Help, 23 kyu would like to get strong, but don't like playing losing games

So I’m 23 kyu or thereabouts. Basically I’ve probably played a total of less than 50 serious real games. Its one of those things where I have known about Go for years (ever since 2015 when the first deepmind thing came about via news that it beat Fan) and spend some money and got a very nice Go set from GoGameGuru before they went bust, signed up for a years subscription of Internet Go School but was honestly turned off by the site due to heavy accents, lack of a app, and in general bad design and the spaced repitition thing didn’t work for me…

Long story short, I am not getting any strong. I do some tsuemgo go app on the pro app thing but I can’t even mentally solve the “basic”/“easy” puzzles in my mind without having to actually click through and kinda brute force them… not to mention many of the problems are in fact buggy and wrong… (and sometimes more than one solution but it only accepts one solution) as for books I’ve read three of the five volumes of Lear to Play Go from Janice Kim and I watched some dywrin videos on ‘fundmantals’ etc

So I want to get stronger, first goal is to get to 20kyu, then 15kyu, then sdk, then 5ku, and eventually dan level (1 dan)…

I’m old man at age 33 so I know that 1 dan is my max ceiling even if under perfect ideal circumstances.

But here is the thing I don’t like playing with humans and losing…

So is there any other way I can get strong? I know AQ recently have a app bot for the phone but it seems not that great. Leela Zero is pro level but the problem is that its weaker networks are terrible at ladders, groups, L&D and still it doesn’t play like a human…

So what are my options to get strong without having to play against humans and without having to review with humans? I don’t believe in the lose 100 games mantra and the whole “dont read books until you get to sdk” advice. I think I prefer a much higher review to play ratio, like one to three, rather than most people are doing which is 1: 10… but how to review when most of the bots don’t explain in plain english? Reading a much of winrate % isn’t the same as someone telling you what you did wrong.

There’s a perfectly antisocial way of improving at Go:

  1. Memorize and replay pro games. The more, the better. At least one per day.
  2. Do 100-200 tsumego a day. And when I say “do”, I mean “force yourself to hallucinate the stones in sequence to solve the problem”. Do NOT click through all possible spots, that’s a surefire way to not improve at all. Your brain is lazy, yes, and if it stays that way, you will stay weak.
  3. Play the strongest bot you can get your hands on. Stop whining and play.

People resort to all sorts of excuses to avoid facing the fact that hard work is required to improve. You can have a million lessons with a pro that will make you feel all fuzzy inside but if you don’t actually, regularly exercise deliberately and tenaciously, you will still suck.

Don’t be one of these people. If you don’t want to play real people and if you don’t want to put in the hours, just quit while you’re ahead. Do something useful with your time, learn to fish instead.


Hard question. I think I’ll settle for just trying to answer a tiny little part of it:

I see two possibilities here. Firstly that the puzzles are incorrectly categorised. That’s particularly possible if they are aimed at someone who has grown up playing Go with grandparents from the age of three. Some of us were just born in the wrong country. The second possibility is that you need to focus more on learning general principles. If you don’t know the difference between a real eye and a false eye, you will still be able to brute-force your way through many simple puzzles but understanding the principle gives your thinking a shortcut.

ps. I don’t think Janice Kim’s books are very good and I’m 99% self/book taught. See:

1 Like

Why not play teaching games?

You don’t have to lose those. When I do teaching games and the student is getting behind, I turn the board around.

Then you can learn from experience of someone who can explain it without suffering that dreaded “you lose”.


Good idea but then I would have to find a teacher that was able/willing and also overcome any logistics in terms of language, time zone etc

Oh, that’s easy.

You just ask here.

I’m up for it :slight_smile:. I know for a fact many others are too: usually “I need a teaching game” posts are replied to within hours.

What’s worked in the past is setting up a correspondence game, and playing it semi-live: when both people are online at the same time the teaching game proceeds quickly, and when they aren’t it goes a little more slowly.

Send me a friend request if you want to take me up on it, or post an “I need a teacher/teaching game” request to get more options.



I can +1 the suggestion for the Level UP books - I am going through these and enjoy it a lot; they have helped my reading quite a lot so far. The idea is that you see many problems with the same theme / underlying technique, so you get a lot better at recognizing them.

“graded go problems for beginners” is also pretty good.

What’s the issue with losing ? Is it that you don’t like to lose (can’t help with that), or is it that you don’t like that the game is too long while you know you’re falling behind ? If it’s the latter, playing 9x9 at first may help a lot. I’m about 15k and only now switching to 19x19, though maybe I should have done that a few kyus ago (but IMO under 20k I would recommend playing mainly 9x9).

1 Like

Let’s be honest here, you don’t just want to lose against humans, you dont want to lose at all! AI or not!

I understand that perfectly, but not losing also means not playing for fear of losing and you are restricting your growth. I suffer from that as well, so I stagnate a lot at my current rank though I actually did suffer to get to where I am, though I have always had a very good win rate so 40% loss rate is still manageable though if you hate losing as much as I, you should know every loss is painful still.

Lucky for you losing 100 games and not reading books is pure BS. Reading books from the start is what got me to SDK in a month or two without ever reaching 100 losses, or 100 matches to say the least. I played little (because I hated to lose) but I studied enough to make up for it. Quality over quantity.

Unfortunately you seem to give off a vibe of someone who hopped onto the Go train and is likely to be a fad for you, seeing your rushed intense initial investment into expensive game materials without even starting to know the game yet, having lessons but refusing to make good use of them because of trivial excuses (yes, not reasons) and wanting to off the bat be Mr Unbeatable. All are signs that don’t point to realistic expectations and are indicators of a short lived interest that is soon to die unless you change that thinking.

If you want to be strong, put in the work. If you have any sort of talent, or work hard enough to overcome that, you may bypass the conventional sayings. If you don’t even do any of that, forget it.


Get used to losing now, cause it never stops no matter how good you get.

Also you learn more by losing than by winning. Losing is more likely to trigger an emotional response. Emotional responses, when tied to events, makes them stick in your memory much more permanently. That’s why women tend to remember EVERYTHING.


What would you recommend as the best way to study tsumego? Do you use an app/site to do problems, or do you just read through them in a book?

Doesn’t matter where you do the problem, as long as you read the problem out completely in your head before trying to solve it. Not just brute forcing through problems.


It indeed doesn’t matter where you get the problems from, as long as you train your vision. If you want to, you can also visualize just n stones in a row for some beginner practice. As I’ve said elsewhere, memorizing and replaying pro games will train your intuition about possible alternatives in a given situation (a.k.a. shape points) whereas to see if any or all of these candidates actually give you a good result, you have to force yourself to visualize candidate > candidate reply cycles, also known as ‘reading’.

If you give up on a problem and start to click random candidates, you reduce the number of candidates and stones you have to visualize, making the problem easier on your lazy brain and preventing you from making progress. It’s similar to weight training - if you never lift heavier things, even if it’s just a few repetitions, you won’t be able to lift those 200kg or whatever, at any point.


I started with Tsumego Pro but got turned off when I tried to buy the “All-Access Pass” (for $9.99) and the app didn’t give me anything at all. So I switched to GoProblems instead. I think it’s much better.

But don’t just randomly click through problems. Don’t play any move until you think you have it solved. Consciously compare the problem to previous ones that you have seen. Is there something about the board position that looks familiar?

You say you don’t like loosing to humans and only want to play bots, but answer me this… when you play on OGS are you playing a human or a bot? The fact of the matter is there is no difference, so just play and pretend they are all bots.

Should you play really strong bots? I don’t think so. If you wanted to learn to dead lift, always trying to lift 500kg would not help you at all. You have to play opponents that are a just a little better than you. Strong enough so they can beat you but weak enough so you can understand their moves. Playing strong people in teaching games, where they intentionally “dumb down” the game to your level would probably help too.

Lastly, review every game you play. This is one area where strong bots can help. I use Go Review Partner. Treat every game as an exam and every move as a question on the exam. Whether you pass or fail (win or loose,) you still go through each question and compare your answer to the correct one. If you can’t figure out why the (more) correct move is better than yours was, then ask here or in some other forum.

1 Like

i would say sadly…losing still the best way to learn.

actually losing is a better way to learn than wining :slight_smile:


The only way not to lose is to play people at least 3 stones in rank below you.

There is a table published from data here that shows the percentage chance of winning vs rank difference. I can’t recall the exact numbers, but to have a reasonable chance of not losing, you need to play 3 stones up I think.

And you can’t rank up significantly without playing people equal or higher to you (because rank increase from wining games of 3 stones below you is almost nil).

Therefore you have to play games against people equal or slightly higher than you to rank up.

Therefore, if you want a high rank (if that’s what “get strong” means) you have to play losing games.


Or you do it like megatrue and do none of that. :smirk:

1 Like

what is your goal in being adamant not to lose to humans? what do you want to get out of the game if you’re not interested in playing games that you might lose?

Let’s say you play various LeelaZero bots a bunch and eventually, you are able to beat a “shodan” LeelaZero bot about half of the time, so you consider yourself roughly 1-Dan. Then what? Will you continue playing that particular bot, for the rest of your life, winning half your games, but not improving? What value do you derive in that case?

I think more likely you want to get to about 1-Dan level, then beat up on some kyu players. But if you get to a certain strength by only facing one opponent, you are likely not actually that strong. You will learn to exploit the particular weaknesses of that particular version of LeelaZero, and you will not be prepared to deal with an unusual move or trick move that a 5-kyu might play. After playing maybe 10 or 20 games against nominally weaker humans, your rank may drop down to a mid SDK or even a DDK. This would be extremely demoralizing and completely defeat your goal of becoming strong for the sole sake of winning games against weaker humans.

I think you’ll find that if you play go, with the goal of becoming stronger, losing and falling and all the pain that go with those are inevitable. You might work real hard to get to 10 kyu, feel real proud of yourself, then lose a couple games to some 13-kyu or 15-kyu players, have your rank drop several stones, and be miserable. I think if you ask most people who are strong, they’ve had many experiences like this. I don’t think there’s any way to avoid this experience, and it’s definitely not avoidable by avoiding playing against humans, because any other type of training will not help you against unusual moves or trick moves.

tl;dr the pain of losing, and falling in rank, is inevitable if you want to get stronger. So if you really really want to get stronger, it’s probably best to accept now that you are inevitably going to experience a lot of frustration.


I’m thinking maybe would like to find a person who is a high dan (3 or above) who can speak good English without accent who can skype and teamviewer with me to help me in person one on one in training/learning games. I can pay them for this. I think for me this is best way to learn. Instead of stumbling around making same mistakes for my first 100 games, would be much better to get instant feedback PER MOVE as to why this is a good /bad move and why I can/can’t play there/here and what is the best way to counter a paritcular move etc etc and then get every game reviewed… so I’ll play very few games, but the learning ratio per game will be high quality. Blindly playing many games is not my style.

1 Like

I heard an interview with a gold medalist from the Olympics who never was a big talent in his youth and always came second or number 15. When asked about how he kept his motivation up during this period he said: “I knew I wouldn’t care about those losses the day I won my first international competition at a senior level, and I knew I would get there”.

When you’re 15 kyu, you won’t care about your losses as a 23 kyu. When your 1d, you won’t care about your losses as an 8 kyu. Just play and know that you will improve.


You ruffled my feathers a bit here, @Bitt.

Remember that many people are never going to win. This guy got to tell his story just because in the end he won. I bet there’re hundreds of people as persistent as he is who never won anything big and because of that no one asked them to tell their stories. And many of these people also thought they “would get there”.

You will improve but at some point you will stop improving. It’s a fact. And it might be extremely frustrating (or not).

“I hate go.”
Cho Chikun