How Fair Are Your Dice?

[Stormwolf]

We've all had lucky (or unlucky) dice before... so for those who have that one die that only comes out of it's padded silk bag for life-or-death rolls that nobody else can ever touch... here's an interesting if slightly over the top article showing the statistical analysis of D20 rolls (http://arstechnica.co.uk/the-multiverse/2016/08/how-fair-is-your-d20/) using dice from various manufacturers.

[Eldonauran]

That was very interesting. Thanks for the link. Now I'm off to check the balance of my dice.

[The_Jackal]

That was very interesting. Thanks for the link. Now I'm off to check the balance of my dice.

https://en.wikipedia.org/wiki/Mersenne_Twister

This algorithm will likely be fair to an extent which far exceeds your requirements for playing a game.

[Stormwolf]

https://en.wikipedia.org/wiki/Mersenne_Twister

This algorithm will likely be fair to an extent which far exceeds your requirements for playing a game.

I'm a big fan of the Mersenne Twister for computer-based gaming but I'm so old school that I'd never use a computer to roll my dice for me in a pen and paper game. Sure they're really useful for storing spreadsheets and keeping game info at your fingertips, instead of scrabbling through binders full of A4 sheets covered in coffee and pizza stains, but there's just something about rolling dice that is satisfying in the way a computer can never simulate... also if you're the DM there's nothing so un-nerving for players as rolling dice behind a screen ;)

[Cheesegear]

We've all had lucky (or unlucky) dice before... so for those who have that one die that only comes out of it's padded silk bag for life-or-death rolls that nobody else can ever touch... here's an interesting if slightly over the top article showing the statistical analysis of D20 rolls (http://arstechnica.co.uk/the-multiverse/2016/08/how-fair-is-your-d20/) using dice from various manufacturers.

The salt-water test is very common in tabletop wargaming circles. People are often rolling upwards of 30 dice (D6s) on a routine basis. At n>30, you begin to get statistically significant numbers, and many TTWGers catch on real quick when someone's dice are regularly not rolling averages. One of the main offenders is Games Workshop, where an experiment was carried out in 2006 (http://www.dakkadakka.com/wiki/en/That's_How_I_Roll_-_A_Scientific_Analysis_of_Dice) that found that GW dice rolled 1s more often. However, GW has changed their dice production method at least twice (?) since then, and the experiment was only conducted the once. But, it was kind of big deal when the original experiment was carried out.

Now, however, the problem lies in GW's current 'limited edition' dice (not their normal D6s, only the LE ones), where ornate designs are engraved gouged out on the 1 side. The internet is much wider than it was in 2006, and there are many reports of pretty much everyone's 'Skull Dice' not rolling averages. Some people's roll 1s and 6s, some roll 2s and 5s. But, the pattern is emerging that you simply shouldn't buy GW's Skull Dice - or, in fact, any dice with huge engravings gouges on only one side.

I've heard stories that X-Wing's dice are also b0rked. But, nothing that I've tested myself.

[Winter_Wolf]

Interesting. But other than one set from ye olde BD&D red box with Elmore art on it, every other set I or anyone I've played with is the same brand: Chessex. Because that's all the game stores we've found ever carry. So we're all in the same boat that way; it's fair in the sense that were all equally blessed/cursed from essentially the same pool of probability. Except d6, that's just straight up Yahtzee dice because a 20d6 fireball was a thing way back. Everyone used them because for some reason counting pips is a lot easier than adding actual numbers. Seriously the fairest dice ever.

[Erloas]

I've had one set of dice (wargaming, so a large block of 12mm D6s) that never seemed to do any good so I stopped using them.

At some point my brother and I ended up with a few of the large red dice from a casino, we never tested them specifically, but we figured that a casino would have to have very good dice because so much money is riding on them rolling as expected. The most obvious things about those is that they are fairly large, (amazon says they are 19mm), they have very sharp edges, and the pips are painted or a flush inset (I can't really tell which) so there should be almost no difference in sides. They are transparent too so you can see they can't be weighted.

In general it seems like the 12mm dice tend to have more issues than the 16mm dice. Which I think would make sense, the lighter the dice the more influence small imperfections will have. My favorite dice are metal though, I've got several of them. Not enough for what many wargames require but they are the ones I always use when I can. They also seem to be the most reliable of the dice I have. I think for the same reason, they have so much weight that minor imperfections have a hard time changing their rolls.

[Crow]

I just did a float test with my d20.

Always came up 1. This explains a lot.

[Cheesegear]

The most obvious things about those is that they are fairly large, (amazon says they are 19mm), they have very sharp edges...

The most important thing about squared edges vs. round edges, is that even if the squared dice are weighted, a squared dice will roll less when it hits the table. A particlar nasty trick when using a squared dice, is that it's much easier to roll the same number again;

1. Roll a 6.
2. Pick up the dice, smoothly. The '6' face should now be touching your palm, with the '1' face visible.
3. Let the dice fall out of your hand; It should roll out of your palm, getting the '6' face, and then dropping straight down, thwunk. It wont roll when it hits the table, because it's bigger/heavier and has square edges, and physics. Another 6.

It looks like a real roll to an observer, because the dice rotates out of your hand. But it's not.


They are transparent too so you can see they can't be weighted.

I've seen plenty of transparent weighted dice. :smallamused:

[Siosilvar]

tl;dr: if fairness matters that much, make everybody use Gamescience dice in a dice tower. :smallwink:

That is somewhat difficult to do in fistful-o'-dice games like 40k or even things like Shadowrun, but if you build your own dice towers then it gets easier.