Aside from losing (heh), the thing I find most frustrating about playing Go at my novice level is an inability to read the board. Part of it is due to my poor memory (can't remember when I've placed imaginary stones), but most of it is because I can't think of where the best opposing moves would be.
It becomes a paradox -- I can't predict my opponent's moves, so I can't protect myself from them. This means I can't work with tesuji problems of any post-early-beginner level. I tried my hand at the Kanazawa Tesuji Series, and had to quit after #5 because they required more reading than I can handle.
Aside from practice, does anyone have any advice?
HolIgor: It comes at the level of the pattern recognition. You gain some experience and then tend to build shapes with known properties (stability, thickness, attacking and defensive power). You learn a lot from your opponents, especially in the case when you cannot find what you can do against his shape. Then you use a similar shape in your own games. Soon you gather information about the strong and weak points of the patterns and tend to hit there or to protect there. The process is not very fast but not so long as well.
Actually you cannot read everything. The trick then is to know what is worth reading. This is also a major drawback of computer programs. They can't figure out what to read. Interesting thing with them is that the programmers build libraries of good patterns, thus programs play them successfully and therefore can beat beginners. The weak point now becomes a transition from local to the whole board and programs are extremely weak there.
If you want to solve some problems then http://www.goproblems.com is a good place. Start from 30th kyu and go on. It is a level indicator as well. I know that I solve the problems up to approximately 5k at a glance but further I have to think and sometimes make mistakes. Dan level problems are difficult and my solution is seldom correct.
--Stefan: Agree with this - the best way to turn yourself into a pattern recognition machine is the "solve a lot of easy problems fast" approach. Books like "Graded go problems for beginners" are ideal for this. Spend a limited amount of time on each problem, looking up the answer if you don't find it, and work your way through the book a few times in succession. When you solve a problem by remembering the solution from last time, you may feel that you haven't exactly improved your reading skills, but that's OK: the benefit from knowing the pattern is bigger anyway.
BillSpight: At the start, I would not worry too much about your reading skill, in the sense of working out lines of play. There is another aspect of reading that is much more important: evaluating the result of the line of play. This takes judgment, which takes experience.
You can start developing this skill right away. Look at the whole board. How does it look to you? (Later on you will be counting territories and assessing influence, but in the beginning you can be satisfied with an overall impression.)
Now look one move ahead. What moves look good to you? Imagine that you have played one. How does the resultant board look?
Don't worry too much about your opponent's reply. If he refutes your play, you will learn something. :-)
If you try this out, you will find that an interesting thing happens:
You will start seeing sequences of play. That's reading, too.
Bob McGuigan Kato Masao, a top Japanese player in the second half of the 20th century, wrote a series of books with (English) titles like "Three Move Tesuji" or "Three Move Life and Death". He took BillSpight's recommendations one step further, i.e. you make a move your opponent replies and you make another move. He felt that if you could get to this level you would soon be single digit kyu.
Scartol: On page 14 of Lessons in the Fundamentals of Go (which I finally got my hands on through inter-library loan), I found some more advice on this matter: "Some will say 'Phooey, that much I know already (that one must learn to read patiently); it's just that it's too much bother to actually do it.' Others will say 'Look, I'm still weak at the game; I can't do anything difficult like reading.' So much for these lazy students, let them do as they please. They are not going to get anywhere. They need to be grabbed by the scruff of the neck and have some sense knocked into them."
A bit strict, perhaps, but by playing many many games (mostly losing), I find that my reading abilities have improved. Thanks for the feedback, all.
Artis: I have difficulty with fine reading. It seems that my brain does not want to like to consider individual stones, but rather the lines they form. On a similar note I also tend to play worse when big lumps of stones are involved.
argybarg: This is just speculation on my part, but I believe that the seeing that happens when we are staring at a board and the "seeing" that has to take place in the mind's eye both make use of the visual processing portions of our brain. The problem is that the visual input is so much more vivid and urgent than the tiny residual image we are trying to build step by step. It's equivalent to trying to remember a melody while a different one is playing loudly in our ears. (Or, to be more precise, to imagine changing a melody and listen to it at the same time -- something trained musicians can probably do and most of us can't).
The task, then, is to build up the strength of the mental image, so that it can be seen even over the strength of the visual input. For this I think that solving tsumego with the eyes closed has some power. Mostly it requires a specialized kind of concentration to fade down the retinal image and bring up the internal one.
Usually, the visual input helps in that you're building on the existing 'picture'; adding a sequence of stones to empty points. What you say is true though. That explains why 'under the stones' problems are more difficult to see (whereas for a computer, it makes no difference at all).
anonymous: It's interesting that Hikaru no Go always shows that intent stare at the board, indicating that visual cortex (and NOT narrative scenario spinning!) are intensely involved in Go reading. This reminds me of certain marine shrimp that perform complex "hand eye" coordinations with virtually no brain at all, but extremely advanced eyes that demonstrated binocular AND panning vision sensitive to colors well outside human capabilities. My own best games (by comparison) are intuitive and non-verbal.
MarkD: I agree with Stefan, solving a lot of simple problems fast is the best training. I did an interview with Nakayama sensei (6 dan pro) two years ago and asked him about the best way for beginners to study Tsumego. His answer was "Study a lot of simple problems. Beginners often make the mistake to try to solve a few hard problems. In games you need to see the solution to simple problems fast. For a shodan I recommend problems about 5 kyu level. Try to understand the moves, not just memorize them.". The interview was done in Japanese and is available in a german translation.
see also : How To Study On Diagrams
ufo: Hi, some great advice from "Elementary go series vol 3: tesuji":
- Always have a goal, dont read out at random... (duh? go is to complex for that)
- Find a line of play that works (achieves your goal).
- Work your way back through that line of play. Take the last move of the opponent and change it. Does your solution still work? Work your way back all the way to your first move. If you make it all the way back, you have read out the problem completely and you can play your move.
- If you find a flaw, change the last of your moves before that and repeat from nr 2.
- Stay consistent with this approach. Use it whenever you are playing friendly games and doing problems... (this garantees consistency in your learning. You dont improve so much if you randomly read out some moves here and there. Your brain will get used to the methodology, and you wont be going over the same branch twice which makes this an effecient way of reading you can keep for the rest of your life!)
Now after trying, some people might say phooey, i can't do that, that is way to hard. Some other people will say that without even trying. Let me illustrate why you are wrong with a little story:
I study the piano. At first i couldn't jump with my hands without looking at them. It goes without saying that sooner or later you end up in some trouble there. So what does one do? I tried playing blind. Gosh no!!! I cant do that, it's way to hard. Even the pro's look at their hands at times! Guess what. I stuck to it. At first it seemed like my level dropped loads. All the things a could more or less play before, i couldn't play any of them anymore. It was hard. It was labor. It took ages to find the right key by feeling groups of 3 and 2 black keys for orientation. But guess what? It improved. Gradually it went faster and more accurate. And now? Now i hit them most of the time right, instantly, almost without thinking about it. The road is still long, but after about 2 years the improvement is already enormous. I now seem to have an internal map of the keyboard, that is not at all visual. The experience is so strange and enriching that i cant really describe it. Probably blind people know what im talking about.
So, what is the moral of the story? I wish that my teacher had insisted when i was a kid that i always play blind. But she didn't. Too bad.
I hope you learn reading with consistency.
Patrick Traill: I agree that the advice in Tesuji is a good summary of how to do reading, i.e. backtracking when one line fails to find the next. The OP mentioned a problem with remembering the stones on the board, but what I find makes it harder is also remembering (a) which moves you have or have not considered at all points in the sequence, and (b) those sequences that yielded optimal results. There may be more than one optimal result, in that locally it may be impossible to decide if you are better of with seki or a ko.
BruceTheHoon: After a long struggle with what looked like a mental block preventing me from making any progress in terms of reading ability, I have realized that I was paying too much attention to the shape of stone groups, to the exclusion of everything else. Turns out I should have focused more on the shape made by empty points (mostly those representing liberties) surrounding the groups. The number of points to be gained or lost can depend on the shape / size of stone groups but the action to watch and predict happens in the space where liberties lie. Empty space around groups has a structure, all empty points aren't equal; everyone encounters the my liberty / your liberty / common liberty categories, but there are many others (e.g. liberties whose occupation does not increase the total number of liberties of the group etc.). Building a personal dictionary of such classifications / patterns of empty points labeled with their classification was a steppingstone for me, and it improved my reading from disastrous to merely inadequate.
Another thing I did wrong and you might want to avoid: don't try to figure out formal methods to decide outcomes (Kageyama warns against these wrt. reading ladders). The idea is extremely tempting for those of us having a certain mindset because intuition tells us that these methods must exist, but they don't: there will always be an apparently infinite vortex of exceptions-inside-exceptions that the method doesn't quite cover, and figuring out whether you face such an exceptional case is usually not simpler than the original task of just reading. Predictably, the idea of the ever elusive method-to-look-at-the-groups-and-see-the-solution-at-once was - and to some extent still is - severely blocking my progress because I had to actively fight the urge to continue working on it whenever I encountered a reading problem, and that can really distract one from playing the game itself.
(ErnestNotGo?): The content of this very interesting discussion goes very well with the theory that Go is even more a thing of the right side of the brain than chess is. Left-side-analysis is part of the process but surely not all of it. Since I learned drawing at high level as a child but started playing Go (as many "Westerners") as an adult I tend to look at things - including Go - from the artist point of view which means recognition of edges, space, compared value, interdependence and "gestalt". Sure, Go is a very complicated game but I have the suspicion that it's only complicated for the left side of the brain. That's the side that allways kicks in and wants to understand everything at once. The right side thinks differently - which means visually - and is far more relaxed. And it's highly train-able. So as any mentaly healthy adult is able to learn to draw accurately I'm fairly sure anyone is able to learn to read in Go as well.