Performing vs. Improving

The part about tsumego is largely based on a Twitch stream by Nate Morse (telegraphgo) from May 2023, with my own notes.

By practicing tsumego you will know what kinds of moves to look for when a capturing race or a life-and-death situation arises during a game.

Tsumego show you techniques and tricks in a very focussed way. It’s more efficient to learn these things from tsumego than from games where they may or may not come up. Games take a long time, and it can be frustrating to lose game after game because of new techniques and tricks that you are not familiar with.

Whether or not you were able to solve a particular tsumego is not so important. But you have to analyze the answer to understand what kind of tsumego it is and why this particular answer works in this case. If you missed anything, like the main line or even a variation, think about what exactly you missed and why you missed it and how you could find it next time. That way you build up your repertoire of techniques.

Improvement does not happen while you are solving tsumego or while you are playing because then you are just performing at your current level. The phase where you improve is when you look at the answer or the game analysis and are thinking and figuring out things afterwards.

Don’t judge your performance based on what happened while you were solving the problem, but rather on what you did after attempting to solving it, when the actual learning takes place. It is no good if you just look at the answer briefly and move on. Don’t say “I’ve seen the answer, so I know it now. I’m done with this problem; there is nothing important left for me to do for this problem”.

The main point is not how well you executed but how well you learned.

Whether in tsumego, analyzing your own games or replaying pro games, when you are what is supposed to be in the learning phase, don’t just mindlessly accept the information. So in tsumego, don’t just look at the answer and continue, hoping you will somehow build up a shape intuition. When analyzing your own games, don’t just accept AI moves without understanding them. When replaying pro games, don’t just mechanically move through the game.

So the goal is not to perfectly solve the tsumego the first time around. Don’t spend many minutes on it, trying to read out the whole variation tree. That’s because it is not very effective to train the execution process.

Whenever you read a wrong variation you misunderstood the task that the problem gave you. Afterwards, when you analyze the answer, you can think about what that task was and how could you have immediately understood that task. So when you have deeply understood all the tsumego in a book you will be able to go through that book and only read the correct variation for each problem.

See the article “David Bull on the ten-thousand hours rule” for the importance of mindful learning.

When you’re drilling problems, you’re only executing at your current level, you are not learning. That is a good idea when you want to warm up, for example, before a game.

When you are replaying pro games, you don’t perform yourself; instead you are purely in the learning phase.