How do I actually improve?

OGS member since 2019 but I still can’t figure out how to have fun playing.

Whenever I play a game it’s like my opponent gets two turns while I only get one. The only way I can win is if my opponent times out, misclicks or make a catastrophic error. Any game that actually resembles a game of go I lose.
I also play chess and shogi, and for those when I lose a game I can generally figure out where it went wrong. As a beginner even brute force experience helped me. With go, I can’t see what I did wrong even after looking at AI analysis. So I can lose as many games but with zero lessons learned.

I’ve tried to read various Go resources but they are either overly simple (explaining the rules to basic concepts like atari) or extremely difficult.

How do I learn from losing, let alone actually win?

How about finding a teacher? They are exactly the kind of people who can point you to mistakes you cannot discover by yourself.

Apart from that, a change in mindset may help:

You’re not losing every game, you only lose a little more than half of them, but you’re also playing mostly stronger opponents. :slight_smile:

Another way of changing your mindset, is to note that, as with chess and shogi, Go is a perfect information two-player game. As a consequence everybody can only win if their opponent makes mistakes*. It’s just that the mistakes of stronger players don’t appear like mistakes to you (yet). Even professional players and superhuman AI only win their games against each other because the opponent is making mistakes, not because their moves are “better”.

*Assuming the correct value of komi and allowing draws

Finally, I guess that you should consider what your goal is: are you playing Go because you enjoy it? If not, then why do you want to become better? I’m pretty sure that the feeling of making stupid mistakes will never disappear, we all feel that, regardless of how strong we are.


you can create sgf with variations collection
as your collection of openings to remember grow bigger, you will be able to apply more and more bot moves in your next games

and you can solve tsumego of your level


I guess I have one last comment, about studying.

It is extremely difficult to get better at something: it takes effort, study and practice. Don’t let difficulty of a resources get in your way, but see it as something you can overcome with effort. The first time you read a text, you may not understand most of it, or only have a superficial idea of what is meant. But, if you then try to study what it would mean to apply the advice in practice, and play some focused games where you specifically try to apply what you think is written, you will slowly come to get a feel of what is written in the text. Then, you go back to read the text again, and you’ll be surprised that you understand more of it. You return to extracting how to apply it, and then put it into practice, and so on, until you feel you have really understood what is written.

It is certainly difficult to do, because brains naturally don’t like to focus on hard things. But, as with many things, there’s an easier way of concentrated practice. You don’t as much “learn” how to get better at Go, it’s more like you “get used to” being better at Go by finding new strategies and practicing them until they become your second nature.



Which books did you try to read and found extremely difficult?

I’m confused. What’s your motivation to play then?
Do you have someone to play with IRL?

I believe that.
It’s a necessary step to enter this game of complexity.
I usually suggest to use a software that allows you to take back your moves and try something different when your previous move didn’t work.
You could do that with any Lizzie, Lizgoban, katrain and so on, but they use very strong AI so it’s impossible to win a game.
My best choice is an android app called “go free” from “AI factory” developer, which is very straightforward and has many levels of difficulty.
That helped me a lot when I started.

Here you definitely need some human advice. You can have it here in the forum: pick a game of yours that you would understand better and ask for reviews. There’s plenty of people willing to help.

I looked at few of your recent games and I can suggest these key points:

  • try not to be surrounded: you need room to make a live group, if you’re in a cramped space you could be killed. This is especially for 19x19. In 9x9 the board itself is cramped and it’s all about killing or being killed
  • beware of cuts: if your stones are separated, they can be easily harassed and killed. Try not to be cut
  • don’t play in too narrow places: if you invade, you need to live when surrounded, and it’s hard.
  • don’t be afraid of your opponent’s big moyo: if they can build, you can build too. Go is a game of balance: you don’t have to destroy your opponent. You can’t avoid that he has some territory. You just need to have a little more. So share the goban with your opponent: if he takes some, you take something else.
  • look for opponents that are about your strength. They’ll make mistakes too and you’ll have some gratification by killing some groups and winning some games.

I wouldn’t say you’re 13k, but your history on OGS is long, so your rank should be reliable. I’m quite confused about that.


If you are a beginner ignore AI analysis. It’s not useful, and might even be detrimental. You need human guidance. Best would be in person if that’s possible in your part of the world, second best is a live review online exploring variations, third best is text commentary from humans on your games.


If you would like to play a teaching game, feel free to send me a challenge (preferably a correspondence game with analysis on).
While playing we can chat and exchange ideas.


I will second @Uberdude 's suggestion on that. AI usually gives advice that it is hard to comprehend, even for dan level players. It is hardly a “teaching tool” when you are trying to learn the game and break into SDK ranks.

Instead, what I would suggest (since you are keen on looking at the finished games and reviewing what happened) is that you look at what the opponent did and try it yourself. For example:

Let’s say you start a game, the opponent plays a fuseki like san-ren-sei or Chinese and you fail to break into it and lose. That’s fine.
Start another game with a higher ranked opponent (another good habit you seem to have) and you give a try to the thing that you just lost to. Play the san-ren-sei or the chinese fuseki you just didn’t manage to deal with and see how the more experienced player deals with it and learn from that.

Treat the game as a teaching game and set to observing what the other player is doing differently from you. That way it will be like you are having doznes of tutors at the same time. That’s one of the things I did to improve and it works. Pose the problem you have to a better player. You will probably lose the game, but now you have suggestions and solutions to the problem. :slight_smile:

Here I will second @Vsotvep 's suggestion by only adding this, which might help you:

a) Try to study a more difficult book slowly, instead of trying to read through it.
b) Try to find a real Go board (a cheap one will do) and try to play the variations and board positions and moves that more difficult books are suggesting. This will make it more easy for you to visualise, understand and imprint it as a “good idea” or something to try in your next game.

If something doesn’t work or you think is beyond you (e.g. the concept of Sabaki), leave it for later.
Sometimes in order to understand one concept, you need to have practiced a different one first, which you might have skipped or might have not totally grasped.

At the end of the day, @_Sofiam is correct that practice is important, but the truth is that there is a modicum of talent involved. If you think you have reached a limit to how much better you can become, then do not worry about it and try to have fun playing the game at the level you are.

Hope this helps. Good luck and have fun :slight_smile:


It may be that he, like me and our inactive Russian friend S. Alexander, does not find go fun, but does think it is fascinating—an intriguing mystery to be probed.

I think that’s the most important point.

If you want to improve, the first step is to find go fun. If it’s not fun, what’s the motivation to improve?

Your OGS profile tells me that you’ve played 187 ranked games against higher-ranked opponents, and only 127 ranked games against lower-ranked opponents. I also see that you’ve played 107 games against the same opponent, Kugutsu [6k].

So maybe the problem is that your opponents are too strong? You could try looking for opponents that are closer to your rank?

1 Like

Just in case anyone feels intimidated: You can reach 4k without understanding what these terms mean. (Doesn’t mean they are useless or you shouldn’t know them, I have not the slightest wish to start a flame war over such things. I’m just mentioning it because these terms might have intimidated me when I was a beginner.)


Stay positive! Go is challenging in ways that chess and shogi aren’t—it’s part of what makes Go so special. Imagine you’re playing chess and blunder a piece. You get immediate feedback on your move from losing the piece, and you know exactly where and how you went wrong. Now imagine you’re playing Go and make an improper move. You may only get feedback on your mistake much later in the game, and sometimes the consequences may not even be obvious at the end of the game. That’s just the nature of Go. It’s also why Go wisdom is encapsulated in “proverbs” and why professional players speak of “proper” play (honte in Japanese). So here’s some tips:

  1. Get a teacher: If Go is important to you, a teacher can be the quickest way to get to the next level. Even if you don’t get a teacher, you can still gain new ideas from books or YouTube videos (Nick Sibicky anyone?).

  2. Study pro games: Flip through professional games and learn from their moves and strategies. This way, you can judge the quality of your own moves by comparison to how the pros play.

  3. Ask for review: If you lose a game, ask your opponent to review the game with you. Many players will be happy to help point out what you could have done better.

Remember, improvement takes time and deliberate practice. So don’t be too hard on yourself, and keep playing! You’ll get there!


its actually very hard to understand game of someone else. To understand game of someone much better is even harder.
There is no difference from kyu perspective between pro moves and bot moves, both are “too hard”.
But when you do AI review of your own game, you at least may understand short alternative simulation from move N, because you understand your game and when simulation is short, game is still similar enough to your game.

So “study pro game” sounds much more useless for kyu than “do few AI simulations on your game”


I understand your perspective and agree that it can be hard to understand others’ games, especially if they’re a lot stronger. But I feel that studying professional games can still benefit a kyu player. It can show them “proper” play in given situations (honte, again), and it can improve their sense of flow—when to stay the course and when to play elsewhere or change strategies.

AI simulations help only to an extent. They can’t show you the broad view of how Go should be played.

But what worked for me may not have worked for you. Different strokes for different folks, everyone learns differently, etc. At worst, you’ll gain experience, which can’t hurt. But it underscores the other points about finding human connection to evaluate games, whether a regular teacher or one-off encounters with fellow players.


Play face to face games. As pointed before, having someone stronger help with comments is priceless and that is much easier done in real life with board and stones. Online is often underestimating and poor in practice on this side.

Books, videos and such can be very enjoyable but the gap between you and these can be has huge as visiting a painting museum and painting by yourself.


I found Dwryn’s ‘back to basics’ series on YouTube very helpful -
I’ve not necessarily won a lot more games - which I like to put down to spending too little time on screen before placing stones - but my enjoyment of all my games has improved regardless and I have a much better sense of what’s going on - shape, direction of play, which are the big points and so much more. I’ve no experience of Shogi but I used to play Chess and I found Go play more strategy oriented than the latter, where tactical play was more crucial. I prefer the head space - so to speak - playing Go but now I get totally thrashed if I play any Chess. Could it be that you’re bringing a Chess approach to your games of Go?


There are many ways to learn and improve our go skills. As others have replied, play reality game or find a tutor or read more go contents, or do more puzzles and etc. They are all useful, I think we could only figure out which is the best approach by ourselves. I hope the materials that I have made could help you. It is the
I have selected content from Japanese and Chinese go books, there should be content that suited you.
And it is free!