New Proof: Taking the Middle Answer Beats Clever Voting Every Time

Imagine five friends each pin their ideal restaurant on a map. You need to pick one spot. You could average all the pins, or you could take the one in the middle. A new paper by researcher Jianhao Jia asks a sharper question: if each friend might exaggerate or lie to pull the result their way, which rule is safest? The answer is the median — the middle pin, measured left-to-right and top-to-bottom.

The key word is 'strategyproof.' Plain English: nobody gets a better outcome by lying about what they want. Averaging fails this test badly. If you want the group to choose a spot near you, just claim you want it twice as far away, and the average slides toward you. The median doesn't budge that way. The paper proves the median also minimizes the worst-case total distance everyone has to travel, meaning it's both cheat-resistant and efficient on its worst day — a combination that's surprisingly rare in this kind of math.

The result is broad. It works for any odd number of people (3, 5, 7, and up) and any number of dimensions — think of dimensions as the number of separate things you're deciding at once, like budget, timeline, and location. With a small tweak, it also covers even-sized groups. The proof works by showing that any hypothetical rule that beat the median would have to collapse into a rigid, predictable form — and then showing that even that form can't win.

What does this mean for you? No app is changing today. But these rules quietly sit inside recommendation systems, delivery routing, shared budgets, crowdsourced maps, and AI tools that blend many people's preferences into one answer. The takeaway is reassurance: the simple, transparent, slightly boring approach isn't a lazy compromise. It's provably the best option available.