I wrote an article about how to get better at Go on Go Game Guru. It's the result of a number of questions we received from readers and several discussions between Younggil and I about how to improve at Go.

Since this is a topic that's discussed often and extensively on L19, I thought some of you might be interested in it.

As usual, I'd be very interested in hearing your comments, especially regarding anything that you think needs to be added or discussed further (even things that might require a separate article).

David

_________________
David

Go Game Guru: Learn Go | How to Get Better at Go | Go Game Shop | Go Boards | Baduk TV

Very interesting, especially the opinion that there is no harm looking at the solution when doing tsumego. There are indeed other opinions that warn not to look at the solution. As a tsumego fan I absolutely agree with you. The comparison of tsumego with sudoku is interesting as well, but for me sudoku is waaaaaaaaaaaaaayyyy easier (even the "difficult" ones). Maybe it is easier to play with numbers than to visualize black and white stones on the board.

Tsumego rulez.

http://senseis.xmp.net/?BenjaminTeuber% ... trong#toc6

One instance I remembered, which explicitly advised not to look at the solutions.

How I've done tsumego for quite a while is to look at the solution once I believe I have solved a problem, or once I have thought quite a bit about a problem and believe that the shape is highly valuable to know. With other problems that I can't solve I'll often put them aside and try them later.

Good article David; I read through all of it. I have a couple of minor punctuation suggestions which you can decide to take or ignore:


I think it would read more smoothly with a comma.


In the article, there are multiple spaces between "games" and "with".


Each of the four subpoints after this headline has slightly weird spacing - three spaces before the dash, one after.

_________________
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.

To respond in turn to each of the three points justifying looking at the answers:

1. How will you know if you got it right?

I of course defer to stronger players, who can study however they damn well please. But if a beginner asked me for advice, I would tell him that "solving" a tsumego doesn't mean finding one line that works, it means going through all the lines that work. If you can't do that, you didn't solve the tsumego. If you do solve it, you got it right. And if you can't solve ~80% of your problems, you should find an easier set.

2. It's fun to know the answer!

It's fun to figure out the answer yourself. Going on Wikipedia to identify "31. Battleship Potemkin director Sergei" is strictly for people with a neurotic streak.

3. What if the author's solution is better?

If you're actually solving the problems (and not just finding one line that works), this shouldn't be a huge problem. Finding the best solution is most important with respect to setting up superior kos, but there failing to find the best line is equivalent to failing to solve the problem.

As for ko threats and aji... these are important, but is it better for a beginner to completely solve a problem, or make a couple of guesses and then choose the guess with the fewest ko threats? Is it better to completely solve five problems, or to completely solve one problem and also count all the ko threats in all the different lines?

4. How will you learn new techniques?

Doing tsumego isn't a great way to learn new techniques (and crossword puzzles are an awfully inefficient way to learn cultural trivia). Tsumego make you competent at applying techniques you already know, and strengthen your reading. If you want to learn a new tesuji or whatever, get the appropriate book.

I'm only being pedantic about this because (i) training your eyes on a solution diagram accomplishes basically nothing, (ii) many people use Looking At The Solution to impose an arbitrary end on the problem-solving process ("Okay, close enough!"), and (iii) looking at solutions is an inefficient use of your time. As an added bonus, when you look at a solution you rob yourself of an epiphany later when you finally solve a really hard problem that was stumping you.

I don't doubt that there are some people who get a huge amount out of looking at the solutions, but for many more it's a mistake.

I think it's just the other way around. Looking at the solution means to find potential blind spots. As David has pointed out, not looking at the solution is for players who are already strong. That means strong enough that they have already internalized all the underlying techniques.

Although the warning that solving means being able to answer all responses is spot on, I take issue with the rest. For one thing, even pros miss lines and so just believing you got every line right is no proof of anything useful. There is reassurance in looking at a solution.

There are also valuable lessons. Telling a beginner to glare at a problem until he sees light at the end of a tunnel is a good way of reinforcing bad habits. He will probably go "if I play there, he plays there, and then I play there, but if goes there I have to answer there, and then - oh where was I?"

Problems are not solved by such bushy tree searches. Restricting ourselves to weak players (i.e. most of us here), they are best solved by identifying the various key shapes and weak points and then superimposing patterns of known sequences on these. Looking at a solution first can help pinpoint the shapes and catalogue the sequences in your brain. With many (all?) of the patterns there is no mystical osmosis required. You can break down the pattern into easily understood chunks. E.g. the tombstone tesuji has three elements: (1) push through and cut, (2) add sacrifice stones, (3) fill the outside liberties. Once you learn that, you can learn the refinements (e.g. part 1 may not be completely applicable, or in part 3 you need to be careful about the order of filling liberties).

I submit that even a beginner could be taught swiftly to solve tombstone problems by this method, but it would take the fabled army of monkeys with a tripewriter, or a genius, to solve these rather long problems with the stare and glare method.

There are also heuristics that can be applied. E.g. "there is often a good move at the 2-1 point". This is backed by research. Some Japanese guy recently calculated that in problems where such a move was possible, it was the right first move about (I think) 17% of the time, and the right second move 9% of the time. Such heuristics are often contained in the texts of solutions.

As I said in my review of Kyushin Tsumego, there is a large body of pros who are happy to recommend peeking at solutions - with the important caveat that you must look at the problem often, so that the patterns get wired into your brain. Even for those pros who recommend stare and glare, we are entitled to ask if what they say is applicable to us. I suspect that they are really addressing a go population (an Oriental one) where they can assume that at least a decent proportion of the readers have already learnt the basic patterns in the way I described for the tombstone. In terms of the same review, they are advising people who are ready to move onto the applications or combat stage. Since, as far as I know, virtually nothing has been presented in English that enables readers to learn the basics (the kata) as above, such advice to western readers is often irrelevant and even harmful. Although we lack books on the kata, by looking at solutions and (importantly) the way they are presented, we can deduce many of the kata ourselves. I don't believe stare and glare is an efficient, or even a plausible, way of deducing the kata, because it induces looking at individual moves rather than complete sequences.

Try learning a karate kata, or dance steps by just watching someone else. It's an awful lot easier when you have a teacher who breaks it all up into small units and shouts out: left foot forward, right arm back with hand to the waist, make a fist, twist and punch. That's how tsumego can and should be taught - the punch bit is especially fun

Good article. I particularly like the advice, "Don't try to win the review"!

_________________
Dave Sigaty
"Short-lived are both the praiser and the praised, and rememberer and the remembered..."
- Marcus Aurelius; Meditations, VIII 21

@DuskEagle, thanks, there are some weird formatting things I need to fix. Trust me, those extra spaces are not in my source text...

@jts, I think John has summed things up pretty well and I agree with what he said. The term 'stare and glare' made me smile .

However, let me make it clear that I'm definitely not suggesting that people should just look at the solutions! I'm saying, once you think you've solved the problem, then look at the solution.

In addition you've either missed what I consider to be an important point, or you're mis-quoting me for rhetorical purposes?

Since not everyone who reads this, will visit our website, I'd like to clarify things. What I said was, looking at solutions means:

A little bit earlier in the article I talked about 'giving yourself a game you can win'. The 'fun' aspect should be understood in this context. I think this is very important on a psychological level. If you view life as a series of games - whether it's work, study or Go - everything goes much better if the games are 'winnable'. Playing an impossible game sucks badly, and it also crushes intrinsic motivation and interest.

Maybe you're exceptional, my impression is that many of the folks on L19 are among the most conscientious when it comes to studying Go. They're hardly a representative group. The majority of Go players I know don't like doing life and death problems. I think that's because they don't know how to give themselves a game they can win.

The 'stare and glare' school of thought makes most people hate doing Go problems. People are wired to move away from pain and towards pleasure... And even with the small percentage of players who can hack that approach, I agree with John that it's inefficient. And Go players love efficiency!

I should add, that as a weaker player I tried both approaches extensively and went with the one that helped me improve faster. Secondly, I do now sometimes solve problems from books that don't have solutions, but I still look at (and learn from) solutions in the books that have them.


There's actually a huge amount of subtlety in killing (or living) a group efficiently and with good aji, especially in a real game. Winning at Go can be viewed as consistently accumulating small gains until you have a big advantage.

For example, here's an interesting and relatively 'easy' Go problem, black to play. Any takers?

Click Here To Show Diagram Code

[go]$$
$$ +-------------------+
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . X X X . |
$$ | . . . X X O O . . |
$$ | . . . X O . O . . |
$$ | . . . . . O . . . |
$$ +-------------------+[/go]

_________________
David

Go Game Guru: Learn Go | How to Get Better at Go | Go Game Shop | Go Boards | Baduk TV

I'll bite:



EDIT:

Interesting ideas Marcus. If you explore your second diagram further, you're onto something. I'll wait awhile before telling you what I think the solution is, so other people can have a go too.

_________________
David

Go Game Guru: Learn Go | How to Get Better at Go | Go Game Shop | Go Boards | Baduk TV

My Solution

OK, so Marcus and Redundant, thanks for having a go . I don't know if this is too boring or if everyone else thinks I'm trying to trick them, but anyway...

Here's is what I think the strongest play is. You both found elements of the correct sequence. And I don't think the first move is at all obvious if you haven't seen this technique before.



I hope this has been interesting and useful to some people. I think strong moves like are hard to learn without looking at the solutions in Go problem books.

This technique is quite generally applicable and there are lots of other techniques like this to learn. You don't know what you're missing out on until you study the solutions provided by strong players.

_________________
David

Go Game Guru: Learn Go | How to Get Better at Go | Go Game Shop | Go Boards | Baduk TV

That was a very interesting item. Thanks for posting it!

Sometimes reading out the different ways to solve a problem reveals a key point like you found here. When that happens, you can try playing that move earlier in the sequence to see if it's tesuji.

This one is fairly subtle though. It's easier to find them when it's the difference between life and death.

_________________
David

Go Game Guru: Learn Go | How to Get Better at Go | Go Game Shop | Go Boards | Baduk TV