How to Tell If a Die Is Fair

Why Die Fairness Matters

Most people never question whether their dice are fair — and for casual play, a slightly biased die probably doesn't matter. But for competitive board games, tabletop RPG sessions where dice rolls have real stakes, or anyone who wants to understand their equipment better, fairness is worth investigating. A die doesn't need to be deliberately loaded to be biased; manufacturing defects, material impurities, and even normal wear can create measurable imbalances over time.

Visual Inspection

Start with your eyes. Place the die on a flat surface and examine each face systematically. What to look for:

• Rounded vs. sharp corners: Mass-market plastic dice are injection-moulded, and rounded or asymmetrically worn corners can bias which face lands down. Precision casino dice have perfectly sharp edges and corners, which is part of what makes them fair.
• Pip consistency: On a standard D6, the pips (dots) on each face are drilled or moulded cavities. These cavities are often not filled with additional material, meaning the face with the most pips (6) has the most material removed and is technically the lightest face. This causes a very slight natural bias toward 6 landing up (since the heavier face 1 sinks down) on some cheap dice — though the effect is usually negligible.
• Flashing: "Flashing" refers to thin plastic protrusions along the mould seam. Visible flashing along an edge means the die has a manufacturing defect that could affect rolling behavior.
• Surface uniformity: Any visible bubbles, deformations, or inconsistencies in the plastic suggest quality control issues that may create bias.

Visual inspection alone cannot definitively prove a die is fair or biased, but it can immediately identify obviously flawed dice worth setting aside.

The Saltwater Float Test

The saltwater float test is a popular method for detecting heavily weighted dice. The idea: dissolve a large amount of table salt into warm water until the die floats (you need a saturated or near-saturated solution — roughly 8–9 tablespoons of salt per cup of warm water). Drop the die in.

A fair die, floating freely, should spin to random orientations when you give it a gentle flick. A biased (loaded) die will consistently orient with the heavy face down and a light face up — meaning the number opposite the heavy face will repeatedly face upward after each spin.

Perform ten to twenty flicks, recording the top face each time. If one number appears as the top face significantly more than 1-in-6 times across many trials, the die may be weighted.

Important caveats: The float test is most useful for detecting heavy loading — cavity loading or inserted weights. It is much less sensitive to subtle biases caused by pip drilling or rounded corners. A die can pass the float test and still be slightly unfair. Conversely, some dice float inconsistently due to air bubbles trapped during manufacturing, producing misleading results. The float test is a good first screen, not a definitive proof.

The Roll Distribution Test

The most statistically rigorous method is simply rolling the die many times and recording the results. For a D6, roll it at least 100 times and tally how many times each face appears. A fair die will produce approximately 16-17 of each face per 100 rolls — but natural variation means you shouldn't expect exactly 16.7 of each.

To interpret your results more formally, you can apply a chi-square goodness-of-fit test. Calculate the expected frequency (100 ÷ 6 ≈ 16.67 per face), then compute chi-square: sum of (observed - expected)² / expected across all faces. A result below 11.07 suggests no statistically significant bias at the 95% confidence level for a D6. A result above that threshold warrants suspicion — though 300+ rolls give a far more reliable test than 100.

For casual players, just eyeballing the tally is often enough. If one face appears 25 times while another appears 8 times across 100 rolls, you have a suspicious die worth replacing.

Precision Dice vs. Consumer Dice

Casino-grade precision dice are machined to tolerances of 1/10,000 of an inch. They're made from transparent cellulose acetate (which prevents cavity loading) and have sharp, machined edges. These dice are objectively fairer than mass-market game dice, which is why casinos use them and why competitive dice games sometimes require casino-grade equipment.

Standard board game and RPG dice are manufactured to much looser tolerances. Most are acceptably fair for recreational play, but none should be considered precision instruments. High-end RPG dice manufacturers do produce tighter-tolerance dice, but even these fall short of casino-grade precision.

What "Fair" Really Means Statistically

No physical die is perfectly fair — the ideal of exactly 1-in-6 per face exists only in theory. What matters practically is whether any bias is large enough to affect gameplay meaningfully. A die biased 1% toward one face is undetectable in normal play. A die biased 15% toward one face would noticeably skew a game over many sessions. Understanding this spectrum helps calibrate how much testing is worth doing for your level of play.