Three-player chess, done right

3Chess is a three-player chess variant. Three players, one symmetric board, every move a dilemma between two opponents. Play it free in your browser or buy the full game for a one-time $9.99.

The kingmaker problem — and how 3Chess fixes it

Every classic attempt at three-player chess hits the same wall. When one player falls behind or gets eliminated, the winner is often decided not by skill. It's decided by which of the remaining two that player chooses to attack. They become a "kingmaker". The endgame feels arbitrary and the whole game feels unserious.

3Chess attacks this from three directions at once:

• Cascading checkmate. A single move can checkmate two opponents simultaneously. A sharp player can simply win rather than being forced to pick a beneficiary.
• Territory scoring. If the game doesn't end in a knockout, it's decided by how much of the board your surviving pieces control. Falling behind still gives you something concrete to fight for, instead of just a spoiler vote.
• The 20-move rule. If 20 moves pass for each player without progress, the game ends and territory decides it. No endless dead positions, no stalling.

Together these mean the strongest player tends to win. Second place is contested on its own merits. There's never a stretch where your only move is to hand someone else the game. There's a longer write-up in The kingmaker problem.

Why a hexagon, not a square?

The hard part of three-player chess isn't picking who goes first. It's picking what shape the board is. A standard 8×8 square has two-fold symmetry. Every line, rank, and file points from one side to its mirror image. There's no third "side" that gets the same treatment. Force three armies onto it and at least one of them ends up cramped, off-axis, or borrowing the same starting cells as another. Past three-player attempts on square boards always feel grafted-on for this reason: one player faces a fair fight, the other two play awkward angles.

A hexagon doesn't have this problem. Six-fold rotational symmetry permits three identical territories arranged at 120° to one another. Every player gets the same edge, the same back-rank shape, the same march toward the opposite face. Take any of the three armies and rotate the board: what you see is structurally identical. That property is the prerequisite to caring about who plays best. No player is at a built-in disadvantage no matter where they sit. Without it, the result is always tainted by where the seat happened to be.

There's a side effect. Every hex cell has six orthogonal neighbours instead of four, plus six "diagonal" neighbours one ring out. Twelve directions in all. That changes how every piece moves. It's the price of three-way fairness. It's where 3Chess stops being "regular chess with one more seat".

Anatomy of a typical game

In two-player chess the phases are familiar: opening, middlegame, endgame. They map roughly to time on the clock. In three-player chess the phases describe something different: the number of opponents you have to think about.

Opening (~moves 1–10). All three armies develop in parallel. You can't be aggressive against one opponent without giving the other one a free initiative. Everyone tends to develop centrally and toward the middle of the board. The classic two-player opening principles still apply: develop your minor pieces, contest the centre, get your king safe. But the centre is now shared three ways.

Middlegame (~moves 10–30). Tension builds along three pairwise fronts. Most moves you make either pressure one opponent or shore up against one of them. Very few moves affect both opponents equally. The strategic question becomes which front to attack and which to defend. The answer almost always involves the other player's incentives. Attack the leader and the third player benefits. Attack the trailer and you waste tempo. The strongest players read the whole triangle.

Endgame. One of three things happens. (1) You deliver a cascading checkmate to both opponents in a single move: outright victory. (2) One opponent is eliminated and the surviving two settle it on territory or by mate. (3) The position reaches one of the four draw triggers (threefold repetition, 20-move rule, insufficient material, two-player stalemate) and territory decides it. Every game has a clear result. There are no true draws.

Three colour classes, not two

A square board has two colour classes: white squares and black squares. A chess bishop spends its life locked to one of them. A hex board has three colour classes, defined by (q − r) mod 3 on axial coordinates. Each bishop is still locked to one class for the whole game. But now that class is one third of the board, not half. Two bishops on opposite-coloured cells can never meet. Three bishops on three different classes never attack the same cell.

The knight goes the other way. On a square board a knight always alternates colours. On a hex board the knight is at distance 3 from its starting cell, which means it always changes colour class. Mating with two knights is harder. Trapping a king with a single knight is easier than you'd expect.

The pawn is the piece most reshaped. Instead of one forward direction and two diagonals, a hex pawn pushes into a pair of forward cells and captures on the two outer-flanking cells. Each army's pawns face a different edge. Gold pushes south, Azure east, Crimson west. All three promote on the opposite player's back rank. The full piece-by-piece geometry is on the rules page.

A short history of three-player chess

People have been trying three-player chess for at least four hundred years. The earliest widely documented attempt is George Hope Verney's 1881 design on a Y-shaped board. Three rectangular wings met in a centre, with players seated at each tip. It worked in a parlour-game sense, but the geometry was visibly unfair. Pieces near the centre had access to three opposing armies. Pieces on the outside had access to one. The result depended heavily on seat.

Twentieth-century attempts trended toward hexagonal boards. Włodzimierz Gliński's Hexagonal Chess (1936) was a two-player design that influenced everything after. Its 91-cell hex board and re-expressed piece movements showed that chess on hexes wasn't a gimmick. Several three-player adaptations followed. None solved what would later get called the kingmaker problem: a losing player gets to pick the winner.

The 1990s and 2000s saw computer-friendly variants: sectorial boards, drop variants, scoring layers. Most leaned on house rules to "discourage" kingmaker plays rather than designing them out. None gained competitive traction. Three-player chess remained a curiosity rather than a game with serious players.

3Chess takes a different approach. Instead of asking players to be polite, the rules of the game itself make the kingmaker option unproductive. Cascading checkmate lets the strongest player simply win. Territory scoring keeps a trailing player competitive instead of reducing them to a spoiler vote. The 20-move rule scores dead positions instead of letting them stall. Together they remove the configurations where one player's irrational move decides another player's first place. The history of three-player chess is mostly a list of attempts that didn't engineer this away. 3Chess does.

Ways to play

• Vs AI. Three difficulty/character bots. The fastest way to learn three-player tactics.
• Pass-and-play. Three people, one device, around a table.
• Online. Timed three-player games against other people. Results feed a rating ladder.

Where to play three-player chess

Right here. The free browser demo runs in any modern browser with no install. It's the real engine, real bots, real rules. The full game is a one-time $9.99 purchase (online play, all modes). A DRM-free desktop build is on itch.io, with Steam and mobile coming.

Frequently asked questions

↑ Back to Learn