"Testing for fair dice" Topic

I have 5 aluminium dice, in their own case with foam cut outs, bought from a "£" store. On critical die rolls I sometimes produce them with a flourish saying "Time for my SPECIAL dice" to howls of claiming they are not fair.

Rather than rolling each one 600 times and recording the results, is their a quick and dirty way to check they are fair.

(My personal belief is minor imperfections are present in most dice, and a proper roll will introduce to many other variables of physics- bounce etc- for these to matter. However a significantly bias dice will skew the stats.)

Rather than rolling each one 600 times and recording the results, is their a quick and dirty way to check they are fair.

No… but if they are really unfair, it may not take 600 rolls to prove it.

Maybe you could announce before the game that you ahve them and that you are saving them for critical die rolls. That might allay any unfounded suspicions. Although I doubt that I would be able to roll any sixes with them.

Or just use a pool of dice for both players and include them.

That way you can both suck with the same dice!

(Yes, I had one of those 'anything but a 1' nights at gaming this week!)

Don't casinos have some type of spindle a die can be put on that pretty quickly determines if a die is loaded?

Found this online:

Loaded Dice

The Pivot Test
A loaded die is one with hidden weights inside. Since the weights are usually spread around the edges of the die so as to conceal any obvious marks as much as possible, the pivot test is perfect to tell whether or not the die is crooked. To test the die, hold it loosely between your thumb and forefinger by two diagonally opposite corners, putting as little pressure as possible and trying different combinations. If the die is loaded and one of the weighted sides is on top, the die will pivot as the heavier side swings around to the bottom. The movement is unmistakeable.

The Sinking Test
This is one of the best tests you can perform on a loaded die. Hold the suspected die just above the surface and drop it gently into the water, with a different number up each time; watch whether or not the cube spins as its sinking to the bottom of the glass and observe what number is facing upwards when the die lands. If the die lands on only one of two or three numbers, the die is loaded.

Bevelled Dice

The Wobble Test
A bevelled die with one of more sides sandpapered so that they are more rounded rather than flat and are more likely to land on one of the flat sides more often than probability should allow. To check for bevelled dice, hold the two dice and rub their sides together, trying different sides and combinations. The dice will 'wobble' back and forth if they are bevelled instead of lying flat against each other.

Allow your opponent to use them as well. What's good for the goose…

I wouldn't allow "special" dice to be pulled out during a game and then put away directly after a roll. Loaded or not.

I suspect that someone is taking this too seriously.
-Kle.

I would in RPGs and games between friends.

Acquaintances at shops, clubs and tournaments, this is a horrible faux-pas.

Unless you're friends it is important to avoid even the appearance of cheating.

We had a certain person that had loaded dice at our old club.

We then brought in a club rule, all players use the same dice. Solved the problem and he left shortly after!

I'm with Tin Dictator. We had a JRIII game where an individual did this. after complaints from his opponent we found 1 die had 2 5s the other 2 6s. "He" was not invited back, ever.

I'm with the `everybody uses the same set' lot here.

I haven't played D&D for years but if anyone did this at one of my D&D games, the dice became mine for the night and were rolled only when used against the joker that brought them.

Funny thing, no one ever did this twice.

mjc

5 dice? What is that, a Yahtzee set?

The issue here is if any dice are actually more lucky than normal, it's the same as if they're loaded -- they won't produce a fair distribution and it's cheating to use them. And if they're not lucky, there's no game reason to pull them out at any special time.

I'd say actually, put the dice in your backgammon set and stop pulling them out during miniatures games.

If you must use them for tabletop gaming, put them out in their case with the other dice at the start of the game. The fact that they're in a silly case all by themselves will signal to other players they are special and you don't want them lost (after all, they cost what, 20 pence each?).

The point is, as others have said, if other players can use them when they want, you've made the game fair -- whether or not the dice themselves are cheater dice.

I have never worried about dice one way or the other. I would be shocked if I ever caught a cheat.

My son did a Science Fair project while in school a few years ago. The premise delt with both probability and 'fairness' in dice. A lot of the teachers were in awe as they were not familiar with any dice other than d6. he won quite a few awards and prizes for the project.

In regards to d6, the only truly balanced, 'fair', dice are those sold by casinos which are tested to ensure balance.

Gaming dice tend to have between a 5% and 15% variance in the number rolled based primarily on the construction of the dice. So the more edges that a dice had the less the variance will affect the overall 'fairness' of a die. This is why, I personally preferd10 dice to d6 dice when I am game designing.

Things such as air bubbles and the different density of of material used to compose dice (swirls) will throw a dice out of balance. Another factor and one which is very popular today are the 'rounded' edge dice. These are less balanced than the 'sharpe-edged' dice.

Same dice for everyone.

This is a fascinating little subject you've stumbled upon.

To add to the aformentioned tests. There's the Caliper Test. Take three 6 sided dice of the same dimensions as the die you're testing and set them end to end on a table. Push the center die just slightly out of line--still touching the outside two but leaving enough space to fit about half the test die. The three dice now form a poor mans caliper that can be used to test the other die. Try placing the test die into the caliper in all 6 facings. The die is 'loaded' if it pushes the caliper apart. You'll want to hold the caliper together with your other hand so you can feel the failure.

Here's an aritcle that goes into some more detail…

link

Psychology plays apart because many people actually cheat by accident. People are generally honest and don't specifically search out loaded dice to use against their oppenents. But people actually do just this all the time. Ever no anyone to lable a d20 as a bad die and throw it away (or smash it with a hammer, or throw it off a balcony, …) People who do this are culling the herd and will start to generate a pool of dice who are either [near] perfect or that are biased in favor of their favorite game. This essentially magnifies manufacturing defects because the bad ones no longer exist to average out the good ones. This person is essentially 'cheating' even if they don't mean to.

Game design is another branch of this discussion. The later versions of D&D have actually gone in the opposite direction of what's needed to make a 'fair' game. While it's certainly easy to understand rules if high dice rolls are good and low rolls are bad. But, it makes it much easier to cheat either by accident or by design. Games that alternate successful restults (high die to hit, low die to save, …) don't reward or punish dice for imbalances--players will actually prefer dice with even distributions. Watch out if players use one die for saves and one for hits :)

Another game design factor is to use big dice (d20s) or little dice (d6s). Big dice actually suffer more from manufactoring defects more then little dice. The surface area for the little dice is larger per number and is less swayed by minor defects in dimension and weight. Big dice with small surface area per number are lot more vulnerable. Therefore little dice do a better job then similar sized big dice for creating a fair game.

What a great topic.

Sumo, very interesting stuff, thanks.

Rocky, one of my friends who is the local Battletech "Commando", uses a bunch of casino dice. Apparently after so many times used, they make a little make on one face and discard them. No casinos around here and on my last trip to Michigan last year, my wife and I went to a new Casino in northern Michigan (Odawa in Petosky) and she tried to get dice for me, but they weren't selling them.
--
Tim

Oh and I missed Andrew's post as I was typing. more interesting stuff.

I think for many of us, me certainly, dice are fascinating objects. I mean just look at what we do with them as wargamers and gamers.
--
Tim

I read an article that convinced me that small round-cornered pip dice have nearly twice as much chance as rolling a 1 as a 6. Not good news if you play DBA!
I promptly ordered 10 x D6 from a Casino supplier – expensive, but well worth it for peace of mind!
Ex Casino dice can be picked up cheaply on E-bay (Casinos bin them after each session).
Ian

In addition to fairness, people who only use their own dice are a pain in the rear after the game. "One of my dice is missing! Who's got my red and white d6 with the gold pips?"

Personally, I would roll the die a couple hundred times. Its the only way to be sure.

If you want to proof it, roll it 180 times, keep a log, date and initial it, present it along with the die at the start of the game.

I wouldn't have a problem with something like this from an opponent. After all, if I didn't trust him somewhat I wouldn't bother playing with him at all. After all, if you can trust them to be pleasant company, good sports, and so on, trusting them not to cheat is easy.

I have rolled various cutesy dice hundreds of times over the years, just because I'm curious and inherently skeptical. I don't trust the die-within-a-die things, but those wooden dice work fine, even though a lot of people doubt them.

I've never had a problem with smaller dice, either.

Andrew

Used them tonight, and no one seemed worried by them. At our club people tend to bring their own dice. Wanted to use them because they are so unusual, and the fact they have their own padded case.

I seemed to be getting a full range of throws out of them. Memory hints there may have been a lot of 1s and 6s, but one thing everyone should know is never trust human memory and perception- it tends to find 'patterns' in randomness (amazing coincidences aren't really). Also I hit '5' 3 times on the trot during initiative, which was perfect for a Chain Reaction 2 game. Given CR2 sometimes needs high, sometimes needs low it didn't matter, but I'd like to be sure so I can keep using them.

Andrew, I think you should throw your dice 100 times, and report back to the forumn! If you throw a die you need to feel confident that there is a one in six chance of rolling a six. Otherwise your rules don't work!
Ian
PS I think this subject was closed by a Professor from Indiana University.

If you want to proof it, roll it 180 times, keep a log, date and initial it, present it along with the die at the start of the game.

180 times isn't even close to being enough times to ensure that it is not tilted slightly. If you got 20 of each result out of 180 rolls then I would be pretty much convinced you WERE cheating.

"Proof" would need to be done by physical testing with precision, not randomised rolling.

A d6 that is fair will average (16.6) 16-17 times per number. So for example a sharpe-edged 5% variance of a d6 means that the #1 may roll .8 times the number, more or less. so the number would fall between 15-16 times or on the high side 17-18 times.
It is even less for a d10 at .05% or a d20 at .025. Not really enough to worry about.

When using rounded dice or other flawed dice the variance can reach 15% which could have a greater impact at a 2.4 number shift so a d6 could run 14 rolls on the low side to 20 rolls on the high side. Those extra rolls will show up in break-down of the other numbers somewhere.

I also work for Zoochi for a period of time which gave me the science project idea for my son.

On a similar subject, don't you hate dice thieves. This happens at my bi-weekly Magic the Gathering Games, usually because almost every wargame I've played one player provides all the dice.

Anyway, I have a container with a collection of very small d6s for marking cards. They're from various sources including Epic 40k, BFG (I think) and Pirates of the Something Something.

While people complain about their small size, it's really hard to place the Pirates die on a specific face, people are always grabbing them.

They don't ask, and it's often in games I'm not involved in, so that I cannot keep track of them. I do wish everyone brought their own dice.

On the other hand, it's probably easier for my sanity to buy a block of small d6s in a couple of colours for communal use.

Back in the 50s some college kids went to Vegas and started recording roulette wheel results. They found that, due to manufacturing and material flaws certain numbers tended to come up more frequently than others. They went back with a bunch of cash, bet those numbers, made a killing, and got banned from the casinos. Now, the casinos regularly swap around their wheels.

That was Zocchi of course. Sometimes 'unbalanced' dice is a game mechanic. There was one WW2 naval game about Pacific battles made by a Japanese company. 'Yamamoto', I think.

Anyway the game came with d6 that had an extra large rising sun as the #1 pip. Being a large groove made the one lighter so a higher percentage of #1s would be rolled. The combat charts had all very low numbers as hits for the Japanese.

It seems to have been a design to produce a sort of play balance. A totlaly different game with other dice as the Allies never lost.

Why do we operate under the assumption that casino dice are fair? Those guys only make money if you lose… and they make a lot of money. So are they motivated to have fair dice, or dice that favor their house games? Hmmm… "Fair play" and "casino boss" don't quite match up in my mind; "casino patron" and "sucker" do.

I've never bothered to test any of my dice. I often use dice for silly "thematic" reasons: If I'm attacking, I like to use red dice (for "blood"). "Evil" forces use black, red and green dice. "Good" forces use blue, yellow and white dice (and maybe green— seems to go with elves). Yellow is used to roll for treasure, etc. And if a die rolls "badly" during a game, I may switch to a different die for that game, but I don't bother to set aside dice either way for future games. (And my switching never seems to help anyway).

Honestly, I think the flaws in commercially available dice amount to a neglible effect either way. The simple action of tumbling should overcome any minor irregularities, at least in the course of a game. I'd be more concerned about someone who tried to "drop" dice for the number they wanted, rather than whether a certain set of dice might be theoretically "unfair."

And if you're playing with a cheat, that should become obvious in a lot of other ways… not the least of which is that he or she will probably also be an insufferable jerk.

The trouble with throwing a die 100/180 times is it could still come up 1 every time. While this is not likely, there is still a probability for it. The die has no memory. Every time you throw it , there is a 1 in 6 chance of 1 coming up.Casino dice are made clear so it can be seen that no lead is inside. Most crooked craps games in fact or fiction, involve switching the dice when it is the cheater's turn, then switching back when play passes to the next player. Fair casino dice does sound like an oxymoron. But the odds say the casino still has a better chance of winning than you do.

Why do we operate under the assumption that casino dice are fair? Those guys only make money if you lose… and they make a lot of money. So are they motivated to have fair dice, or dice that favor their house games? Hmmm… "Fair play" and "casino boss" don't quite match up in my mind; "casino patron" and "sucker" do.

I can't speak for casinos elsewhere, but in the USA there are rather serious legal penalties for any Casino that gets caught using unfair dice (or anything else that could be construed as cheating). By serious penalties, I mean large fines for relatively small problems – fines large enough that the potential cost of cheating far outweighs the potential benefit. For major problems, they are looking at jail time and/or having the establishment shut down.

They don't make their money from cheating, but rather by setting the payoffs such that they end up winning in the long run.

As the simplest example, look at Roulette. The payoff for any given bet is exactly the same as the odds of that result happening if there were no Green (0 or 00) results on the wheel. Add the 0/00 in, and the house gets a small margin. 2.7%(1 in 37) for a wheel with only a 0, 5.3%(2 in 38) for one with both a 0 and a 00.

For blackjack, their edge comes from "ties go to the dealer".

And so on… they make money because the odds are slightly stacked in their favor, and given the amount of money wagered.. slight * billions = big profit.

Oh, I know. I was being a tad facetious. It just struck me as ironic that given the general image of casino owners as being closely tied to organized crime at the worst, or at least on a hand shaking basis with the same at the best, the first place everyone turns to as a source of "honest" dice would be casinos.

Oh the dice are fair- its the ODDS that are stacked. Otherwise why would they get so upset about card-counters.

I'd be more concerned about someone who tried to "drop" dice for the number they wanted, rather than whether a certain set of dice might be theoretically "unfair."

Ooohhh, now you've hit one of my buttons!!!!

I hate the 'drop dice' brigade.

I've encountered a number of these players (balancing dice on their fingertips and then using this odd 'wrist flick'!).
Really irritates me!

I've been close, several times, to bringing a cup to games and insisting that dice rolls are performed from the cup.

Guys,

I haven't been to Vegas in 10 years. In the day in blackjack ties were pushes with no one winning. I suspect it is still that way. Casinos have gotten rich off all of the fools who believed they could count cards. I was one who very quickly learned that to count cards you had to work at it 365 days a year.

Because players can bet the come or don't come, biased dice in craps would be death to the casinos if they did not frequently change them out. The smart guys would immediately figure it out and bet appropriately.

When I went to college in the late 60s I'd supplement my income by playing street craps behind the Lexington Market in Baltimore. In street craps there was no bar 12 for no bettors. Thus, don't come or don't passs was a winning bet.

It never ceased to amaze how humans couldn't bet no. The trick was to win just enough but not too much while complaining about losing and burying your winnings in your pocket. Of course the blacks raised on street craps always look on whites as marks. Thus, the guys always thought of me as a loser and I was always welcomed.

Another college kid, a math major at Johns Hopkins, gave me the best, though not statistically supportable, gambling advice. If the dice have passed 4 times in a row, all things being equal, you have to bet pass or walk away.

What all things being equal meant was subjective. Ten years ago I walked in on a guy who had rolled 5 passes at the Rio. I bet the pass and come and maximum odds and he made 3 more numbers. At that point his wife came to the table and put her hand on his back asking when they were going to dinner. He started and would have jumped a foot in the air if he were not drunk. I immediately took my odds down and watched him roll a 7 on the pass, craps. Such is gambling. And such are odds.

There is no guarantee in die rolling and there is no guarantee in life.

Good die rolling.

Bob Coggins

I have written a VB program to count the rolls- How far from 16% would you say a set of rolls has to go before you would suspect the dice

and also
TMP link

The program relies on 'buttons' on screen at the moment (its the most attractive code I have written, but using the keyboard would be quicker)