Dots and Boxes Variants: From Square Lattices to Triangular and Hex Grids

Most people only know classic dots and boxes — played on a square grid, with 1×1 squares as the capturable cells. But the game has a family of variants played on other lattices: triangular grids, hexagonal grids, and rectangular grids of different aspect ratios. Each variant changes the strategic character of the game in interesting ways.

This post is a tour through the variants. What changes, what stays the same, and what playing them teaches you about the core game.

The standard version: square grid

In standard dots and boxes, dots are arranged in a square lattice — a grid of evenly spaced points. The lines connect adjacent dots horizontally or vertically. The cells between lines are 1×1 squares.

Every box has 4 sides. Chains form along rows and columns of boxes. Loops form around single "holes" in the structure.

This is the version everyone knows, and the version all the strategy theory has been worked out for.

Triangular dots and boxes

The triangular variant is played on a triangular lattice — dots arranged so that each interior dot has 6 neighbors rather than 4. The cells formed by adjacent dots are triangles, not squares.

What changes:

• Each triangle has 3 sides rather than 4. This makes each cell easier to complete (one-third fewer moves per cell) and radically shortens the game.
• Chains are shorter on average. A chain of triangles terminates faster because each triangle closes with fewer lines.
• The double-cross works differently. With 3 sides, the math of sacrificing "all but two" does not cleanly apply. Variants of the double-cross exist but are specific to triangular geometry.
• Loops are more common. The hexagonal structure of the lattice creates natural ring shapes.

Triangular dots and boxes feels faster and less deep than the square version. It is fun but has less room for complex strategy.

Hex dots and boxes

Less common, but interesting. Played on a hexagonal lattice, where dots are arranged at the centers of a honeycomb pattern. Lines connect adjacent hex centers, and the cells between lines are... well, it depends on how you define the rules.

The most common hex variant turns the game inside out — lines are drawn between adjacent hexagons, and hexagonal cells are claimed when all 6 sides are drawn. This creates very long chains (because each hexagon has 6 sides) and a slow-building game.

Hex dots and boxes is rare enough that there is almost no strategic theory for it. If you are curious, it makes a good occasional variant for experimentation.

Rectangular grids of different aspect ratios

Not strictly a new lattice, but a variant worth mentioning. A standard 5×6 box grid is different from a 5×5 or 4×7, even though they use the same square lattice.

What changes with aspect ratio:

• Asymmetric boards favor different players. The chain rule's parity target depends on board dimensions, so different rectangular dimensions change which parity is winning.
• Long, thin boards encourage single long chains. A 3×10 board tends to produce one or two very long chains spanning the length of the board.
• Square boards produce balanced chain structure. A 5×5 board tends to have more varied regions.

Playing the same "dots and boxes" on different rectangular aspect ratios is a mild form of variant. The strategic ideas transfer, but the specific parity numbers change.

Other variants worth knowing

Beyond lattice changes, several rule-level variants exist:

Scoring variants

• Bonus captures — each box is worth points, but specific boxes are marked double-value. Adds tactical complexity.
• Race to a target — first to claim N boxes wins, ending the game early. Makes endgame structure matter less.
• Penalty for completing a box — deliberately unusual rule set where completing boxes is neutral or penalized, forcing players to avoid captures.

Dot Clash uses the "race to target" variant by default — first to a score target wins, rather than playing until the board fills.

Three-player variants

Standard dots and boxes is two-player. Three-player variants exist where three players take turns on a single board. The strategic dynamics change dramatically because two players can implicitly cooperate against a third.

Three-player dots and boxes is a neat exercise but rarely played seriously.

Misère variants

Normal dots and boxes: most boxes wins. Misère: fewest boxes wins. Same rules, opposite goal.

Misère dots and boxes is strategically fascinating because the techniques flip. The double-cross is still a tempo-buying move, but now you want to give the opponent chains rather than receive them. The chain rule parity flips.

Misère variants are good puzzles because they force you to re-examine every strategic concept from the opposite direction.

Partisan variants

Standard dots and boxes is symmetric — both players have the same available moves. Partisan variants restrict some moves to specific players (e.g., "player A can only draw horizontal lines, player B can only draw vertical lines"). The resulting game has different strategic character.

These are mostly theoretical curiosities, but mathematicians have studied them as extensions of combinatorial game theory.

What variants teach you about the core game

Playing variants is not just for novelty. It teaches you things about standard dots and boxes that pure standard play does not:

Chain length matters more on large cells. Playing on hex grids (6-sided cells) makes chains slower to form and emphasizes how chain length interacts with strategy. You come back to the 4-sided standard game with a sharper sense of how much chain length drives outcomes.

Parity is fundamental. Trying variants with different board dimensions makes you acutely aware that parity is always an issue, just with different specific values. You stop relying on memorized parity numbers and start understanding parity as a concept.

Double-cross generalizes. Playing the triangular variant forces you to reinvent sacrifice techniques, which illuminates what the square-grid double-cross is actually doing. The pattern is "sacrifice material for tempo," and the specific implementation differs by geometry.

Rules are arbitrary, structure is not. Variants emphasize that rules determine a game's surface but do not determine its strategic structure. Two games with similar structural properties (chain formation, parity, forced captures) play similarly even if the specific rules differ.

Dot Clash as a variant

Dot Clash can be seen as a variant of dots and boxes, though a dramatic one. The lattice is still square, but the capture mechanic is totally different — you capture by enclosing opponent dots rather than by drawing the fourth side of a box.

Despite the mechanical difference, many strategic ideas transfer. Chain-like structures form, parity-like dynamics apply, and a sacrifice-tempo trade analogous to the double-cross exists. Playing Dot Clash alongside classic dots and boxes illustrates how the surface mechanics can change while the deep strategic ideas persist.

Try them for yourself

If you want to try variants:

• Triangular dots and boxes: draw a triangular lattice on paper. 5 or 6 rows. Play a normal game with 3-sided cells.
• Aspect ratio variants: play 4×6, 3×7, 5×5, 6×6. Notice how different they feel.
• Misère: same rules, try to get the fewest boxes. Play one game of it and see what happens.
• Race to target: pick a score (say, 6 boxes) and play first-to-target. Notice how the pace changes.
• Dot Clash: try the digital grid-capture game with its different capture mechanic.

Each variant reveals something different about the core game. Half an hour of variant play teaches you something that a week of standard play might not.

The mathematical side of variants

For the mathematically inclined, variants are not just fun — they are test cases for combinatorial game theory. By solving specific variants (small triangular boards, small misère boards, etc.), mathematicians learn which properties of the standard game are general and which are specific to square lattice geometry.

Some results are surprising. The misère variant is harder to solve than the normal variant, even though the rules are the same — "hardest" in the sense that the known algorithms for solving normal play do not easily adapt to misère.

Triangular and hex variants have been less studied but are promising directions for future research.

A compact taxonomy of variants

To summarize:

• Standard — square lattice, 4-sided cells, normal scoring.
• Triangular — triangular lattice, 3-sided cells, faster games.
• Hexagonal — hex lattice, 6-sided cells, slow games.
• Rectangular aspect ratios — same square lattice, different dimensions, different parity.
• Misère — fewest boxes wins.
• Race to target — first to N boxes wins.
• Three-player — adds cooperative dynamics.
• Partisan — asymmetric move availability.
• Dot Clash — different capture mechanic but shares strategic structure.

Each teaches something different. Each is worth trying at least once.

The takeaway

Variants are the laboratory of game design. Playing them expands your understanding of what "strategy" means across a family of related games, rather than just getting deep in one specific version. A player who has only played standard 5×5 dots and boxes is narrower than a player who has tried triangular, misère, and rectangular variants — even if the first player has logged more total games.

Variety in play is variety in thinking. Try one variant this week.