"How do you know a die is fair" Topic

How many rolls is it accepted you need to make to ensure a die is fair. What is the range of elasticity in the distribution?

To complicate things, I want to test Fudge dice. These have two + sides, two – sides, and 2 blank. I can't necessarily tell which symbol has come up – one + might be over represented, the other underrepresented, but average at 1/3 of the rolls. Does this matter?

Get some 'casino' dice.

Casino dice are very expensive. Do not jump and get one that has been drilled by the casino before drilling.
I worked with Lou Zocchi at Gamescience for a few months in 1984 helping him with a new retail shop.
Sharp edged dice are better than rounded edge dice. The rolling machine cannot ensure that each edge is hit an equal number of times.
Any die with additives such as glitter should be avoided. The dispersal of the flakes will not be even.
Combo color, swirls. Cannot not be balanced if each color has a different weight.
Look for and avoid dice with air bubbles. They are almost always unbalanced.
I tend to buy opaque and for them a number of times to check balance.

When they roll 6 at the beginning of the game, and 1 at the end.
Life is too short to worry about die rolls. 🤷

Jim, also Casinos don't use Fudge dice!

I would expect that 98% of commercial dice are fair enough for gaming.
I expect more bias is introduced by gamers who are 'flexible" when measuring or remembering to implement rules.
It would be very time consuming (and irritating) to ask all players to roll their own dice 100 times and note down the uniformity or otherwise of their scores.

martin

The general rule for statistics is you need 20 data points, but the more the better your assurance. For a 6 sided die I'd go with a multiple of 6 so role it 24 times if you don't get at least 3 of every number or you get 6+ of a number it's possibly biased, roll it another 24 times, if that repeats the previous data you have a bias, if it doesn't you probably don't.

I always try and account for possible dice bias in my rules; somewhere you need high scores, somewhere else you need low ones, but it doesn't seem to be something commercial rules writers think about.

When it comes to dice, does size matter? 12mm better than 20mm or worse or doesn't it matter?

DBDD

(1) If you have three possible outcomes, and each one comes up 1/3 of the time, the die is fine for gaming purposes.
(2) If you have 12 identical dice, minor defects will even out and only serious trouble in manufacturing--fairly obvious--or deliberate cheating can cause troubles.

Buy your dice by the dozen, and don't mark or segregate the "lucky" die. Then focus on tactics.

I always offer to allow my opponent to use my dice if they want.

Everyone knows they only roll hot for their owners.

I once read a blog post by someone complaining how hard it was to create biased dice. Even adding weights seemed to have little impact. So don't worry.

I was once told to float them in a bowl of water; do they float? If they do they should float happily with any face uppermost.

There are videos on YouTube about testing dice in salt-water solutions: they must be of a certain size, and density for this to work; if size and density are correct, the die will float, so you spin it, in the salt-water solution, and see if it comes up to the same number, too often (unbalanced/biased towards that number), or not (comes up to a different number very frequently).

Otherwise, there is a technique called chi-square, I think. This requires the recording of 10,000 rolls to track the results. By the time you roll a given die 10,000 times, all of the corners will have worn off, making it mostly round, and so the entire test will be ruined, as the die is no longer intact…

To be honest, I roll my dice through a hardwood Dice Tower. My numbers are consistently terrible to the point where my gamer friends told me, often, to not use the Tower. My results were staggeringly random -- their die rolls, dropping them onto the tabletop from the palms of their hands, were staggering favorable, in comparison! I am being absolutely serious here.

Take a look at these two videos: one die at a time, handful at once. These were made by the fellow who built my Dice Towers (Hickory and Padauck hardwoods; the videos show Bloodwood, which is even harder than Hickory and Padauck). Hardness of the wood matters: the harder the wood, the higher the pitch of the clattering of the dice inside the Towers, and the more they bounce and spin inside of it, as they travel through it.

Any hardwood Dice Tower, with at least two baffles, will give you truly random dice rolls. I've used my Towers for almost 10 years, rolling single dice, and handfuls at a time. They are truly random, every time. By the way, I have three Hickory, and one Padauck; my two sons have a Wenge, and a Black Walnut Dice Tower. They, too, are truly random in the die results they get, as well

The really nice thing about our Towers, are the trays. I can roll, literally, 50 d6, at once, and they stay in the tray, without knocking over figures on the tabletop, and they won't roll off of the table onto the floor, to God knows where… I love my Dice Towers. Cheers!

Because in my solo games they roll equally bad for both sides
My home made dice tower dice go into together kicked to opposite side, the back to other side then to back wall to get kicked out to a tray

"Casino dice are very expensive. "

And also flammable. ;-)
I set fire to an old beat-up pair. They burned with intense flame. Not something you'd want to keep in the house… ;-)

Oh, just go to Dollar Tree and buy a few new packs if you're so worried. 🤷

A lot of "crooked" dice issues come up from not enough time in curing before being packaged and sent to the stores. your most likely dice for this are going to be Bicycle dice you get from target, Wal-Mart, etc. I know someone who bought half a dozen boxes and tested to find the suspect dice and there was always at least one and often two per pack that continuously came up the same number when doing the water test.

I've actually never thought about it in the wargaming context.

Gambling, on the other hand …

It is highly unlikely that any commercial dice are significantly biased enough to overcome the chaotic effect of being shaken in a hand or cup and tossed or dropped onto a table (or into a dice tower). They're just too small, and whatever differences might be made by internal weight distribution or edge size and length will be so marginal as to have little overall impact, if any.
Seriously, "glitter"? If glitter dice contained so much as a single milligram of glitter material throughout the die, I'd be stunned. There's no way that minuscule amount of mass will affect the energetic effect of rolling a die to any meaningful or measurable degree.

As for the chi-square test, you get all sorts of people claiming all sorts of things as to how many rolls "have" to be made to accurately assess a die. I don't really give them credence.

So I wouldn't worry about it. Use a dice tower or dice cup.
And wood isn't necessary. I designed my own paper dice tower built out of printed cardstock. Works great, and the dice rattle around quite effectively. Which is quite sufficient to produce an acceptable level of randomness in my book. It's not about the dice, it's about the kinetic energy when they are rolled. Shake ‘em and shoot ‘em.

Technically, for chi-squared or similar statistical analysis, the number of rolls required depends on the variability. So really wonky dice will be identified quicker

I like the point that the die will wear during testing, which if uneven, will "wonk" the die all by itself

If this stuff really upsets you, I have seen dice dungeons available where you can lock your dice away until they repent and see the error of their ways

I sometimes wonder if there is a market for a dice crushing machine. With or without an evil laugh as you reduce it to fragments

John

Billy, I think you're confusing dice with witches.

My dice take bribes. They are mean little gits.

Billy, I think you're confusing dice with witches.

So if the dice weigh as much as a duck…

Silurian! No I'm not, if the dice float you burn them!

No-one has mentioned the person rolling the dice!

is there any die rolling when it comes to delivering goods that a customer has paid for ? asking Martin Goddard

link

Takes your money but doesnt send the goods
Martin goddard of Peter Pig is the type of small businessman who takes payment for an order, confirms the order, then doesnt send the order, then says they have issued a refund, then doesnt issue the refund and then refuses to communicate with the customer about what is happening to the point where by the customer has to issue a final demand letter but then still does not issue a refund or send the goods which have been paid for.