Blog
Dice Superstition and the Tables That Keep It
Retiring a cursed d20 is irrational, harmless, and probably good for the game.
Benching a cursed d20, rolling only in the tray, not touching another player's set — table superstition is irrational, mostly harmless, and probably good for the game.
There is one version of it that is not harmless, and it is worth separating out.
The rituals, and why they survive
Every long-running table develops them. A die that rolled three ones in a session gets benched. A particular tray is lucky. Dice are not shared. Some players roll left-handed after a bad run.
None of it affects anything. Each roll is independent, dice have no memory, and the benched d20 was behaving exactly as a fair d20 does — three ones in twenty rolls has a probability of about 7.5%, which happens somewhere at most tables most sessions.
They survive because they are cheap and they do something real: they give a player a response to a run of bad luck other than being annoyed. That is a genuine function, and it is not a mathematical one.
What three ones actually means
In twenty rolls of a d20, getting three or more ones has a probability of about 7.5%. Across four players and a full session that is close to certain to happen to somebody.
Distinguishing a genuinely biased die from ordinary variance takes hundreds of recorded rolls, not one memorable evening. A frequency count over a few hundred rolls will detect a serious manufacturing defect; a session will not detect anything.
Casino-grade dice exist precisely because ordinary dice are not perfectly fair — moulded pips remove material unevenly and the difference is measurable with enough rolls. It is also far too small to be what anyone at a table is reacting to.
The version that is not harmless
There is a line, and it is when superstition stops being about a player's own dice and starts being about someone else's conduct. A table where a player is blamed for a run of bad rolls, or told they broke the luck, has converted a harmless ritual into a way of making one person responsible for chance.
It also shows up as a rule that disadvantages someone — insisting a particular player rolls in the open while others do not, or attributing a party wipe to whoever rolled last.
That is worth interrupting, and the interruption is a probability argument for once: the roll was independent, nobody caused it, and the alternative is a table where being unlucky is a social offence.
The genuinely useful ritual
One convention is not superstition and is worth keeping for real reasons: rolling in a tray, in the open, where everyone can see.
Dice that roll off a table are re-rolled, and a re-roll is a small unaudited decision. Dice rolled behind a screen or into a lap produce results only one person saw. Neither is a fairness disaster and both are the shape of thing that erodes trust slowly.
So keep the tray. Not because it is lucky, but because a visible roll is a settled roll, which is the same reason a spun wheel ends an argument that a silent draw does not.
The one measurement worth doing
If a table genuinely wants to know whether a die is fair, the test is straightforward and almost nobody runs it, which tells you something about what the superstition is for.
Roll the suspect die two hundred times and tally the faces. For a d20 that is ten expected per face, which is too few to conclude anything — you need closer to two thousand rolls before a real manufacturing defect separates from noise. That is about an hour of rolling, and the fact that nobody does it is the clearest evidence that the ritual is doing social work rather than empirical work.
Casino dice exist because the empirical question has a real answer: precision-machined, sharp-edged, with filled pips so the faces have equal mass. Ordinary moulded dice are measurably worse and the difference is far too small to be what anyone at a table is reacting to.
The rituals that are actually about pace
Some table conventions look superstitious and are doing something practical, and it is worth telling them apart before abolishing anything.
Rolling in a tray keeps dice on the table and settles disputes about a cocked die. Announcing the target before rolling stops the argument about whether a 14 was good enough. Passing a set of dice rather than reaching across the table is about not knocking over other people's minis. None of these are about luck and all of them are followed with roughly the same ritual seriousness.
The genuinely superstitious ones are the ones about outcomes: benching a die, changing hands, refusing to let anyone else touch a set. Those are harmless and worth leaving alone, because they cost nothing and they give a player something to do about a run of bad luck other than be irritated at the table.
Frequently asked questions
Is a die that rolled three ones cursed?
No. Three or more ones in twenty d20 rolls has a probability of about 7.5%, which happens to somebody at most tables most sessions.
Could a die be genuinely biased?
Ordinary dice are not perfectly fair — moulded pips remove material unevenly — and the effect is far too small to be what anyone notices at a table.
How would I detect a biased die?
A frequency count over several hundred recorded rolls. A single session detects nothing.
Is superstition harmful?
Usually not. It gives a player a response to bad luck other than being annoyed, which is a real function.
When is it harmful?
When it becomes about someone else's conduct — blaming a player for a run of bad rolls, or applying a rule only to them.
Is rolling in a tray superstition?
No, that one is practical. A visible roll is a settled roll, and rolls only one person saw erode trust slowly.