The Yahtzee Problem

My wife and I play Yahtzee every now and then. At some point I started noticing I was rolling Yahtzees more than seemed reasonable — and almost always with fives. At first I chalked it up to confirmation bias. But the nagging feeling didn't go away.

So I started asking the more interesting question: how would you even measure that?

You can't just roll a die a few times and declare it biased. You need volume — hundreds of rolls, tracked faithfully, with something that can actually read the result. Manual counting was a non-starter. I wanted data, not a weekend of tally marks.

That question opened the door to a bigger one: could I build something to answer it? What would it take — mechanically, optically, computationally? And once I was building it, why limit it to one kind of dice? The Yahtzee set sparked the idea, but the machine I was imagining could work on any type of dice with any number of faces.

This project is the attempt to find out.

Paste your Autodesk embed src URL into the iframe in the HTML to show the 3D model here.

Paste your Autodesk embed src URL into the iframe in the HTML to show the 3D model here.