DieterVerhofstadt/Ideas on improvement

Paths: <= Teaching Methods =>   ·   <= Study =>
Sub-page of DieterVerhofstadt

Table of contents


When it comes to improvement, there are four categories of players:

  1. Those who are satisfied with their level , are not prepared to do what's necessary to improve and play for fun.
  2. Those who are not satisfied with their level, but are not prepared either to do what's necessary to improve.
  3. Those who are not satisfied with their level, who ARE prepared to do what's necessary to improve but do not know what that is.
  4. Those who improve because they both know what should be done and do it.

Many players are in the second category, whether they know what it takes to improve or not. They just don't want to do that and so they should be satisfied with their current level.

Many other players are in number three. I am writing this article so that people of category 2 and 3 should know what it takes to improve and then can decide whether they want to become happy players of their current level (cat 1), or people who improve (cat 4).

I am not a great Go player so what I write is aimed at lower kyu players. It is quite probable that to become a higher dan, something else is needed than what's below. What I do know, is that I would have become 2d much faster if I had applied it.

The keys to improvement

If you want to improve in the game of Go, and this I have distilled out of many other teachers' pages, there are four major things you have to do.

  1. Improve reading skill.
  2. Get a source of new ideas.
  3. Play games, fast and slow.
  4. Review games.

Here's more on each item.

1. Improve reading skill.

1.1 Tsumego

The main instrument for improving reading skill are tsumego. Exercising tsumego will also enhance intuition for vital points of groups, for possible actions against groups, for status of groups. This kind of intuition is developed by doing many easy tsumego.

Tsumego go one step further than ladders. In ladders, you read very deeply (say 30 moves) but do not branch at all. In tsumego, depending on the level, you read 3/5/10/20 moves deep and you branch into a few variations at each move. In whole board fighting, the branching is much broader and there is usually not one single solution. So, when tsumego exercises are done, one has to apply the acquired skill to the game. This is where many people go wrong: they become excellent problem solvers, but fail to extend their ability to the game.

It is a matter of discipline. "I'm going to read this like I would read a tsumego". It is arguable whether every single move is a problem setting. As a first approach, apply your "problem solving mode" to fighting where the life of groups is at stake.

1.2 Remembering games

Remembering your own games after they've ended, becomes a second nature to every strong player. Professionals can often remember games all their life. Actively trying to remember games, is another instrument to improve Go memory. It also serves as a test for understanding the game, as remembering a game becomes easier as patterns fit predictions.

1.3 Counting the score

Positional judgment is an important strategic tool. In the later stages of the game, it boils down to calculating the score. Strong players can score the game very accurately even long before it has ended. Working this aspect of go is yet another instrument to improve the more rote side of Go memory.

2. New ideas and knowledge

Of course, if you want to improve, you need to acquire knowledge. The best thing I think is a teacher, because he or she will know best what your game is lacking and how much new stuff you can handle. The teacher should give you a few new ideas and insist that you exercise them in your game. Replaying professional games can also serve as a source of new ideas.

Many teachers claim replaying professional games is an important step to improve. I have been on and off about this one. I believe it can be good for those who are aiming high. It is definitely a source of new ideas, often through the much undervalued learning method of imitation.

Mistakes: people without a teacher usually read too many books. Secondly, and consequently, they fail to consciously apply the new ideas in their games. Also, some spend hours trying to understand the pro game, which is well beyond reach for players below 5d. It doesn't harm your game, but it is time consuming and inefficient.

3. Play fast and slow

Even more obvious is the fact that you should play to improve. I know many players who just played an enormous amount of games and improved very fast. But this is also time consuming and out of reach for many people, particularly those with a working schedule.

In order to make effective use of the time invested in playing, one should do two things:

  • play fast games to test your intuition
  • play some slow games to test your reading ability and comprehension of the game.


  • many players get worked up about the final result; positive feedback through the game result is an easy mistake to make. In Go, many things can lead to victory or defeat. One should try to find positive feedback through the review.
  • many people play games, applying the same attitude, the same old ideas, without any intention to correct their mistakes. Which takes me again to the review:

4. Reviews

In my opinion, this is the activity with the heaviest impact on one's improvement. With your teacher or someone else or alone, review the fast games and try to find which of your intuitively played moves gave a disadvantage. You can do some tactical analysis, even if you didn't in the game, but not too much: the purpose is to find flaws in the intuition.

At the same rate, review your slow games and dive into tactics. Where did you read badly ? Where did you NOT read !? Which principles, ideas or concepts did you apply ? Where did you forget to apply which principle ? Where did you misjudge a position you anticipated in the game ? Identify your mistakes and try to consciously avoid them in your next (slow) game.

Incidentally, it is a good idea to occasionally review games of players just a little higher ranked than yourself. It will be easy to spot mistakes, which can psyche you up a level.

Mistakes: Reviews are often carried out randomly, depending on what the viewer thinks is interesting about it. Instead teacher or self reviewer should focus on the aspects one is practicing at this moment and on basic mistakes that should not occur anymore. In particular I've seen too many reviews that glorify the knowledge of the reviewer or even drift into pointlessly long variations that create another game.

Kindly read my article The philosophy of mistakes

The allocated time

Here's a schedule I propose for an ideal student. The only thing you have to do when you have not the necessary amount of time, is take ratios.

  • Play one or two fast game a week. 1 hour per week mostly.
  • Play two to three slow games a week. 4 hours per week.
  • Analyse the fast games during half an hour. Analyse the slow game for one hour. 2 hours per week.
  • Do a couple of very easy (3 moves deep) tsumego a day. 15 min a day, 1,5 hours a week. Do 2 or 3 difficult tsumego a week, should take half an hour in total.
  • Have an interview with your teacher, read some articles in a book, on the web, replay a pro game etc. Not more than 1 hour a week and 2 or 3 new ideas.

This gives a total amount of 10 hours. If you have 20 hours to spend, make it double. If you have 5 hours to spend, cut every aspect in half.

To make a comparison, imagine you spend 10 hours the following way:

  • Once a week, you spend 4 hours playing two long games.
  • Every evening you spend one hour reading some Go book.

You may even try to maximize the effort and apply the ideas in the book to your two games, but:

  • You don't analyse, so you won't know whether you did, let alone correctly.
  • You don't analyse, so you keep making the same old mistakes.
  • You don't do tsumego, so your new ideas will get lost in failure to read and bad technique.

An important note about level

When speaking of level, I deliberately do not speak of rank. While rank is an indicator of level, preoccupation with rank too often distracts of the proper goal: improve. People who care about their rank will try everything to win or not to lose (which is the same thing but quite another approach) and tend to ignore the value of lost games, namely the analysis.

Positive feedback

One day I'm going to rewrite this article but for now I want to make an addition about positive feedback. The problem with improvement in Go is that in absence of a teacher the only real positive feedback you have is the game result and its accumulated effect in your rank. Obsession with rank is a perversion of a longing for positive feedback. Players who don't get a pat on the back for a good move, will become frustrated if their efforts do not immediately bare fruit in the game result.

I've been looking for alternative ways to get positive feedback. The main reason why the game result is poorly linked to improvement efforts is that efforts in one aspect of the game are cancelled out by many other factors. A first way of dealing with this is to tackle those factors first which have the biggest impact on the game result. I wrote a page about /basic laws of gamesmanship for this.

Once such high impact issues are dealt with, one can concentrate on one to three aspects at a time and make a checklist for those. The idea has been explored at compass. Again, it's important to identify the area with the highest impact on the game result. As an example consider the following strategy:

A checklist for positive feedback can then be

  • did any big group die
  • did I regularly check the score
  • did I consciously switch to endgame mode
  • how many endgame tesuji could I apply
  • did I effectively end up within 5 points of my opponent

This strategy focuses on balance and making the difference small. Positive feedback has increased probability because losing small is also good. The checklist can be another albeit weak source of positive feedback. More importantly, I believe that such a strategy is the one with the highest probability of yielding wins. It also focuses on one source of ideas, the great endgame exercises book by Bozulich, which holds a rare programme for improvement in itself.

Other checklists are conceivable.

  • If you're a slack player, count the number of times you've taken sente
  • If you're weak in punishing weaknesses, try killing big groups (which is difficult) and count that score
  • If your groups often die or live small, count the number of groups you end up with
  • If you tend to play dame, do the same but give opposite scores

etc etc

Bottom line: without a teacher, you should come up with credible positive feedback for yourself. Focus on improvements that have high probability to produce positive results, as this is the most primitive positive feedback.

See also

Paths: <= Teaching Methods =>   ·   <= Study =>
DieterVerhofstadt/Ideas on improvement last edited by Dieter on February 5, 2014 - 16:02
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