Go vs chess: the short answer
Chess is a battle to trap one piece. Go is a negotiation over the whole board.
In chess, two armies start face to face and the game ends when a king is trapped. In Go, the board starts empty, players take turns placing black and white stones, and the winner is whoever surrounds more of the board at the end. Stones never move once placed; they are only removed when captured.
Go’s rules fit on an index card, yet the game is so large that it was the last classic board game where humans beat computers. Chess has more rules to learn but a smaller game tree. If you want a game you can learn in ten minutes and study for a lifetime, both qualify; Go is just more so.
How Go works, for chess players
- Placing stones. Black moves first. On your turn, place one stone of your colour on any empty intersection (point).
- Liberties. The empty points next to a stone, or to a connected group of stones, are its liberties.
- Capturing. When you take a group’s last liberty, the whole group is removed and kept as prisoners.
- No suicide. You can’t place a stone where it would have no liberties, unless it captures something.
- Ko. You can’t make a move that repeats the previous board position.
- Scoring. When both players pass, each counts the territory they surround (plus captures or stones, depending on the rules). White gets komi, extra points for moving second, usually 6.5 or 7.5, so a game can’t be drawn.
The white stone has one liberty left (marked).
Black fills it and the white stone is captured.
Ko: the rule against endless recapture
Black captures the white stone on the left by playing the marked point.
Retaking at once would restore the previous position, so White must play elsewhere first.
Without the ko rule, the two players could capture back and forth forever; our engine confirms that White’s recapture would recreate the exact position from one move earlier. Chess handles repetition differently, as a draw (see draws in chess).
Life and death: why two eyes matter
White cannot play in either empty point: each would be suicide. The group can never be captured.
With one eye, White plays inside it and captures all 7 stones.
This idea has no direct equivalent in chess. In Go, a group that can make two separate eyes is permanently safe; one that can’t will eventually be captured. Much of Go strategy is about making your groups live and the opponent’s die.
Try it: play Go on a 9×9 board
Beginners usually start on a 9×9 board. Click or tap a point to place a stone; colours alternate, captures happen automatically, and illegal moves (suicide or ko) are refused. Try surrounding a stone, or recreating the ko above.
Black to play
- Captured by Black
- 0
- Captured by White
- 0
- Moves
- 0
Uses the same rules engine that verified every diagram on this page. It does not score the game; agree with your opponent when it is over.
Go vs chess, side by side
| Chess | Go | |
|---|---|---|
| Board | 8×8 squares (64) | 19×19 lines, 361 points |
| Pieces | 6 kinds, 16 each | Identical stones, about 180 each |
| Start | Full set-up | Empty board |
| A move | Move a piece | Place a stone; it never moves |
| Who moves first | White | Black |
| Goal | Checkmate the king | Surround more territory |
| Capturing | Land on a piece | Take a group’s last liberty |
| Draws | Common | Rare (komi usually has a half point) |
| Handicaps | Uncommon | Standard: extra stones for the weaker player |
| First moves available | 20 | 361 (55 different by symmetry) |
Chess first moves counted with chess.js 1.4.0. Go: 361 empty points; 55 remain once rotations and reflections of the board are treated as the same move (computed). Rules from Wikipedia.
Is Go harder than chess?
On every measure of size, Go is far bigger.
| Game | Board | Positions (log₁₀) | Game tree (log₁₀) | Avg length (plies) | Choices per move |
|---|---|---|---|---|---|
| Chess | 64 | 44 | 123 | 70 | 35 |
| Go (19×19) | 361 | 170 | 505 | 211 | 250 |
Rough estimates from the game complexity comparison on Wikipedia.
Go is also one of the few games where the number of legal positions is known exactly. In 2016 John Tromp and colleagues finished counting them for the full 19×19 board: about 2.08 × 10¹⁷⁰, a 171-digit number (Tromp). A position is legal if every group of stones has at least one liberty. We checked the method on small boards by brute force:
| Board | Ways to colour the points | Legal positions (our count) | Share legal |
|---|---|---|---|
| 1×1 | 3 | 1 | 33.3% |
| 2×2 | 81 | 57 | 70.4% |
| 3×3 | 19,683 | 12,675 | 64.4% |
| 4×4 | 43,046,721 | 24,318,165 | 56.5% |
| 19×19 | 3³⁶¹ ≈ 1.74 × 10¹⁷² | ≈ 2.08 × 10¹⁷⁰ (Tromp, 2016) | 1.2% |
Each point can be empty, black or white. Our counts match the published values (Go and mathematics). Chess positions can’t be counted this simply, which is why its figure is an estimate; see how many possible chess games.
But bigger isn’t the same as harder to play. Chess punishes a single tactical slip more brutally: one missed fork can end the game. Go is more forgiving of small mistakes, because the score is the sum of many local fights. Most players find Go’s rules easier and its strategy harder to put into words. For chess’s side of the question, see is chess hard to learn?
Go, chess and computers
- Chess, 1997: IBM’s Deep Blue beat world champion Garry Kasparov in a six-game match, mainly by searching millions of positions per second.
- Go, 2016: brute-force search doesn’t work with 250 choices per move. DeepMind’s AlphaGo, which combined neural networks with tree search, beat Lee Sedol 4–1 in Seoul in March 2016 (match record). Lee won game four; the American Go E-Journal noted the result came a decade earlier than many had predicted (AGA).
- Today: engines are far stronger than any human in both games. In chess, Stockfish uses neural-network evaluation too, and it is the engine behind our best-move finder.
History and popularity
Go was invented in China more than 2,500 years ago and is believed to be the oldest board game still played. It is called weiqi in China, baduk in Korea and igo in Japan. A 2016 International Go Federation survey of its member countries found over 46 million people who know how to play and over 20 million current players, most of them in East Asia (Wikipedia).
Chess developed from the Indian game chaturanga about 1,500 years ago (see who invented chess). It is more widely played worldwide: see how many people play chess. Chess’s East Asian relatives are a different family from Go: xiangqi and shogi.
Should you learn Go or chess?
There’s no wrong choice, and many people play both. Some practical pointers:
- Choose chess if you want the easiest time finding opponents, clubs and lessons wherever you live, or you like sharp tactics where one move can decide everything. Start with how to play chess.
- Choose Go if you like the idea of learning the full rules in one sitting, enjoy big-picture planning, or want a game where handicap stones let a beginner and an expert have a close game.
- Start small in Go. A 9×9 game takes about as long as a casual chess game and teaches capturing, ko and life and death before you move up to 13×13 and 19×19.
- Expect different first lessons. In chess, beginners lose to quick tactics and early checkmates. In Go, beginners lose by building groups that can’t make two eyes.
Skills transfer both ways: reading ahead, staying calm in a long game and reviewing your losses help in either. To see how a chess engine judges a position, try our best-move finder or read how to read engine evaluations.
FAQ
Is Go harder than chess?
Go is much bigger: about 2.08 × 10¹⁷⁰ legal positions against roughly 10⁴⁴ in chess, and around 250 choices per move against 35. That is why computers mastered chess 19 years earlier. For a human, both take a lifetime to master; Go’s rules are simpler to learn.
Is Go older than chess?
Yes. Go was invented in China more than 2,500 years ago and is believed to be the oldest board game still played. Chess developed from the Indian game chaturanga around the 6th century AD, roughly 1,500 years ago.
Which is more popular, Go or chess?
Chess, by most measures. The International Go Federation’s 2016 survey found over 46 million people who know how to play Go and over 20 million current players, mostly in East Asia. Chess.com alone has over 267 million accounts.
Are Go rules simpler than chess rules?
Yes. Go has one kind of piece and a handful of rules: place a stone, capture groups with no liberties, don’t repeat the position (ko), and count territory at the end. Chess has six piece types plus castling, en passant and promotion.
When did a computer beat a Go champion?
In March 2016 DeepMind’s AlphaGo beat Lee Sedol, one of the strongest players in Go history, 4–1 in Seoul. In chess, IBM’s Deep Blue beat world champion Garry Kasparov in 1997.
Does chess help with Go?
Partly. Reading ahead and calculating carry over, but Go rewards a different kind of judgement: weighing territory and influence across the whole board rather than attacking one king.
How long is a game of Go?
An average game lasts about 211 moves by the game-complexity estimate, against 70 for chess (each player’s turn counts as one move). Casual games take under an hour; some professional games run over two days.
Sources and how we checked
- Go (game), Wikipedia: rules, history, IGF 2016 player survey.
- Game complexity and Go and mathematics, Wikipedia; John Tromp, Counting legal positions in Go.
- AlphaGo versus Lee Sedol; American Go E-Journal match report.
- How we checked: a Go rules engine written for this page (the same code runs the playground) counts every legal position on 1×1 to 4×4 boards by brute force, 43,046,721 colourings for 4×4, matching the published values. It also verifies the capture, ko and two-eye diagrams and the 55 symmetry-distinct first moves. Chess figures from chess.js 1.4.0.