In plain English. Every die is fair. We rolled each one 100,000 times and counted every face. No die ever comes up exactly equally — luck always wobbles a little — so the fair question is whether a die wobbles more than a perfect die would. For example, a perfect d20 rolled 100,000 times is usually about 2.9% off on its most-off face, and up to 3.9% off one time in ten; ours was 3.4% off. Every die passes the fairness test below.
Fairness is the first thing a tabletop player should ask of any dice roller, so here are the numbers. We threw every die in Dice Drop 100,000 times in its 3D physics simulation and counted how often each face came up: 1,600,000 rolls in all, last run on 7 October 2026.
How the test works
- The same dice and tray you roll in the app. Dice Drop’s dice are a physics simulation: each die is a 3D shape with weight that is thrown, bounces and comes to rest, and the result is read from the face that ends up on top. The test runs the app’s own code for all of it: the same throw, the same physics, the same rules for deciding when a die has stopped and which face is up. The only difference is that nothing is drawn on screen, so it runs much faster than real time.
- One die per throw, so every result is independent of the others, which is what the statistics assume. (Dice thrown together knock into each other, so their results are not quite independent; see below.)
- Every throw counts. Each one produced a result; none were left out or re-thrown.
- The d100 and d120 are different. They are so close to spheres that they roll as a ball, and their number is a fair random draw that the ball is turned to show. So for those two we tested the draw itself, 1,000,000 times each.
How to read the results
A fair die doesn’t give exactly equal counts; genuine randomness always wobbles a little. The question is whether the wobble is bigger than chance would produce. The standard test for that is chi-square, which gives a p-value:
- p above 0.05: the counts are consistent with a fair die.
- p between 0.0028 and 0.05: a little uneven, but well within chance. Even a perfect die lands here one time in twenty, and with 18 dice tested, often one of them will.
- p below 0.0028: unlikely to be chance. That would be a die to look into.
The 0.05 is a probability, not a percentage a face is allowed to be off. How far a face can be from “exactly fair” depends on how many times the die was rolled, which is why every die below is shown beside what a perfect die rolled the same number of times would do.
That last threshold is 0.05 divided by the 18 dice: the standard way to judge many tests at once, so that ordinary chance across 18 dice isn’t mistaken for a fault.
The results
| Die | Rolls | Most-off face | A perfect die: usually | p | Verdict |
| d2 (coin) | 100,000 | 0.16% | 0.33% (up to 0.62%) | 0.608 | Fair |
| d3 | 100,000 | 0.38% | 0.56% (up to 0.94%) | 0.695 | Fair |
| d4 | 100,000 | 0.50% | 0.76% (up to 1.2%) | 0.718 | Fair |
| d5 | 100,000 | 0.77% | 0.95% (up to 1.4%) | 0.472 | Fair |
| d6 | 100,000 | 0.52% | 1.1% (up to 1.7%) | 0.984 | Fair |
| d7 | 100,000 | 1.7% | 1.3% (up to 1.9%) | 0.198 | Fair |
| d8 | 100,000 | 1.0% | 1.5% (up to 2.1%) | 0.851 | Fair |
| d10 | 100,000 | 1.3% | 1.7% (up to 2.4%) | 0.738 | Fair |
| d12 | 100,000 | 1.1% | 2.0% (up to 2.8%) | 0.971 | Fair |
| d14 | 100,000 | 3.1% | 2.2% (up to 3.1%) | 0.037 | Fair |
| d16 | 100,000 | 2.0% | 2.5% (up to 3.3%) | 0.749 | Fair |
| d20 | 100,000 | 3.4% | 2.9% (up to 3.9%) | 0.266 | Fair |
| d24 | 100,000 | 3.7% | 3.3% (up to 4.3%) | 0.693 | Fair |
| d30 | 100,000 | 4.1% | 3.9% (up to 5.0%) | 0.297 | Fair |
| d48 | 100,000 | 6.5% | 5.4% (up to 6.7%) | 0.179 | Fair |
| d60 | 100,000 | 7.0% | 6.1% (up to 7.6%) | 0.011 | Fair |
| d100 | 1,000,000 | 2.5% | 2.7% (up to 3.3%) | 0.987 | Fair |
| d120 | 1,000,000 | 3.7% | 3.0% (up to 3.6%) | 0.199 | Fair |
“Most-off face” is how far the die’s least typical face was from an exactly equal share. “A perfect die: usually” is the same figure for a flawless die rolled the same number of times, typically and at most nine times in ten.
Every die passes. 2 of 18 came in below 0.05, about what chance alone produces across this many dice; each one has a note below its chart explaining why it is still fair.
The prisms: do the ends come up fairly?
A d5 or d7 is a prism: its sides are identical, but how often it stands on an end depends entirely on how long the prism is. The length is set so that a fair share of throws land on an end, and this is the check that it is right:
- d5: landed on an end 40.16% of the time. A perfect d5 thrown 100,000 times lands on an end between 39.69% and 40.31% (95 times in 100), so ours is right inside that range.
- d7: landed on an end 28.46% of the time. A perfect d7 thrown 100,000 times lands on an end between 28.28% and 28.86% (95 times in 100), so ours is right inside that range.
Leaning dice
A die that comes to rest cocked against the tray wall or another die isn’t re-rolled or straightened. It stays where it stopped and is read by the face closest to pointing up, the same way a player would read a die that came to rest at an angle. Those rests are included in every count above. The one exception is a coin: one left standing on its rim isn’t a result, so it is laid flat.
What this test has already fixed
Publishing these numbers is only worth anything if we act on them, so here is everything the test has turned up so far, and what we did about it.
- Every die: an uneven starting spin. Each throw began from a random orientation made from three random angles. That is a well-known trap: those orientations are not evenly spread, and a short tumble can carry a little of that through to the result (it showed as a slightly uneven d24). Every throw now starts from an evenly random orientation.
- Most dice: how they spin. The physics engine worked out how every die spins as if it were a box. That is exactly right for the d4, d6, d12 and d20, but slightly wrong for the other shapes, and it made the d5 tumble a little unevenly, so one of its sides came up a little less often than the rest. Every die now spins the way its true shape would. You won’t see a difference in how they roll; the counts above show the difference it makes.
- The coin: one face slightly favoured. Tossed on its own, the coin came down one way 54% of the time. We traced it to the physics engine, which treats the two flat sides of its cylinder shape a little differently. Each coin’s collision shape is now built either way up at random, so that difference favours heads half the time and tails the other half, and cancels out.
- The d5 and d7: tuned for a single throw. How often a prism stands on an end depends on its length. Their lengths had been set by throwing ten at once; thrown singly, the d5 stood on an end a little too often and the d7 a little too rarely. Both have been re-tuned.
- How we test: one die at a time. Our early runs threw several dice at once, and the d20 and d60 came in slightly uneven, but with different faces each run, which is the mark of chance rather than a loaded die. Dice thrown together collide, so their results aren’t fully independent, and the statistics read too much into ordinary wobble. Every result above comes from one die per throw.
All of these fixes are live on the website, and in the iPhone and Android apps from version 1.0.4.
Every die, face by face
Each bar is one face. The line across is what a perfectly fair die would give, and the shaded band is where a perfect die’s faces land 95% of the time when rolled the same number of times. Hover over or tap a bar for its exact count.
d2 (coin)100,000 rolls · most-off face 0.16% (a perfect die: usually 0.33%) · p = 0.608 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 49,919 | -0.2% |
| 2 | 50,081 | 0.2% |
For comparison, a perfect d2 (coin) rolled 100,000 times: 95% of its faces land within ±0.63% of fair, and its most-off face is usually about 0.33% away.
d3100,000 rolls · most-off face 0.38% (a perfect die: usually 0.56%) · p = 0.695 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 33,459 | 0.4% |
| 2 | 33,253 | -0.2% |
| 3 | 33,288 | -0.1% |
For comparison, a perfect d3 rolled 100,000 times: 95% of its faces land within ±0.89% of fair, and its most-off face is usually about 0.56% away.
d4100,000 rolls · most-off face 0.50% (a perfect die: usually 0.76%) · p = 0.718 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 25,124 | 0.5% |
| 2 | 25,052 | 0.2% |
| 3 | 24,906 | -0.4% |
| 4 | 24,918 | -0.3% |
For comparison, a perfect d4 rolled 100,000 times: 95% of its faces land within ±1.1% of fair, and its most-off face is usually about 0.76% away.
d5100,000 rolls · most-off face 0.77% (a perfect die: usually 0.95%) · p = 0.472 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 20,104 | 0.5% |
| 2 | 19,854 | -0.7% |
| 3 | 19,878 | -0.6% |
| 4 | 20,154 | 0.8% |
| 5 | 20,010 | 0.1% |
For comparison, a perfect d5 rolled 100,000 times: 95% of its faces land within ±1.3% of fair, and its most-off face is usually about 0.95% away.
d6100,000 rolls · most-off face 0.52% (a perfect die: usually 1.1%) · p = 0.984 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 16,674 | 0.0% |
| 2 | 16,753 | 0.5% |
| 3 | 16,673 | 0.0% |
| 4 | 16,623 | -0.3% |
| 5 | 16,629 | -0.2% |
| 6 | 16,648 | -0.1% |
For comparison, a perfect d6 rolled 100,000 times: 95% of its faces land within ±1.4% of fair, and its most-off face is usually about 1.1% away.
d7100,000 rolls · most-off face 1.7% (a perfect die: usually 1.3%) · p = 0.198 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 14,375 | 0.6% |
| 2 | 14,241 | -0.3% |
| 3 | 14,463 | 1.2% |
| 4 | 14,241 | -0.3% |
| 5 | 14,217 | -0.5% |
| 6 | 14,417 | 0.9% |
| 7 | 14,046 | -1.7% |
For comparison, a perfect d7 rolled 100,000 times: 95% of its faces land within ±1.5% of fair, and its most-off face is usually about 1.3% away.
d8100,000 rolls · most-off face 1.0% (a perfect die: usually 1.5%) · p = 0.851 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 12,375 | -1.0% |
| 2 | 12,590 | 0.7% |
| 3 | 12,456 | -0.4% |
| 4 | 12,532 | 0.3% |
| 5 | 12,430 | -0.6% |
| 6 | 12,557 | 0.5% |
| 7 | 12,479 | -0.2% |
| 8 | 12,581 | 0.6% |
For comparison, a perfect d8 rolled 100,000 times: 95% of its faces land within ±1.7% of fair, and its most-off face is usually about 1.5% away.
d10100,000 rolls · most-off face 1.3% (a perfect die: usually 1.7%) · p = 0.738 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 9,866 | -1.3% |
| 2 | 9,976 | -0.2% |
| 3 | 10,071 | 0.7% |
| 4 | 10,011 | 0.1% |
| 5 | 9,986 | -0.1% |
| 6 | 9,893 | -1.1% |
| 7 | 9,968 | -0.3% |
| 8 | 10,015 | 0.1% |
| 9 | 10,088 | 0.9% |
| 10 | 10,126 | 1.3% |
For comparison, a perfect d10 rolled 100,000 times: 95% of its faces land within ±1.9% of fair, and its most-off face is usually about 1.7% away.
d12100,000 rolls · most-off face 1.1% (a perfect die: usually 2.0%) · p = 0.971 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 8,311 | -0.3% |
| 2 | 8,399 | 0.8% |
| 3 | 8,328 | -0.1% |
| 4 | 8,338 | 0.1% |
| 5 | 8,272 | -0.7% |
| 6 | 8,335 | 0.0% |
| 7 | 8,427 | 1.1% |
| 8 | 8,335 | 0.0% |
| 9 | 8,400 | 0.8% |
| 10 | 8,239 | -1.1% |
| 11 | 8,287 | -0.6% |
| 12 | 8,329 | -0.1% |
For comparison, a perfect d12 rolled 100,000 times: 95% of its faces land within ±2.1% of fair, and its most-off face is usually about 2.0% away.
d14100,000 rolls · most-off face 3.1% (a perfect die: usually 2.2%) · p = 0.037 · Fair
Why this is still fair. Its most-off face was 3.1% from an exactly equal share, a little more than a perfect d14 usually shows (2.2%). But a perfect d14 rolled 100,000 times is 3.1% off or more about one time in 11, simply through luck. Its p-value of 0.037 means a perfect d14 would give counts at least this uneven about one time in 27. With 18 dice tested, one or two results like that are expected in every run, and they move to different dice from one run to the next. The test that decides fairness takes all of that into account, and this die passes it.
See the numbers
| Face | Times | vs fair |
| 1 | 7,363 | 3.1% |
| 2 | 7,114 | -0.4% |
| 3 | 7,116 | -0.4% |
| 4 | 7,044 | -1.4% |
| 5 | 7,298 | 2.2% |
| 6 | 7,037 | -1.5% |
| 7 | 7,202 | 0.8% |
| 8 | 7,119 | -0.3% |
| 9 | 7,121 | -0.3% |
| 10 | 7,151 | 0.1% |
| 11 | 7,111 | -0.4% |
| 12 | 6,955 | -2.6% |
| 13 | 7,302 | 2.2% |
| 14 | 7,067 | -1.1% |
For comparison, a perfect d14 rolled 100,000 times: 95% of its faces land within ±2.3% of fair, and its most-off face is usually about 2.2% away.
d16100,000 rolls · most-off face 2.0% (a perfect die: usually 2.5%) · p = 0.749 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 6,340 | 1.4% |
| 2 | 6,375 | 2.0% |
| 3 | 6,302 | 0.8% |
| 4 | 6,196 | -0.9% |
| 5 | 6,244 | -0.1% |
| 6 | 6,198 | -0.8% |
| 7 | 6,217 | -0.5% |
| 8 | 6,275 | 0.4% |
| 9 | 6,153 | -1.6% |
| 10 | 6,324 | 1.2% |
| 11 | 6,282 | 0.5% |
| 12 | 6,126 | -2.0% |
| 13 | 6,259 | 0.1% |
| 14 | 6,259 | 0.1% |
| 15 | 6,261 | 0.2% |
| 16 | 6,189 | -1.0% |
For comparison, a perfect d16 rolled 100,000 times: 95% of its faces land within ±2.4% of fair, and its most-off face is usually about 2.5% away.
d20100,000 rolls · most-off face 3.4% (a perfect die: usually 2.9%) · p = 0.266 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 4,958 | -0.8% |
| 2 | 5,086 | 1.7% |
| 3 | 4,930 | -1.4% |
| 4 | 4,972 | -0.6% |
| 5 | 5,076 | 1.5% |
| 6 | 5,010 | 0.2% |
| 7 | 4,918 | -1.6% |
| 8 | 5,069 | 1.4% |
| 9 | 5,081 | 1.6% |
| 10 | 5,020 | 0.4% |
| 11 | 4,981 | -0.4% |
| 12 | 4,946 | -1.1% |
| 13 | 4,872 | -2.6% |
| 14 | 5,003 | 0.1% |
| 15 | 5,016 | 0.3% |
| 16 | 5,169 | 3.4% |
| 17 | 5,053 | 1.1% |
| 18 | 5,024 | 0.5% |
| 19 | 4,952 | -1.0% |
| 20 | 4,864 | -2.7% |
For comparison, a perfect d20 rolled 100,000 times: 95% of its faces land within ±2.8% of fair, and its most-off face is usually about 2.9% away.
d24100,000 rolls · most-off face 3.7% (a perfect die: usually 3.3%) · p = 0.693 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 4,130 | -0.9% |
| 2 | 4,185 | 0.4% |
| 3 | 4,124 | -1.0% |
| 4 | 4,198 | 0.8% |
| 5 | 4,180 | 0.3% |
| 6 | 4,168 | 0.0% |
| 7 | 4,130 | -0.9% |
| 8 | 4,114 | -1.3% |
| 9 | 4,237 | 1.7% |
| 10 | 4,179 | 0.3% |
| 11 | 4,208 | 1.0% |
| 12 | 4,057 | -2.6% |
| 13 | 4,113 | -1.3% |
| 14 | 4,322 | 3.7% |
| 15 | 4,173 | 0.2% |
| 16 | 4,147 | -0.5% |
| 17 | 4,169 | 0.1% |
| 18 | 4,238 | 1.7% |
| 19 | 4,243 | 1.8% |
| 20 | 4,065 | -2.4% |
| 21 | 4,175 | 0.2% |
| 22 | 4,185 | 0.4% |
| 23 | 4,143 | -0.6% |
| 24 | 4,117 | -1.2% |
For comparison, a perfect d24 rolled 100,000 times: 95% of its faces land within ±3.0% of fair, and its most-off face is usually about 3.3% away.
d30100,000 rolls · most-off face 4.1% (a perfect die: usually 3.9%) · p = 0.297 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 3,281 | -1.6% |
| 2 | 3,367 | 1.0% |
| 3 | 3,294 | -1.2% |
| 4 | 3,407 | 2.2% |
| 5 | 3,241 | -2.8% |
| 6 | 3,323 | -0.3% |
| 7 | 3,358 | 0.7% |
| 8 | 3,322 | -0.3% |
| 9 | 3,353 | 0.6% |
| 10 | 3,277 | -1.7% |
| 11 | 3,304 | -0.9% |
| 12 | 3,197 | -4.1% |
| 13 | 3,425 | 2.7% |
| 14 | 3,463 | 3.9% |
| 15 | 3,302 | -0.9% |
| 16 | 3,257 | -2.3% |
| 17 | 3,428 | 2.8% |
| 18 | 3,346 | 0.4% |
| 19 | 3,297 | -1.1% |
| 20 | 3,377 | 1.3% |
| 21 | 3,321 | -0.4% |
| 22 | 3,328 | -0.2% |
| 23 | 3,417 | 2.5% |
| 24 | 3,240 | -2.8% |
| 25 | 3,317 | -0.5% |
| 26 | 3,337 | 0.1% |
| 27 | 3,363 | 0.9% |
| 28 | 3,355 | 0.6% |
| 29 | 3,315 | -0.6% |
| 30 | 3,388 | 1.6% |
For comparison, a perfect d30 rolled 100,000 times: 95% of its faces land within ±3.4% of fair, and its most-off face is usually about 3.9% away.
d48100,000 rolls · most-off face 6.5% (a perfect die: usually 5.4%) · p = 0.179 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 2,041 | -2.0% |
| 2 | 2,076 | -0.4% |
| 3 | 2,109 | 1.2% |
| 4 | 2,162 | 3.8% |
| 5 | 1,947 | -6.5% |
| 6 | 2,026 | -2.8% |
| 7 | 2,087 | 0.2% |
| 8 | 2,099 | 0.8% |
| 9 | 2,092 | 0.4% |
| 10 | 2,068 | -0.7% |
| 11 | 2,085 | 0.1% |
| 12 | 2,093 | 0.5% |
| 13 | 2,059 | -1.2% |
| 14 | 2,068 | -0.7% |
| 15 | 2,044 | -1.9% |
| 16 | 2,149 | 3.2% |
| 17 | 2,105 | 1.0% |
| 18 | 2,063 | -1.0% |
| 19 | 2,089 | 0.3% |
| 20 | 2,106 | 1.1% |
| 21 | 2,107 | 1.1% |
| 22 | 2,055 | -1.4% |
| 23 | 2,014 | -3.3% |
| 24 | 2,198 | 5.5% |
| 25 | 2,058 | -1.2% |
| 26 | 2,020 | -3.0% |
| 27 | 2,055 | -1.4% |
| 28 | 2,106 | 1.1% |
| 29 | 2,080 | -0.2% |
| 30 | 2,042 | -2.0% |
| 31 | 2,042 | -2.0% |
| 32 | 2,160 | 3.7% |
| 33 | 2,077 | -0.3% |
| 34 | 2,038 | -2.2% |
| 35 | 2,039 | -2.1% |
| 36 | 2,184 | 4.8% |
| 37 | 2,045 | -1.8% |
| 38 | 2,119 | 1.7% |
| 39 | 2,168 | 4.1% |
| 40 | 2,151 | 3.2% |
| 41 | 2,137 | 2.6% |
| 42 | 2,035 | -2.3% |
| 43 | 2,111 | 1.3% |
| 44 | 2,090 | 0.3% |
| 45 | 2,087 | 0.2% |
| 46 | 2,066 | -0.8% |
| 47 | 2,133 | 2.4% |
| 48 | 2,015 | -3.3% |
For comparison, a perfect d48 rolled 100,000 times: 95% of its faces land within ±4.3% of fair, and its most-off face is usually about 5.4% away.
d60100,000 rolls · most-off face 7.0% (a perfect die: usually 6.1%) · p = 0.011 · Fair
Why this is still fair. Its p-value of 0.011 means a perfect d60 would give counts at least this uneven about one time in 94. With 18 dice tested, one or two results like that are expected in every run, and they move to different dice from one run to the next. The test that decides fairness takes all of that into account, and this die passes it.
See the numbers
| Face | Times | vs fair |
| 1 | 1,645 | -1.3% |
| 2 | 1,741 | 4.5% |
| 3 | 1,613 | -3.2% |
| 4 | 1,609 | -3.5% |
| 5 | 1,689 | 1.3% |
| 6 | 1,588 | -4.7% |
| 7 | 1,699 | 1.9% |
| 8 | 1,637 | -1.8% |
| 9 | 1,716 | 3.0% |
| 10 | 1,703 | 2.2% |
| 11 | 1,699 | 1.9% |
| 12 | 1,731 | 3.9% |
| 13 | 1,646 | -1.2% |
| 14 | 1,695 | 1.7% |
| 15 | 1,656 | -0.6% |
| 16 | 1,677 | 0.6% |
| 17 | 1,732 | 3.9% |
| 18 | 1,692 | 1.5% |
| 19 | 1,685 | 1.1% |
| 20 | 1,596 | -4.2% |
| 21 | 1,767 | 6.0% |
| 22 | 1,643 | -1.4% |
| 23 | 1,705 | 2.3% |
| 24 | 1,655 | -0.7% |
| 25 | 1,724 | 3.4% |
| 26 | 1,695 | 1.7% |
| 27 | 1,614 | -3.2% |
| 28 | 1,716 | 3.0% |
| 29 | 1,609 | -3.5% |
| 30 | 1,714 | 2.8% |
| 31 | 1,700 | 2.0% |
| 32 | 1,655 | -0.7% |
| 33 | 1,660 | -0.4% |
| 34 | 1,578 | -5.3% |
| 35 | 1,676 | 0.6% |
| 36 | 1,644 | -1.4% |
| 37 | 1,640 | -1.6% |
| 38 | 1,632 | -2.1% |
| 39 | 1,610 | -3.4% |
| 40 | 1,725 | 3.5% |
| 41 | 1,695 | 1.7% |
| 42 | 1,627 | -2.4% |
| 43 | 1,564 | -6.2% |
| 44 | 1,719 | 3.1% |
| 45 | 1,691 | 1.5% |
| 46 | 1,681 | 0.9% |
| 47 | 1,716 | 3.0% |
| 48 | 1,639 | -1.7% |
| 49 | 1,638 | -1.7% |
| 50 | 1,630 | -2.2% |
| 51 | 1,680 | 0.8% |
| 52 | 1,628 | -2.3% |
| 53 | 1,672 | 0.3% |
| 54 | 1,697 | 1.8% |
| 55 | 1,553 | -6.8% |
| 56 | 1,733 | 4.0% |
| 57 | 1,550 | -7.0% |
| 58 | 1,684 | 1.0% |
| 59 | 1,713 | 2.8% |
| 60 | 1,679 | 0.7% |
For comparison, a perfect d60 rolled 100,000 times: 95% of its faces land within ±4.9% of fair, and its most-off face is usually about 6.1% away.
d1001,000,000 draws · most-off face 2.5% (a perfect die: usually 2.7%) · p = 0.987 · Fair
See the numbers
| Face | Times | vs fair |
| 1 | 10,074 | 0.7% |
| 2 | 10,015 | 0.1% |
| 3 | 9,852 | -1.5% |
| 4 | 9,945 | -0.5% |
| 5 | 9,964 | -0.4% |
| 6 | 9,890 | -1.1% |
| 7 | 9,933 | -0.7% |
| 8 | 9,936 | -0.6% |
| 9 | 9,861 | -1.4% |
| 10 | 9,882 | -1.2% |
| 11 | 10,083 | 0.8% |
| 12 | 9,911 | -0.9% |
| 13 | 10,038 | 0.4% |
| 14 | 10,078 | 0.8% |
| 15 | 10,033 | 0.3% |
| 16 | 9,892 | -1.1% |
| 17 | 10,149 | 1.5% |
| 18 | 10,001 | 0.0% |
| 19 | 10,026 | 0.3% |
| 20 | 9,933 | -0.7% |
| 21 | 10,141 | 1.4% |
| 22 | 9,963 | -0.4% |
| 23 | 10,035 | 0.4% |
| 24 | 10,005 | 0.1% |
| 25 | 10,095 | 0.9% |
| 26 | 10,134 | 1.3% |
| 27 | 10,046 | 0.5% |
| 28 | 10,062 | 0.6% |
| 29 | 10,038 | 0.4% |
| 30 | 9,978 | -0.2% |
| 31 | 10,130 | 1.3% |
| 32 | 10,069 | 0.7% |
| 33 | 9,865 | -1.4% |
| 34 | 9,913 | -0.9% |
| 35 | 9,826 | -1.7% |
| 36 | 9,977 | -0.2% |
| 37 | 9,963 | -0.4% |
| 38 | 9,873 | -1.3% |
| 39 | 9,957 | -0.4% |
| 40 | 10,038 | 0.4% |
| 41 | 10,026 | 0.3% |
| 42 | 10,027 | 0.3% |
| 43 | 9,949 | -0.5% |
| 44 | 9,954 | -0.5% |
| 45 | 9,854 | -1.5% |
| 46 | 9,990 | -0.1% |
| 47 | 9,871 | -1.3% |
| 48 | 9,998 | -0.0% |
| 49 | 10,015 | 0.1% |
| 50 | 10,021 | 0.2% |
| 51 | 10,251 | 2.5% |
| 52 | 10,053 | 0.5% |
| 53 | 10,086 | 0.9% |
| 54 | 9,893 | -1.1% |
| 55 | 10,011 | 0.1% |
| 56 | 10,021 | 0.2% |
| 57 | 10,036 | 0.4% |
| 58 | 10,080 | 0.8% |
| 59 | 10,023 | 0.2% |
| 60 | 10,024 | 0.2% |
| 61 | 9,934 | -0.7% |
| 62 | 10,086 | 0.9% |
| 63 | 10,090 | 0.9% |
| 64 | 10,134 | 1.3% |
| 65 | 9,862 | -1.4% |
| 66 | 9,950 | -0.5% |
| 67 | 9,867 | -1.3% |
| 68 | 10,062 | 0.6% |
| 69 | 10,009 | 0.1% |
| 70 | 10,019 | 0.2% |
| 71 | 9,861 | -1.4% |
| 72 | 9,919 | -0.8% |
| 73 | 9,928 | -0.7% |
| 74 | 9,927 | -0.7% |
| 75 | 10,006 | 0.1% |
| 76 | 10,012 | 0.1% |
| 77 | 10,151 | 1.5% |
| 78 | 9,959 | -0.4% |
| 79 | 10,004 | 0.0% |
| 80 | 9,992 | -0.1% |
| 81 | 10,078 | 0.8% |
| 82 | 10,075 | 0.8% |
| 83 | 9,935 | -0.7% |
| 84 | 10,049 | 0.5% |
| 85 | 10,152 | 1.5% |
| 86 | 10,095 | 0.9% |
| 87 | 10,026 | 0.3% |
| 88 | 10,018 | 0.2% |
| 89 | 9,865 | -1.4% |
| 90 | 10,068 | 0.7% |
| 91 | 9,970 | -0.3% |
| 92 | 10,067 | 0.7% |
| 93 | 9,921 | -0.8% |
| 94 | 10,078 | 0.8% |
| 95 | 10,034 | 0.3% |
| 96 | 10,001 | 0.0% |
| 97 | 10,138 | 1.4% |
| 98 | 9,945 | -0.5% |
| 99 | 9,950 | -0.5% |
| 100 | 9,976 | -0.2% |
For comparison, a perfect d100 rolled 1,000,000 times: 95% of its faces land within ±2.0% of fair, and its most-off face is usually about 2.7% away.
d1201,000,000 draws · most-off face 3.7% (a perfect die: usually 3.0%) · p = 0.199 · Fair
Why this is still fair. Its most-off face was 3.7% from an exactly equal share, a little more than a perfect d120 usually shows (3.0%). But a perfect d120 rolled 1,000,000 times is 3.7% off or more about one time in 12, simply through luck. The test that decides fairness takes all of that into account, and this die passes it.
See the numbers
| Face | Times | vs fair |
| 1 | 8,428 | 1.1% |
| 2 | 8,362 | 0.3% |
| 3 | 8,411 | 0.9% |
| 4 | 8,577 | 2.9% |
| 5 | 8,339 | 0.1% |
| 6 | 8,338 | 0.1% |
| 7 | 8,293 | -0.5% |
| 8 | 8,176 | -1.9% |
| 9 | 8,333 | -0.0% |
| 10 | 8,308 | -0.3% |
| 11 | 8,023 | -3.7% |
| 12 | 8,140 | -2.3% |
| 13 | 8,291 | -0.5% |
| 14 | 8,348 | 0.2% |
| 15 | 8,119 | -2.6% |
| 16 | 8,335 | 0.0% |
| 17 | 8,250 | -1.0% |
| 18 | 8,284 | -0.6% |
| 19 | 8,394 | 0.7% |
| 20 | 8,326 | -0.1% |
| 21 | 8,139 | -2.3% |
| 22 | 8,524 | 2.3% |
| 23 | 8,304 | -0.4% |
| 24 | 8,312 | -0.3% |
| 25 | 8,453 | 1.4% |
| 26 | 8,381 | 0.6% |
| 27 | 8,242 | -1.1% |
| 28 | 8,219 | -1.4% |
| 29 | 8,145 | -2.3% |
| 30 | 8,344 | 0.1% |
| 31 | 8,219 | -1.4% |
| 32 | 8,288 | -0.5% |
| 33 | 8,522 | 2.3% |
| 34 | 8,357 | 0.3% |
| 35 | 8,185 | -1.8% |
| 36 | 8,251 | -1.0% |
| 37 | 8,418 | 1.0% |
| 38 | 8,392 | 0.7% |
| 39 | 8,435 | 1.2% |
| 40 | 8,336 | 0.0% |
| 41 | 8,368 | 0.4% |
| 42 | 8,273 | -0.7% |
| 43 | 8,510 | 2.1% |
| 44 | 8,231 | -1.2% |
| 45 | 8,461 | 1.5% |
| 46 | 8,359 | 0.3% |
| 47 | 8,296 | -0.4% |
| 48 | 8,368 | 0.4% |
| 49 | 8,311 | -0.3% |
| 50 | 8,357 | 0.3% |
| 51 | 8,336 | 0.0% |
| 52 | 8,267 | -0.8% |
| 53 | 8,310 | -0.3% |
| 54 | 8,415 | 1.0% |
| 55 | 8,412 | 0.9% |
| 56 | 8,372 | 0.5% |
| 57 | 8,207 | -1.5% |
| 58 | 8,328 | -0.1% |
| 59 | 8,185 | -1.8% |
| 60 | 8,439 | 1.3% |
| 61 | 8,279 | -0.7% |
| 62 | 8,283 | -0.6% |
| 63 | 8,144 | -2.3% |
| 64 | 8,234 | -1.2% |
| 65 | 8,477 | 1.7% |
| 66 | 8,342 | 0.1% |
| 67 | 8,343 | 0.1% |
| 68 | 8,555 | 2.7% |
| 69 | 8,478 | 1.7% |
| 70 | 8,303 | -0.4% |
| 71 | 8,377 | 0.5% |
| 72 | 8,407 | 0.9% |
| 73 | 8,235 | -1.2% |
| 74 | 8,450 | 1.4% |
| 75 | 8,409 | 0.9% |
| 76 | 8,300 | -0.4% |
| 77 | 8,274 | -0.7% |
| 78 | 8,509 | 2.1% |
| 79 | 8,212 | -1.5% |
| 80 | 8,329 | -0.1% |
| 81 | 8,365 | 0.4% |
| 82 | 8,405 | 0.9% |
| 83 | 8,258 | -0.9% |
| 84 | 8,339 | 0.1% |
| 85 | 8,421 | 1.1% |
| 86 | 8,348 | 0.2% |
| 87 | 8,344 | 0.1% |
| 88 | 8,313 | -0.2% |
| 89 | 8,320 | -0.2% |
| 90 | 8,518 | 2.2% |
| 91 | 8,418 | 1.0% |
| 92 | 8,393 | 0.7% |
| 93 | 8,301 | -0.4% |
| 94 | 8,514 | 2.2% |
| 95 | 8,295 | -0.5% |
| 96 | 8,350 | 0.2% |
| 97 | 8,334 | 0.0% |
| 98 | 8,369 | 0.4% |
| 99 | 8,354 | 0.2% |
| 100 | 8,471 | 1.7% |
| 101 | 8,334 | 0.0% |
| 102 | 8,312 | -0.3% |
| 103 | 8,422 | 1.1% |
| 104 | 8,290 | -0.5% |
| 105 | 8,295 | -0.5% |
| 106 | 8,342 | 0.1% |
| 107 | 8,365 | 0.4% |
| 108 | 8,390 | 0.7% |
| 109 | 8,384 | 0.6% |
| 110 | 8,302 | -0.4% |
| 111 | 8,236 | -1.2% |
| 112 | 8,240 | -1.1% |
| 113 | 8,348 | 0.2% |
| 114 | 8,250 | -1.0% |
| 115 | 8,341 | 0.1% |
| 116 | 8,320 | -0.2% |
| 117 | 8,268 | -0.8% |
| 118 | 8,317 | -0.2% |
| 119 | 8,318 | -0.2% |
| 120 | 8,210 | -1.5% |
For comparison, a perfect d120 rolled 1,000,000 times: 95% of its faces land within ±2.2% of fair, and its most-off face is usually about 3.0% away.
Test it yourself
You don’t have to take our word for it. Open the roller and tap FAIR DICE: pick any die and roll it 1,000, 10,000 or 100,000 times on your own computer or phone. It throws the same 3D dice you roll in the app, off-screen and as fast as your device allows, and draws the histogram as the results come in. Open the roller, or read more about whether digital dice can be fair.
