Using AI change a lot the parameters. Even if there is no explicit use of theory and concepts.
Without access to theory and playing with players who don’t access it too, and without AI, my guess is like middle DDK.
I met IRL once a player who only explored the game with a friend and that’s what they were after a few years of playing together. He lacked mostly of distance, like he was pretty good in very local fights.
Same. I think a very talented player could coast to mid-high sdk on raw talent, but for most people, I think they would hit their first wall at mid-ddk, and need some study aids to get over it
I’m not sure. Discovering the use of influence is for example quite a challenge.
I’m sure by playing a lot of 9x9 blitz without review its easy to reach sdk. Then it would take some time to stop fear 19x19…
You have to be sure that the pool of players you play with didn’t have any contact with the go culture at first. I insist on this point, because i met so many so proud of their self study, like hey i never open a joseki book, but they don’t acknowledge that their opponents did and they benefit from this.
Then that they can develop conceptual tools from 9x9 and export them on the big board.
There are a huge mountain of knowledge we access (and even without reviewing) which is going to be pretty difficult to get by experimentation. Take just a one point jump. Use of the corner, starting at the 3d or 4th line… Reducing the size to 9x9 should help, for example in fighting abilities, short reading… but may confuse when looking at the bigger board too.
I’ve known a player in our club since when he started in his early 20s. I think he only plays offline in our clubs and some tournaments, and I don’t think he ever studied much, but he loves to play and he reached dan within 10 years. His progress slowed significantly since he started working and started a family, but it looks like he’s now 2-3d EGF.
But did he review games with his partner after playing them? While that’s not formal studying, it could be considered as an indirect way of studying.
I’ve never seen him do that except maybe a few minutes to check some game deciding moment or some joseki mistake.
I suppose he has a good ability to learn from his opponents during play.
Interesting topic. I wouldn’t have gotten very far up the ranks without reading books and doing problems. I didn’t get very far with them!
On the one hand, I have enjoyed playing similarly ranked players, always, as I ranked up.
On the other hand, it’s nice to learn, rank up, play the game at a higher level.
I think if you pursue the “all on my own” path, you’ll plateau. And then you’ll know what that plateau is and whether you want to start studying or not.
I think one can pick up a lot from playing a lot, but it depends on one’s opponents and their skill level too. ^^
My own progression from first reaching 2d KGS-5d KGS (in several months), and 6d KGS (in less than a year from first reaching 2d) many years ago was during a period in which I played many, many slow games for many hours per day (reading deeply then reviewing thoroughly) & didn’t study really anything in particular.
I found that it wasn’t necessary for me to do tsumego etc., as I learnt from really doing my best in games and reading deeply in “applied” situations ( I did my best to read out even difficult opening situations) – and I’m sure seeing other players’ shapes, styles etc. also helped me get a feeling of what was best in terms of how the stones move, strong shapes, etc.
(At least that’s my feeling of it).
(I’ve talked with other strong players who felt that one doesn’t really necessarily need tsumego if playing a lot either, which is interesting and matches how I felt.)
The most active studying I did at the time was occasional play over sections of very few pro game records I found inspiring/Fairbairn commentaries I found inspiring between the stages of 5-6d. (no joseki, books apart from etc.)
Of course, I had somewhat read some books before reaching 1/2d KGS + been exposed simply to some other players teaching me basic concepts as a DDK/SDK, and how others+dan players think when reviewing before, so I’m sure that helped.
But after reaching dan level KGS or so, most of the improvement was from my own experience and effort.
(playing and focussing deeply during games, self-reviews as it was pre-strong-AI era etc)
If one learns from simply playing a lot and with decent players (e.g. dan-level players from time to time for SDKs, or SDK player if one is DDK), I think it is possible to improve quite a bit.
But even playing 1 or 2 serious games per day and really putting effort into reviewing & during the game, I think can help one a lot.
(and applying things learnt or trying to play to learn seems to help too – trying things one is unsure of to learn from & get experience, learning and playing differently in future games)
I suppose you could say that the stronger players in our club did also mentor him. No formal lessons, but still you’ll pick up a thing or two from it.
That sort of informal/casual instruction/sparrring with senpai was also quite important in my own initial progress.
Definitely. But you can get to 5k without that. Living proof speaking. 
What an interesting training regimen! Next you will tell me this player doesn’t aim to get the most points!
I suppose one could ask how strong AlphaZero was after 2000 self play games (~3-4 games per week for 10 years)… I think the answer is “doesn’t know the rules”.
I believe its not possible for average adult person to reach dan ever by just playing. If someone do, they are not average. Review without AI that is also helpful is not easy to invent without someone else to explain how.
https://forums.online-go.com/t/what-is-the-real-limit-of-being-self-taught/51388/7
What is average?
When you’re a go player, you’re already not an average person.
When you’re 5k you’re already a stronger go player.
Even when you’re 10k I think you’re a better player than most casual players, except you won’t find most casual players playing go online.
Also I think all stronger players did get explanations from other players. When you’re completely on your own, without anyone teaching or explaining you anything and giving feedback, then I think even the most talented person wouldn’t get very far. It doesn’t matter if it’s about playing go, playing the violin or speaking a foreign language.
Even a 15k giving 9 stones handicap to a 25k is a stronger player.
I think the debate is a bit biaised by the fact that go is a 2 players game, a communication between 2 people so to pretend to be isolated from the go world you need a partner who is isolated too. So this is very rare cases (i guess especially online where people meet on go servers)
I know a few people who have played in a club for years, but otherwise don’t study, almost never play tournament games and don’t play on the internet. They are about 13k EGF (so I guess about 10k OGS).