Gomoku Tactics: Open Threes, Double Threats, Forced Wins
Gomoku has a vocabulary, and it's tiny. Learn five shapes and you can read any position on the board; skip them and you'll lose, repeatedly, to anyone who didn't, without ever quite seeing how. The rules take ten seconds — on Inkplay's gomoku you place white stones, the computer answers in black, and the first unbroken row of five in any direction wins. Everything past that sentence is grammar.
The alphabet: threes and fours
An open three is three stones in a row with both ends empty. It isn't a win; it's a promise. Left alone it becomes an open four next turn, so it demands a reply. A simple four — four in a row with one open end — is stronger: the defender must plug the single completion point immediately or lose. Once plugged, though, those are spent stones. An open four — four with both ends empty — ends the argument. Two completion points, one block. Unless the defender can win outright on their very next move, an open four wins by force.
That hierarchy gives the game a currency of urgency: an open four beats everything, a four demands one specific answer, an open three demands some answer, and everything else is quiet. When threats coexist, the faster one rules — fours outrun threes. That one sentence settles more gomoku games than any other.
Two footnotes to the alphabet. A blocked three is nearly worthless, and beginners build them constantly — three stones butting into an enemy stone, threatening nothing. And the broken shapes count too: a split three (two stones, a gap, then a third on the same line) matures into a four just as fast as the solid version and gets overlooked far more.
Building a double threat
Against anyone paying attention, one line never wins. Every three gets blocked, every four gets plugged, and a single-minded attack just donates stones to the defence. Wins are manufactured, and the unit of manufacture is the move that creates two threats at once. The classic is the 4-3: a single stone that completes a simple four and an open three simultaneously. The four demands an instant block; while the defender plugs it, the three grows into an open four, and no move covers both. The 3-3 (two open threes from one stone) and the 4-4 (two fours) do the same job.
You don't find double threats by staring harder at move twenty. You build them from move five. They live where lines cross, so place early stones at intersections that serve several directions at once — every stone sits on four lines, and a stone quietly participating in two developing shapes is worth two stones that aren't. Loose, connected clusters near the centre beat long straight pushes, both because the centre fits the most lines of five and because clusters breed crossings. Keep your stones within touching distance of each other and the double threats start appearing on their own.
1993: someone finished the game
In 1993 the Dutch researcher Victor Allis proved that free-style gomoku on a 15×15 board is a first-player win — with perfect play, the second player is simply lost. (He'd already done the same to Connect Four in 1988.) The method matters more than the trophy. His program didn't grind through every position; it used threat-space search, following almost nothing but chains of forcing moves — fours and threes stitched together, each compelling its reply, ending in an unavoidable double threat. The proof of the entire game is the tactic from the previous section, industrialised.
The gomoku world took the result seriously. Renju, the tournament version, handicaps the first player outright — no double-threes, no double-fours, no overlines. Modern tournament gomoku instead uses swap openings, where one player builds the opening and the other chooses which side to take, which turns a too-strong opening into a liability. When a game needs rule surgery to stay fair, that tells you exactly how heavy the first move is.
On Inkplay you play free-style, and you move first. Theory says your seat wins. Theory also assumes perfect play, which nobody has — so the proof's practical residue is a directive about initiative. The first player's edge is tempo, and tempo compounds: every forcing shape you make is a move the computer spends defending instead of building. Attack with shapes that demand answers. As the first player, passivity is the only real way to lose your inheritance.
Defence is counting
Defence in gomoku is bookkeeping, done in strict order:
- Can you win this move? Then nothing else on the list matters.
- Do they have a four? There's one legal answer. Play it — unless you can finish a faster win of your own.
- Do they have an open three? Block it now, before it grows up — and since either end usually works, block on the end that also feeds your own shapes.
- None of the above? Only now are you free to build.
The counting extends across whole sequences. A chain of your own fours suspends every threat the opponent owns — they spend each turn plugging — but make sure the chain ends in a double threat and not in an apology, because when the fours run out, their position is still sitting there.
The loss that stings most in casual play isn't missing a threat; it's counting one line at a time. The stone that made the three you're busy blocking may have quietly finished a second shape you never read, so when any stone lands, check all four lines through it before replying. Chess forgives — drop a pawn and the game goes on. Gomoku holds no material to hide behind, only tempo, which is why five minutes of it feels like a knife fight. It's the same muscle that decides checkers endgames: whoever counts one move deeper takes the point. Gomoku is just tactics with everything else stripped away. Learn the alphabet and the board starts talking.