death to learn the basic shapes.

As you say - you see them play it. Once they do it once, they should not do so again if they study their own games. This is how I learned. And it was more fun than going over tsumego which might or might not ever come up in my games.

The only way something is difficult to learn from playing is when it never comes up. But then you might not really need it.

[Bill Spight]

I just find it so striking how widely divergent tsumego practice is, from not at all to intensely.

I think it's a jump to describe tsumego practice as widely divergent. On the one hand, you've got a massive (overwhelming) amount of strong players saying tsumego is the key to go strength, and on the other hand, you've got Bantari. Sounds a bit like the global warming thing: >99% of scientists say that we are warming the earth, less than 1% say we are not. You cannot conclude from that there is widely divergent opinion on warming.

The fact that Bantari developed his reading skills without tsumego is fascinating (I'd certainly like to hear more), but it's probable that he is some kind of savant rather than an inspiration to us all.

Strong players know their own experience and know what their teachers and other strong players say. That is not to be discounted, but it is opinion, not scientific knowledge. Informed opinion, of course, but opinion nonetheless. Tsumego, while important, has not always been advised as the key to go strength. Honinbo Shusai Meijin advised studying pro games. Hard to beat that advice, since pro games have everything. In one New Year's issue of Kido magazine back in the 1970s, each Nihon Kiin pro was asked to give advice on how to improve at go. Some of them said tsumego, but it was a small minority. The main advice was to play a lot, and there were other variants, such as to get a rival and to play Black with three stones. More specific advice was to play thickly. But again, these are opinions.

Were I to give advice, based upon my own experience, it would be to play the whole board. That's pretty much the opposite of tsumego. Next I would say, sacrifice stones. That's a kind of thick play, OC. I do not recall any tsumego which said, Black to play and lose the corner.

There is a saying that chess is 98% tactics. But even for chess Znosko-Borovsky says, "Do not entangle yourself in a maze of calculations" (How Not to Play Chess). By contrast with chess, go is largely strategy. But when I was learning go, many, if not most high dan Japanese amateurs were fighters, not strategists. You hear that today about high dan Chinese amateurs. Not much has changed in that regard. It does not follow, however, that focusing on tsumego is the best way to improve.

As a sometime student of the brain, I am confident that there are dozens of go skills, most of which have no name. Tsumego improves some of them. But by definition it cannot improve all of them. Why not study everything?

[Elom]

If you want to be a 5 dan who reads better than a pro, do thousands of Go puzzles.

If you want to be a 5 dan who knows every trickplay and has better positional judgement than a pro, go through hundreds of pro games and understand hundreds of Joseki.

If you want to be a pro, quit wasting time with the above. Instead of breakfast, do 100 Go puzzles (and do ten pushups for every one you get wrong). Instead of lunch, read 100 Go books. Instead of dinner, replay 100 pro games. Rinse and repeat for 10 years and you might stand a chance.

Since I assume you're a normal human being like me and therefore not trying to becoma a pro, let's not worry too badly

Jokes aside, your safest bet is to do mostly problems which are not too challenging, and solve lots of them. If you want to change things up, throw in the occassional hard puzzle. Seems to be a method that stood the test of time

Maybe studying is more student-specific than we like to admit.

[Bill Spight]

Well, Bantari isn't the first I've heard of who didn't do much or any tsumego before reaching somewhere in the dans.

I'm still fully convinced it's what I need to improve, but apparently this isn't a self-evident truth which can be generalized to everyone.

Go is a mountain with many paths to reach the top. Some prefer the most traveled path, others a lesser one, and still others to blaze their own trail. Some of the paths are quite direct, some are more circuitous, some are steep and some are gentle, but all of them can lead higher.

There are 360° in the martial arts.

[Uberdude]

Just to add another case. I don't do tsumego as a habit, but wouldn't claim to have done zero like Bantari (really? you never did the play in the middle of a 3 space eye to kill problem as a beginner?). I guess over my ~10 year Go life I've probably done ~2000 tsumego. I suppose I've replayed more pro games, and that number is increasing faster. If I did more tsumego I would be stronger.

[quantumf]

By contrast with chess, go is largely strategy.

I disagree with most of Bill's post, which is a bit worrying, but this statement in particular I think deserves further scrutiny. Bill, you're stating this as a fact, rather than an opinion (even if intended as such), but on what do you base this fact/opinion?

My view, admittedly that of a weak dan, is that while strategic elements might be critical to separate the absolute elite from the merely very strong, it is tactical ability (reading strength) that makes up the overwhelming majority of go ability. Yes, strategic awareness can help compensate for tactical weaknesses, but ultimately, you're going to get involved in games where your opponent will create problems where no degree of sacrifice or exchange is going to help you. You have to be able to outread them to win.

I would suggest that these pros you referred to in the survey were either being kind to their readers who found tsumego too boring, or had forgotten or discounted their efforts to achieve strong reading, and were thinking more about what it took to reach elite strength (strategy).

[tentano]

@quantumf:

I think you might be on to something. Most go players I know would call tsumego very dull. I learned very quickly that people might like to play the game, but doing problems was usually seen as off-putting. It's likely that people trying to promote the game would want to advertise the sweet parts and not the bitter pill.

[Bill Spight]

By contrast with chess, go is largely strategy.

I disagree with most of Bill's post, which is a bit worrying, but this statement in particular I think deserves further scrutiny. Bill, you're stating this as a fact, rather than an opinion (even if intended as such), but on what do you base this fact/opinion?

Sorry. In editing an earlier statement I left out the IMO. But when I was learning go, the opinion that go was much more strategic than chess was what I heard from different knowledgeable sources and seemed to be generally held. I started off believing that, and I still do.

My view, admittedly that of a weak dan, is that while strategic elements might be critical to separate the absolute elite from the merely very strong, it is tactical ability (reading strength) that makes up the overwhelming majority of go ability. Yes, strategic awareness can help compensate for tactical weaknesses, but ultimately, you're going to get involved in games where your opponent will create problems where no degree of sacrifice or exchange is going to help you. You have to be able to outread them to win.

As for reading, it is not just about tsumego. Tsumego has the advantage that the goal is fairly clear, even when not stated (which, as a rule, it is). But, as Sakata made a point of saying, reading in general requires judgement to assess the resulting positions. It also requires the ability to choose candidate moves, to prune the tree and to include good options. Tsumego does not develop judgement, and the choice of moves is also more restricted that in general in a game.

As a 4 kyu, about the only thing I was at all good at was sacrifice, in addition to whatever unnamed skills I had developed. My life and death sucked, my tesuji was almost non-existent, as was my shape. I knew nothing of opening theory and had picked up only a little joseki. I had a feel for sente and gote, and for the whole board. (As it turned out, I also had a feel for the size of endgame plays, but that had not been tested by problems yet.) Still, if we take 30 kyu as the floor, I had advanced some 26 stones. Since then I have advanced only 8 stones. That despite studying and practicing the usual things. As I said, I would like to credit tsumego with 4 stones of that, but that's too much.

As for your opinion, I expect that it accords with your experience. My experience is different. To be sure, you must have some tactical skill to back up your strategy, but it is easier to win a won game than to come back from a losing position.

I recall when, as a 4 dan, I took 3 stones from the US champion. In the open corner we played a large joseki in which he took the corner at the expense of getting some floating stones in the center. That left me with an easy strategically won game. Near the end he found some tactical probes that surprised me. But, since I was good at sacrificing, I gave up a few stones or points and held on to the win. (He was not happy about the loss. A friend told me that he gave the local 4 dans 4 stones. ) He was about the tactically strongest amateur that I have ever faced.

When I was a weak 3 dan I took 2 stones from a 5 dan. I made a wall in the top right corner. In the bottom left corner he also ended up with some floating stones. This game was not so easy, strategically. I attacked his floating group and built up frameworks. Which he invaded, and proved that I was no killer. In fact, a couple of times I felt like he was insulting me by his invasions, but he managed to live in each case. He definitely outplayed me tactically. OC, I took sente and renewed my attack. In the end his group ran into my wall and died a glorious death for the fatherland. I was lucky to win, I suppose, especially since I was on the borderline of 2 dan. But this game is about the purest example of strategy vs. tactics that I have played.

Edit: Another game I recall, from a few year ago. My opponent, a strong player, invaded my corner at the end of the middle game. I could have let him make small life, but I thought that I could kill. The game proceeded as I had read, but when the time came for me to make the key play, I saw that I could not, because of damezumari. How embarrassing! Not only did he live, but he captured a number of my stones. OC, I took sente and made some territory in the center. I only won by 10 points.

OC, none of this proves the superiority of anything. But that is my point.

I would suggest that these pros you referred to in the survey were either being kind to their readers who found tsumego too boring, or had forgotten or discounted their efforts to achieve strong reading, and were thinking more about what it took to reach elite strength (strategy).

Well, we are talking about nearly all of the Nihon Kiin pros at the time. I have heard tsumego proposed as the one true path only in recent years, and mostly by amateurs, but I am sure that a number of pros stress it, as well.

I have played contract bridge professionally -- best results, winning a nationally rated charity game and 11th overall in national pairs event -- and I would give similar advice to that given by the go pros. Play against stronger players.

[Boidhre]

Well, we are talking about nearly all of the Nihon Kiin pros at the time. I have heard tsumego proposed as the one true path only in recent years, and mostly by amateurs, but I am sure that a number of pros stress it, as well.

I have played contract bridge professionally -- best results, winning a nationally rated charity game and 11th overall in national pairs event -- and I would give similar advice to that given by the go pros. Play against stronger players.

There's probably selection bias going on there I think. If you're a Nihon Kiin pro you got to devote a lot more of your time to playing games than any amateur could and you also from a fairly young age had ready access to extremely strong opponents. Tsumego may not be needed if you can play very strong players a lot but this might be an impossible condition for most amateurs.

Idle thought.

[Bill Spight]

Well, we are talking about nearly all of the Nihon Kiin pros at the time. I have heard tsumego proposed as the one true path only in recent years, and mostly by amateurs, but I am sure that a number of pros stress it, as well.

I have played contract bridge professionally -- best results, winning a nationally rated charity game and 11th overall in national pairs event -- and I would give similar advice to that given by the go pros. Play against stronger players.

There's probably selection bias going on there I think. If you're a Nihon Kiin pro you got to devote a lot more of your time to playing games than any amateur could and you also from a fairly young age had ready access to extremely strong opponents. Tsumego may not be needed if you can play very strong players a lot but this might be an impossible condition for most amateurs.

Idle thought.

Almost all of those who proposed playing against stronger opponents suggested three stone games. For DDKs I would suggest 5 stone to 9 stone games. It's not too hard to find players in that range online until you get really good.

[Boidhre]

Almost all of those who proposed playing against stronger opponents suggested three stone games. For DDKs I would suggest 5 stone to 9 stone games. It's not too hard to find players in that range online until you get really good.

True.

[DrStraw]

I think you might be on to something. Most go players I know would call tsumego very dull. I learned very quickly that people might like to play the game, but doing problems was usually seen as off-putting. It's likely that people trying to promote the game would want to advertise the sweet parts and not the bitter pill.

When reading this I thought of an analogy in mathematics. Learning one's multiplication tables is a chore for many, much like tsumego. But those who master them at an early age find mathematics much easier later in life.

[tentano]

I think nearly everyone I know sees maths as 100% bitter pill, without any sweet parts.

One of the most useful skills I have though.