# Where to get cuboid dice?

Anybody know where I could get some cuboid (but not cube) dice for use with teaching probability? Or advice on how to make some without a 3d printer?

Edit: In response to comments seeking clarification, I am indeed in search of the type of dice referred to in this question

https://math.stackexchange.com/questions/581900/probability-of-rolling-a-cuboid-dice

Eichler, A., & Vogel, M. (2014). Three Approaches for Modelling Situations with Randomness. In Probabilistic Thinking (pp. 75-99). Springer Netherlands.

Specific use case: I am leading a workshop for teachers on the 7th grade probability related common core standards, and such dice seem like a very nice way to illustrate the transition from the classical approach to probability to the frequentist approach, while also hitting standard 7.SP.C.7.b.

• You can get a "100+ Pack of Random Polyhedral Dice in Multiple Colors Plus Free Pouch Set by Wiz Dice", all for \$20 on Amazon. amazon.com/Random-Polyhedral-Dice-Multiple-Wiz/dp/B009R6J8RY/…
– user173
May 2 '14 at 2:43
• @MattF: I am not sure this answers the question. These dice typically are not biased, while it appears this is what OP wants, asking for cuboid dice. // Jason: could you clarify what exactly you are looking for and what is your intended usecase more precisely.
– quid
May 2 '14 at 8:06
• I suggest to steal a block out of the nearest childs building block set and putting stickers with numbers on it. May 2 '14 at 10:44
• @MattF. I do not only want non-uniform probabilities. I also want a random generator whose probability distribution cannot be deduced. In the examples you mention one could easily write down the probability distribution without having to collect data. May 3 '14 at 2:45

I can think of a couple of possibilities, though I know of no ready-made source of cuboid dice.

• Eric Harshbarger makes dice for fun and profit. I, myself, have purchased a number of his math-related dice, including the fun "Go First" dice, four separate 12-sided dice which are guaranteed to give you a turn order for four people with one roll. Eric has a special page for custom dice, since he gets requests. I think your cuboid dice fall outside of his usual orders, but it may be worth contacting him.
• You could craft your own dice out of Super Sculpey (Firm). I and my daughters use this stuff all the time to make small, semi-durable objects. The firm version is a little tougher to work with (like heavy clay), but it holds its shape very well while curing. You could make a jig of some sort with the shape you want, cut off the approximate amount of Sculpey, force it into the jig, unmold, then bake. Possibly paint and lacquer as a finishing step. This is what I would do if I needed cuboid dice tomorrow (I keep this stuff on hand for crafting and object-making).

Making the dice: In thinking about the "jig" I'm not sure there is an easy way to do that. Perhaps a better method would be to use flat surfaces to form the dice. Start by cutting off the Sculpey from the block at the approximate size and shape you want. Weigh the thing with a decent scale that gives you fractions of grams so that you can more reliably make a second one of the approximate same size. Weight is going to be more helpful than anything else here. Then, use flat surfaces to gently square off the sides (it should already be the approximate cuboid shape when you cut it). Be gentle, so as not to deform it. Once you have the desired number of dice, bake.

• Great tips - Super Sculpey it is! May 2 '14 at 15:40
• Perhaps make dice of random shape with plane faces, not just six faces. Mark each face, and consider a throw as giving the face on the table, not the one up (for a pyramid or so there won't be a upper face). Probably a roundish form with many faces will work best May 3 '14 at 9:43
• @vonbrand - That's how tetrahedral 4-sided dice work, although they stamp the numbers at the base of the triangular faces. Fancy. May 3 '14 at 13:44

How about not using dice, but a top or roulette? Mark sectors with colors/numbers; by adjusting the sizes you get whatever probability distribution you want.

• That's a good idea - thanks for the suggestion. May 2 '14 at 19:06