Learning the rules of go takes about ten minutes. Nobody hides this; it is one of the game’s best selling points, and if you have not done it yet, our beginner’s guide to go will get you there. But most people who quit go do not quit because the rules were hard. They quit in the gap that comes after the rules: you know what a liberty is, you know how territory works, and yet every game feels like being lost in fog. Nobody handed you a map for the part between “I know the rules” and “I know what I’m doing.”
This article is that map, for roughly your first few months. It covers the board-size progression that beginner resources commonly recommend, the point at which each step up makes sense, the three study habits that reliably move people forward, how the kyu and dan ranking system works so the numbers stop being demoralizing, and where to actually do all of this online for free.
One thing before we start, because I would rather you hear it from me than sense it between the lines: I am not a strong go player. I play shogi, I live in Japan where go sits on the same shelf, but I hold no go rank. What follows is not one player’s pet training theory. It is a synthesis of what the established beginner resources and the English-speaking go community consistently agree on, organized into an order you can follow. Where the consensus is a convention rather than a proven law, I will say so.
Why Start on 9×9
The full go board is 19×19 lines, 361 intersections. The commonly recommended path for new players, echoed across community forums, teaching sites, and equipment guides alike, is to not start there: begin on a 9×9 board, move up to 13×13, and only then graduate to 19×19. I have not found a governing body that decrees this as official doctrine, so treat it as strong convention rather than law. But the convention exists for reasons that hold up when you examine them.
Small games end fast. A 9×9 game can be over in ten or fifteen minutes. When you are new, you learn far more from playing ten short games than from wandering through one long one, because every game gives you a complete arc: opening, fight, counting, result. Fast feedback is the whole engine of early improvement, and 9×9 maximizes it.
Small boards force the fundamental skill. On 81 intersections, there is nowhere to hide. Stones come into contact almost immediately, which means every 9×9 game is a crash course in the mechanic that decides all go games at every level: whether a group of stones can survive or will be captured. Players call this life and death, and it rests directly on the liberty-and-capture rule you learned on day one. The 19×19 board, by contrast, is so large that a beginner can drift for fifty moves without ever being punished for a weak group, which feels pleasant and teaches almost nothing.
Small boards keep the decision space survivable. The standard board’s 361 points are precisely what makes go legendary among strategy games; if that scale is part of what drew you here, our go vs chess comparison goes deep on it. But handing a brand-new player 361 choices per move is how you get decision paralysis. On a 9×9 board, the candidate moves are few enough that you can actually reason about them, be wrong, and see why.
So play on 9×9 first, and play a lot. This stage is not a tutorial to rush through. It is where the core instincts get built.
When to Move to 13×13, Then 19×19
The honest answer to “when should I move up?” is that there is no certified threshold, and anyone who gives you an exact rank number is stating a personal opinion with confidence. Community discussions on this question have run for years without converging on a rule. What they do converge on is the idea that the trigger should be how the small board feels, not a number next to your name. So here are feel-based checkpoints instead.
Move from 9×9 to 13×13 when 9×9 games have stopped surprising you. Concretely: you are no longer losing groups without noticing they were in danger, you can usually tell before the counting phase who is ahead, and games have started feeling like they follow patterns you recognize. You do not need to be winning consistently. You need to understand what happened consistently. Those are different skills, and the second one is the prerequisite for the bigger board.
The 13×13 board adds one new problem: the whole board exists now. On 9×9, nearly every move is a fighting move. On 13×13, for the first time, there is enough room that you must choose between continuing a local fight and playing somewhere else entirely. That skill, judging when a local position is settled enough to leave, is the single biggest thing 13×13 teaches, and it is exactly the skill 19×19 will demand of you on every move.
Move to 19×19 when you feel the itch, and slightly before you feel ready. A reasonable sign is that 13×13 games are starting to feel cramped: you find yourself wanting more room to build something before the fighting starts. You will not feel fully prepared, and that is fine. Nobody does. The first dozen 19×19 games are disorienting for everyone, because the opening phase, where you sketch loose claims on a mostly empty board, barely exists on smaller boards and has to be learned fresh. Expect the fog to return temporarily. It lifts.
One more thing worth saying plainly: moving up is not a one-way door. Strong players still play 9×9 for fun and for sharpening close combat. Dropping back to a smaller board when you want fast games is not regression, it is just choosing the right tool.
The Three Study Habits That Matter
Strip away the mystique, and improvement advice for go beginners keeps reducing to the same three activities. None of them requires a teacher or money.
1. Play real games, and lose a lot of them
There is no substitute for playing, and volume beats agonizing. The classic advice in go culture (often quoted as “lose your first fifty games as quickly as possible”) captures something true even if you treat the exact number as folklore: early losses are the tuition, and paying it quickly is cheaper than paying it slowly. Play against humans when you can. Playing against bots has real uses, and the Online Go Server (OGS) offers computer opponents across roughly thirty difficulty levels, which makes them a genuinely useful pressure-free warmup. But bots at beginner settings lose in strange, repetitive ways, and it is possible to get good at beating a specific bot without getting good at go. Humans keep you honest.
2. Solve life and death problems (tsumego)
Tsumego are go’s puzzle form: a local position where you must find the move that kills the enemy group or saves your own. They are to go what tactics puzzles are to chess, and every improvement tradition in the game endorses them, because they train the one skill with no ceiling on its usefulness: reading out sequences before you play them. Ten minutes of problems a day genuinely outworks an hour once a week; reading is a muscle, and muscles respond to frequency. For a beginner, the free app BadukPop is a comfortable on-ramp, with a library of over 5,000 tsumego plus interactive lessons, in a phone-native format that makes the daily-habit part easy. OGS also has problems built into the site if you would rather stay in one place.
3. Review your games, especially the losses
The habit that separates people who improve from people who plateau is looking at a finished game and asking where it was actually decided. You do not need deep analysis. After each game, try to name one move you would take back, and one moment where you should have played elsewhere on the board. On OGS, every finished game is stored and replayable, and the site has an integrated AI review feature that will flag the biggest swings in a game for you (full AI analysis is reserved for paying supporters, but the free tier is plenty at this stage). A caution that experienced players give consistently, and that I will pass along: use the AI to find where the game turned, then think about why yourself. Copying engine moves you do not understand teaches you nothing durable.
Games give you experience, tsumego give you reading, review turns both into understanding. Any two without the third works slower.
Understanding Ranks (Kyu and Dan)
Go’s ranking system confuses newcomers mostly because it runs backwards at first, so here is the shape of it.
Beginners who have just learned the rules start around 30 kyu. As you improve, the kyu number counts down: 25k, 20k, 15k, and so on toward 1 kyu. Past 1 kyu, the scale flips into dan ranks, which count up: 1 dan (shodan, the “beginning rank”), 2 dan, onward to 7 dan at the top of the amateur ladder. Professional players use a separate dan scale, written 1p through 9p, where the gaps between ranks are far smaller than on the amateur scale. If you have read our shogi coverage, the kyu/dan structure will look familiar; it is the same ranking family used across Japanese arts from martial arts to calligraphy.
The genuinely useful thing about go ranks, though, is not the ladder itself but what one step on it means: one rank of difference corresponds to roughly one handicap stone between amateur players. Go ranks are not an abstract rating. They are a unit of measurement with a physical meaning on the board, and that is what makes the next section possible.
Two practical notes for online play. First, each server calibrates ranks its own way; a new account on one server may be placed at a different starting number than on another (OGS, for instance, shows a ”?” until you have played enough games to be calibrated). So never compare raw numbers across sites. Second, watching your rank number refuse to move for weeks is normal and says nothing about whether you are learning. Rank movement lags understanding. Play, solve, review, and let the number catch up on its own schedule.
Handicap Games Are Your Friend
Here is a cultural difference worth absorbing early, because it changes what kinds of games are available to you.
In chess culture, a serious game between a strong player and a weak one is basically pointless for both sides, so mismatched players mostly just do not play. Go solved this problem centuries ago. In a handicap game, the weaker player starts with extra stones already placed on the board, one stone for each rank of difference between the players. A 5-kyu playing a 9-kyu gives four stones, and what results is not charity but a genuinely balanced, genuinely tense game that both players have to fight to win.
Newcomers from other games sometimes read handicap stones as an embarrassment, a public admission of weakness. Flip that around: the handicap system is the reason you, at 25 kyu, can play a real, competitive, mutually interesting game against someone vastly stronger, starting tonight. It is not a badge of shame. It is the game’s built-in matchmaking, and it is one of go’s best ideas.
Handicap games are also quietly excellent training. Playing with stones against a stronger opponent teaches you to use influence and attack; giving stones (eventually) teaches you to invade and live under pressure. Take handicap games gladly, from either side.
Where to Do All This
Everything above is free, and you can run the whole program on two or three services.
OGS (online-go.com) is the natural home base for English-speaking players and the server most consistently described as welcoming to Western newcomers. It runs in the browser with nothing to install, and it covers every piece of this roadmap in one place: 9×9, 13×13, and 19×19 boards, live games, correspondence games that run over days if real-time play does not fit your schedule, bots across about thirty difficulty levels, built-in tutorials and problems, and stored games with AI-assisted review (fuller AI analysis is a supporter perk, but the free tier covers a beginner’s needs).
KGS (gokgs.com) is the venerable alternative, run under the American Go Foundation, completely free and donation-supported, with an archive of pro lecture material and support for over twenty languages. Its culture skews long-form and traditional, and some players simply prefer its atmosphere.
BadukPop (iOS/Android) is the study companion rather than the arena: its 5,000-plus tsumego and interactive lessons are the easiest way to make the daily problem habit stick, since the practice lives on your phone.
A workable setup is simply OGS for games and review, BadukPop for daily problems, and KGS if you find OGS’s vibe is not yours. If you still need the rules themselves, start with the beginner’s guide to go; if you are still deciding whether go is your game at all, the go vs chess comparison is the honest version of that argument.
Then play your first 9×9 game tonight. The fog is part of the path, and every strong player walked into it exactly as blind as you are about to.