Tsumego/ Discussion

Sub-page of Tsumego

Table of contents

This page contains discussions regarding tsumego, as well as all the discussions from the original page Beginner's Guide to Tsumego, when I WMEed the encyclopidic content into Beginners' Guide to Go Problems.

Terminology and definitions

(Sebastian:) There seems to be a fundamental difference in opinion. Does "tsumego" cover all sorts of problems or is it synonymous to "L&D problem"?

Bill: Tsumego refers to Life and Death problems, as Bob says. The original author made a mistake.

(Sebastian:) Thank you. Is the current statement "Typical examples include: {...}, connection tsumego, and ladder tsumego" then wrong, as well?

Bill: Connection and ladder problems are fairly specialized. Since connection or a ladder may be crucial to life and death, some of those problems might qualify as tsumego problems, as well. But I think the original author meant all connection problems and ladder problems, and that would be a mistake.

Tsumego is the Japanese term to describe "go problems." [1]

[1] BobMcGuigan: Not just any go problems. "Tsumego" are life-and-death group status problems. There are other type of problems such as fuseki, tesuji, etc., which are not tsumego. Study of these problems is also beneficial, but the great benefit of tsumego study comes from improving your perception of life-and-death status of groups and improving your reading ability. SInce there is a well defined goal and a limited number of moves to consider tsumego problems are good for developing reading skills.

Possible classification

(compiled from different pages:)

Tsumego problems can be classified by their educational purpose:

Generic tsumego (life and death)
These problems are designed to give the player an insight into the life and death of stones. They usually involve a minor clump of stones in a corner and depending on whose turn there is only one way to create "two eyes" and live or gouge that one point and kill the stones. These problems are good for learning about what shapes are strong and how to keep your stones alive when you invade a tight enemy territory. See also How to approach a life and death problem.
These are sequence moves that allow you to capture attacking enemy stones or chasing them into a brick wall and taking the whole lot. These moves come in handy in the middle of the game and also in "fights." Not knowing tesuji allows the opponent to trick you into losing stones that are rightfully yours or putting you into a corner.
Yose is literally, "gathering together." These are moves designed to help you tighten loose corners and reduce enemy territory by a couple stones. They are especially handy when you get "sente" at the end of the game and will make the difference between victory and loss in a close matching.
This literally translates as "fighting each other." This is basically fighting tesuji. Very helpful if you have an aggressive style or in dealing with aggressive opponents.

connection tsumego;

ladder tsumego
Although the latter is not really local, there is only one solution.

other bits we might keep in the main page:

Professional tsumego only leaves one correct choice (two at most) of moves that give the player an advantage.

tsume" is the Chinese character for "tightening," "concentrating(as in Juice)," or "clamping down."

Why a Beginner's Guide to Tsumego

Dieter: If this is an article with a proprietary feel, then I'd like it to be signed. If this page is a core part of the Wiki, then I'd like to make a lot of comments and what I think to be improvements. Soon this will generate a lot of discussion. Both ways are ok for me, but I'd like to know what the intention of this page is. For the time being, I have added page links where possible, because even if it is an article, it is on this wiki for a reason.

Drrichstyles: As the originator of this article I would welcome any direct improvements to the article, however drastic. I just wanted to see a more detailed article on tsumego, because I had more questions about tsumego before committing to a program myself.

Discussion of “Two Schools of Thought”

Personally, I like the quick and fast. Everyone should first assess their learning style and stick to it.

Alex Weldon: Two questions. First, where does the "three month" figure come from, and what exactly does it mean (that 3 months after you've learned something, if you keep using it, it becomes natural?)? I feel like my cycle for absorbing things is faster than that, but I'm not a psychologist. Second, why do you say you should pick one learning style and stick to it? Surely those two methods train different things, the fast version teaching instinct and the slow teaching deep reading. Good go requires both. I tend towards fast, but I don't think it's good for me to keep doing problems that way. I've been trying to do the slow version to teach myself patience, and how to really read things out...

SnotNose: Alex Weldon raises good questions that I share. I've often seen it advised that one try to solve go problems quickly and to look at the solution if it cannot be found in a relatively short period. This smacks of self-indulgence and is counter to the situation one finds oneself in in a game. Reading and reading and reading and reading until one is sure of the correct line to a difficult L&D problem upon which the game may depend must be of some value and probably deserves more credit as a study technique than it seems to be given (not on this page, which is nearly balanced, but elsewhere). Drrichstyles: The "three month" figure comes from a book on scientific learning methods and is written by a Japanese neuroscientist. I think it basically applies to learning completely new skills. I included the figure just in case beginners get discouraged along the way, because go and tsumego is really rich in learning experiences but requires an entirely new way of thinking to some people. I solve the same cycle of problems month after month and I still find something new or gain a new insight. I do believe you make more progress if you focus on one aspect rather than try everything at once. Also, if you are like me and get impatient with reading out every possible variation, in the end, the routine will tire you. So I basically mean that you should stick with a program that suits you and is realistic rather than trying to imitate someone else. However, I will mention that there is a limit to how "long" you should spend reading a problem. I mention this because Korean professionals are known for their intensive practice of "reading" using tsumego and other exercises. In international tournaments they make unique and sometimes unprecedented moves but spend relatively little time thinking about moves. I've read about Korean pros forcing Japanese professionals into resignation since they play difficult to read moves that require lots of thought. Part of the advantage of tsumego is that it is not a game where you win or lose. Sometimes making the wrong move in a tsumego problem is more valuable as a learning experience. It is an issue of "quality" and solving a set number of problems quickly and repetitively is a good way for novices to get a good foundation that can be used to improve the "quality" of reading.

Personally, I would be very interested to read something about "go and mathematics" from BobMcGuigan

a third way

jfc: If you read the original blurb for the KoreanProblemAcademy page it says that the original korean books did not include solutions, just positions. A third way is to work easy problems that can be solved very quickly but instead of looking at the solution examine the problem again but very carefully and methodically.

I remember hearing Jiang Jujo say that you should never look at a problem solution. Instead you should just work it out on a board. If you can't get to the point where you are 100% certain of the solution then the problem is too hard. This is somewhat in keeping with the spirit of the KoreanProblemAcademy problems as they appear in their original form.

Arjen: I agree with Alex Weldon. You need both instinct and reading ability. You need instinct to find the right first move and a likely continuation. You need reading to back up your instinct, and, in the worst case, to find the right moves if your instincts fail. Conversely, good instincts make your reading more efficient, because you can "see" more clearly which continuation works and which does not.

When I work on a problem book, first I go through the book quickly. I spend about 10 seconds on each problem just to see if I can find the correct first move. If I get it wrong, I skim the answer without memorizing it and go on to the next problem. This way I can tell if the book is within my grasp, challenging or way over my head. I also pick up a lot of new ideas and moves I would not have considered before.
The second time through I take a different approach. If I can find a likely candidate for my first move within 10 seconds, I try reading it out for about 1-2 minutes. If I still cannot find the right first move within 10 seconds (or get the continuation wrong), I skim the answer again, before moving on. Usually on my second try, my correct-first-move-rate improves drastically.
I go trough the book like this several times, until I can answer about 90-95% of the problems right (including the continuation).

This way I improve both my instinct and my reading skills.

I am not sure about the 3 month cycle, though. If you play go regularly (say, 1 game every day), you pick up things much faster. Also, because the difficulty increases gradually - advanced concepts building on more elementary skills- you get constant practice in these elementary skills. I do find it useful to go through my old problem books once every couple of months. This way I reinforce these basic skills and get a clearer view on fundamentals (whatever they may be). This absolutely improves my view on more advanced concepts as well.

Are Korean players better because they study tsumego?

(Moved from after “Pros and especially insei are on a very strict regimen of tsumego.”:) In Korea, this is even more so, since it gives them an edge in reading ability. This is why the Korean pros have a monopoly on world titles and making new contributions to go theory and the evolution of the game in general.

jfc: while is is universally acknowledged that developing good tsumego skills is an essential component of becoming a strong player, attributing the recent korean dominance of the international Go scene to their greater attention to tsumego is, at best, conjecture.

Charles The bit about Korean pros is probably misleading. Top pros read very fast (30-move sequence in under a second). Essentially all Korean games are played at what would be short time limits for Japanese pros; that means that Koreans regard 3 hours on the clock as normal. The consequence is probably that Korean pros expect to research openings away from the board, Japanese pros expect at least 6 hours clock time and a chance to think quite long at the board in the opening. It would be astonishing if there were really a gap in reading speed.

Tsumego/ Discussion last edited by Sebastian on September 19, 2004 - 23:21
RecentChanges · StartingPoints · About
Edit page ·Search · Related · Page info · Latest diff
[Welcome to Sensei's Library!]
Search position
Page history
Latest page diff
Partner sites:
Go Teaching Ladder
Login / Prefs
Sensei's Library