Are Mini Kraken's Dice Rolls Fair? A Look Inside the Dice Engine

Sooner or later, someone at the table says it out loud: "There's no way that's random. This thing hates me." A digital die can't clatter across real wood, so when a run of bad luck hits, it's natural to wonder whether the app is putting a thumb on the scale.

Fair question. It deserves a straight answer, so here it is up front: Mini Kraken's dice are fair. Every face has an equal chance, nothing is nudged, and — unusually — you can actually verify a roll for yourself. This article walks through exactly how that works, in plain language, so the next time the topic comes up you have something concrete to point to.

What "fair" actually means

Before we talk about our engine, it helps to pin down what a fair die even is.

A fair d20 means one thing: each of its twenty faces is equally likely — a clean 1-in-20 chance, every single roll. Not "roughly even." Not "even on average, with a little house flavor." Exactly even. A fair d6 gives each face a 1-in-6 chance; a fair d100, 1-in-100. That's the whole definition.

Notice what fairness does not promise. It doesn't promise you'll see all twenty numbers if you roll twenty times. It doesn't promise a bad streak will "balance out" soon. It doesn't promise the result will feel even. Fairness is only ever a statement about the odds on each roll — and on that, we don't compromise.

The engine under the hood

Mini Kraken doesn't use a homemade shortcut for rolling dice. Every roll — on the website, in your character sheets, on the virtual tabletop, and through the Fortuna Discord bot — runs through the same shared dice engine, built on a mature, open-source dice library that has been rolling dice for tabletop players for years.

That matters for two reasons. First, an open-source foundation means the math isn't a mystery box; it's code that anyone can read, test, and pick apart. Second, it means we're not reinventing something delicate. Randomness is one of those areas where clever homebrew tends to introduce subtle bias, and battle-tested libraries exist precisely so nobody has to.

The important part is what the engine does when it needs a number.

How a number gets picked — and why there's no hidden bias

Here's the part most people never hear about, and it's genuinely the heart of "is this fair?"

The naive way to fake a die roll is to grab a random value and chop it down to size with leftover division — the programming equivalent of "divide and take the remainder." It's quick, and it's almost fair. But it has a famous flaw called modulo bias: because the range you start from rarely divides evenly by the number of die faces, a few faces end up very slightly more likely than the others. On a single d20 the skew is tiny, but it's real, and it's exactly the kind of invisible thumb-on-the-scale that makes "rigged" accusations stick.

Our engine refuses that shortcut. Instead of forcing an awkward number to fit, it uses a technique often called rejection sampling. In plain terms: if a raw random value falls into the tiny leftover sliver that would skew the odds, the engine throws it away and quietly draws again. It only keeps values that map perfectly evenly onto the die's faces. The result is a distribution with no bias at all — every face lands on its exact, honest probability. You never notice the occasional redraw; it happens in a fraction of a millisecond. But because of it, a Mini Kraken d20 is as even as mathematics allows.

That's the technical spine of the whole answer. Fair isn't a marketing word here; it's a property of how the number is chosen.

The 3D dice are honest theater

Mini Kraken is known for its real 3D physics dice that tumble across the screen and settle — and this is where a reasonable person might get suspicious. If a physics simulation decides where the die lands, couldn't the animation be tuned to land on low numbers?

No — and the reason is the order of events. The engine picks the number first. The animation is then told to land on it. The tumbling die you watch is a faithful performance of a result that was already decided fairly, a beat earlier, by the same unbiased engine described above. The physics is there for the feel — the satisfying clatter and roll that flat text results flatten out — not to determine the outcome. It's showmanship on top of honest math, and it literally cannot nudge the number, because the number already exists before the die starts moving.

One engine, everywhere — so a roll means the same thing

A subtle source of trust is consistency. On some platforms, a roll in one place behaves differently from a roll in another. On Mini Kraken, a d20 is a d20 whether you tap it on your character sheet, drop it on the virtual tabletop, or type it into a Discord channel where Fortuna is listening.

That's because all of these surfaces call the same engine. There isn't a "friendly" roller for the website and a different one for Discord. A roll carries the exact same meaning across the whole ecosystem, which is both a convenience and a guarantee: there's no special mode where the odds quietly change.

You can actually verify a roll

Here's the feature that turns "trust us" into "check for yourself."

The engine supports deterministic replay. In plain terms, a roll can be tied to a small starting value called a seed, and the same seed always reproduces the exact same sequence of results. Feed the engine the same seed twice and you get identical dice, every time. This is the opposite of a rigged system: a rigged roller has to hide how it works, while ours can hand you the recipe and prove that the same inputs always produce the same, reproducible output. Randomness you can replay is randomness you can audit.

What we never do

Being fair is as much about restraint as technique. So, plainly, here is what Mini Kraken's dice do not do:

• No "dice karma." We never track that you've been unlucky and secretly hand you a good roll to compensate — nor the reverse. Each roll is independent, exactly as real dice are.
• No hidden fudging. There's no invisible "smoothing" that trims your extremes to make results feel nicer. Some tools quietly do this. We don't; a streak is a streak.
• No secret modifiers. The only bonuses applied are the ones you can see in the roll — your sheet's modifiers, advantage, and so on. Nothing is added or subtracted in the dark.
• No targeting. The engine doesn't know or care who you are. It isn't nicer to the Game Master or crueler to the new player. It has no idea what a "player" even is.

The dice are indifferent. That indifference is the point.

"But my rolls are still weird"

They might genuinely feel that way — and that feeling is real even when the dice are perfect. True randomness is lumpy: it produces streaks, clusters, and cruel timing far more often than our intuition expects, and our memories cling to the disasters while quietly discarding the forgettable average rolls. A fair d20 will absolutely hand you three 1s in a row sometimes; that's not evidence of bias, it's evidence of randomness doing its job.

That gap between "fair" and "feels fair" is big enough to deserve its own article — see Why Fair Dice Feel Rigged for the psychology behind bad-luck streaks. And if you want to satisfy the skeptic in you empirically, the honest test is volume: roll a d20 a few hundred times and watch the spread flatten toward even. A dozen rolls will always look lumpy; a thousand tell the truth.

Roll with confidence

So, is Mini Kraken fair? Yes — by design, not by promise. Every face is equally likely, the math behind each number is unbiased down to the decimal, the 3D animation only performs a result already chosen honestly, the same engine runs everywhere, and a roll can be reproduced and checked rather than taken on faith.

The dice don't love you and they don't hate you. They just roll. And on Mini Kraken, they roll straight — so you can stop wondering about the machinery and get back to the story, where the good stuff actually happens.