Mastering Go: Opening Principles and Life-and-Death

Go is the oldest continuously played strategy game in the world. The rules fit on a postcard. The complexity is famously deeper than chess — so deep that strong AI play eluded computer scientists until DeepMind’s AlphaGo solved it in 2016. Mastering Go means absorbing a small set of principles (corners before sides, sides before center; good shape and bad shape; life and death) that take years to apply well but that anyone can start using immediately. This guide covers the core.

The basics in one paragraph

Go is played on a 19×19 grid (smaller boards — 9×9 or 13×13 — are used for learning and faster games). Two players, Black and White, alternately place stones on the intersections of the grid. Black moves first. Stones do not move once placed. A stone or group of stones is captured and removed when all its liberties (adjacent empty intersections) are filled by the opponent. The game ends when both players pass; the winner is whoever has surrounded more territory plus captured more stones (with komi — a small score adjustment — compensating Black’s first-move advantage).

The opening: corners, sides, center

The single most important principle for new Go players: play corners first, then sides, then center.

The reason is geometric. A corner is enclosed on two sides by the edge of the board. Building a territory in the corner requires defending only two open directions. A side is enclosed on one side by the board edge — three open directions to defend. The center is open in all four directions. Corners are vastly more efficient places to make territory.

Look at any professional Go game’s first moves: both players will play their first stones near the corners, usually on the 4-4 (star point) or 3-4 (small-knight) point. The 3-3 point invades the corner outright; the 5-4 point claims influence over the center at the cost of pure corner territory.

The opening points

The classical opening points and what they emphasize:

• 3-3 (san-san): Claims the corner with minimal influence outward. Strong for territory, weak for center fights.
• 3-4 (komoku): A balanced corner point that can extend either direction depending on later development.
• 4-4 (hoshi, the star point): Emphasizes influence and potential territory; doesn’t actually secure the corner. Modern professional play favors hoshi.
• 5-4 (takamoku): Faces strongly toward the center and sides; weaker on corner territory.
• 5-3 (mokuhazushi): Side-focused, leaves the corner open.

Opening sequences (joseki) where two players negotiate corner positions are studied in their thousands. You don’t need to memorize joseki to play well — you need to understand the principles behind them.

Shape: good and bad

“Shape” in Go means the geometric configuration of your stones. Good shape is efficient — every stone supports its neighbors and contributes to forming territory or attacking the opponent. Bad shape is inefficient — stones are too close (overconcentrated) or too far apart (vulnerable to invasion).

Good shape examples

• The tiger’s mouth (or empty triangle, but specifically the strong version): Three stones forming a kind of “C” shape that defends a key intersection.
• The bamboo joint: Two pairs of stones forming a connected, hard-to-cut group.
• The diagonal extension (knight’s move and one-space jump): Stones spread efficiently along a side.

Bad shape examples

• The empty triangle: Three stones forming an L without the inside intersection of a tiger’s mouth. Notoriously inefficient — it uses three stones for what two could accomplish.
• The dumpling: Four stones in a 2×2 cluster. All four stones share liberties; one captured stone often condemns all four.
• Stones too close together: Overconcentrated; you’re using more stones than necessary to defend.

Strong play recognizes shape reflexively. The shape vocabulary (tigers, bamboo, ponnuki, etc.) is genuinely useful — once you know the names, the patterns become visible.

Life and death

A group of stones is “alive” when it has two separate eyes — two enclosed empty spaces of one intersection each (or one larger enclosed shape that the opponent cannot reduce to fewer than two eyes). A group with two eyes can never be captured. A group with only one eye can always be killed eventually.

Life-and-death problems (tsumego) are the heart of Go skill. Given a group with uncertain status, can you kill it or save it? The answer often involves a precise sequence of three to seven moves.

Common life-and-death principles:

• If a group has only one eye, look for where a second eye can be made — or for how to prevent the second eye if you’re attacking.
• Vital points — single squares whose occupation either creates or destroys a second eye — are the high-leverage moves.
• Sacrifice stones to gain eye space. The “throwing-in” technique (sacrificing a stone inside the opponent’s potential eye) is a fundamental life-and-death tactic.
• Three-in-a-row inside a confined space cannot make two eyes. Many tsumego rely on forcing this configuration.

Improvement at Go comes almost entirely from doing tsumego. A few hundred problems at your current level, repeated until the patterns are reflexive, will raise your strength more than any opening study.

Influence vs territory

Stones placed near the edges and corners tend to secure direct territory. Stones placed facing the center tend to project influence — they don’t directly own any squares, but they make the surrounding center area easier to claim and harder for the opponent to invade.

The trade-off is fundamental. A 3-3 corner stone takes ~7 points of territory. A 4-4 corner stone takes zero points immediately but projects influence that may turn into 20+ points later if the rest of the position develops favorably.

Professional players balance both. They don’t commit fully to either territory or influence; they navigate the position based on what the rest of the board looks like.

Reading: the calculation side

“Reading” in Go means calculating sequences — if I play here, what will the opponent play, then what will I play, and so on for the next five to ten moves. Strong Go play requires reading complex tactical sequences, particularly in:

• Cutting and connecting: Can I cut the opponent’s group into two pieces? Can I connect my two groups?
• Capturing races (semeai): Two groups attacking each other; whichever runs out of liberties first dies.
• Ko fights: The ko rule prevents immediate recapture; ko fights are positional battles with elaborate threat sequences.
• Endgame sente (yose): Endgame moves that force a response from the opponent (sente) vs moves that don’t (gote). Sente moves are worth their face value plus future opportunities.

Reading skill develops with practice. The standard exercise is to solve tsumego mentally without placing stones, calculating the full sequence in your head before checking.

The endgame (yose)

Go endgames are often technical and matter more than beginners realize. Two players can settle most of the middlegame and still see a five- or ten-point swing in the endgame depending on technique.

The two key endgame ideas:

• Sente vs gote: Take sente (move-forcing) endgames before gote (non-forcing) ones. Sente moves preserve your move advantage; gote moves give it to the opponent.
• Counting: Each endgame move has a point value (its “miai value”). Play the largest endgame first.

Counting and sente analysis is what separates 5-kyu from dan-level play in many endgame positions.

Komi and the structure of the game

Komi is the score adjustment given to White to compensate for Black’s first-move advantage. Standard komi varies (6.5 in most modern professional play, 7.5 in some Japanese tournaments). The .5 prevents tied games.

Without komi, Black would have a slight but significant advantage. With proper komi, the game is approximately balanced — and the exact value of komi has been debated and adjusted over decades as theory has advanced.

Improvement path

The standard recommendation for Go improvement:

1. Learn the rules and play 50 games on a 9×9 or 13×13 board.
2. Start doing tsumego (life-and-death problems) daily — start with easy ones, advance gradually.
3. Study shape — the named good and bad shapes give you a vocabulary for evaluating your own moves.
4. Learn a few core joseki for the corners — not to memorize, but to understand what reasonable play looks like.
5. Play 19×19 games against opponents slightly stronger than you.
6. Review your losses, particularly the moments where you lost a group or got cut.
7. Repeat indefinitely. Go skill develops over years.

For a beginner-friendly introduction to the rules and first principles, see our how to play Go guide. For the citation-supported history and theory, the Wikipedia article on Go is a strong starting point.

Frequently asked questions

What is the most important opening principle in Go?

Corners first, then sides, then center. The geometry of the board makes corner territory dramatically more efficient than center territory.

What does “two eyes” mean in Go?

A group needs two separate enclosed empty spaces to be permanently alive. With one eye, the opponent can fill the rest of the group’s liberties and capture it. With two eyes, that’s impossible.

How do I get better at Go?

Solve life-and-death problems (tsumego) daily. Beyond the rules and a few opening principles, tsumego is the highest-return improvement activity. A few hundred problems at your level, drilled until automatic, raises your strength more than opening study.

What’s the difference between joseki and fuseki?

Joseki refers to standard corner sequences — two-player exchanges in a corner that yield a roughly balanced outcome. Fuseki refers to whole-board opening strategy — how you place your stones across all four corners and the early sides.

Did AlphaGo really master Go?

Yes. DeepMind’s AlphaGo defeated top human professional Lee Sedol in 2016. Successor systems (AlphaGo Zero, KataGo) have continued to advance and now outplay all human players definitively.

The bottom line

Mastering Go is a multi-year project for any serious player, but the foundations are surprisingly small. Corners first, then sides, then center. Good shape, not bad shape. Two eyes mean life. Tsumego daily. The game’s depth is famous, but the principles that take you most of the way through amateur ranks are concrete and learnable. For a lighter browser break that doesn’t require any board, the Chrome Dino game is one-button arcade relief.