Math Shows Why Perfectly Fair Sharing Could Be Impossible

Imagine splitting a box of chocolates with friends. You want everyone to get a fair share, but people have different tastes. Now imagine you must hand out chocolates one at a time, and later gifts are less valuable than earlier ones. This is called residual maximin share (what you can still guarantee after the best items are gone). Mathematicians wondered: how close to "perfect fairness" can you actually promise? This paper proves a surprising answer: you can only guarantee two-thirds of what would be truly fair.

That number isn't a rough estimate. The author proved it's the exact limit—no fair division method can do better, ever, in any scenario where items have different values to different people. To show why, she built the smallest possible "worst case" examples. Think of it like discovering the shortest possible unfair argument: once you see it, you understand why everyday fair sharing sometimes falls short.

The paper also explains the structure behind this limit. It shows that the worst situations are surprisingly simple, not complicated lists of thousands of items. And it gives a clear test to check whether any proposed sharing rule meets the fairness bar. While this is pure math, it has practical meaning: any algorithm that divides goods—from inheritance, to server space, to donated organs—cannot promise more than 67% fairness in dynamic settings.

So what does this mean for you? If you're in charge of sharing something valuable over time, don't chase perfect fairness. Aim for about two-thirds, and know that the remaining gap is not your failure—it's math.