Eight hundred participants enter the market. Each holds a list: who they’d like to team up with. The goal: split into pairs so that no one later wants a do-over. In mathematics, this is known as the problem of stable matchings and its roots go back to the work of John von Neumann. He described a ‘club of chosen unions’ — a special set of pairs that satisfies everyone. But exactly how to assemble such a club long remained a mystery.

Now the answer is found: just clean up the preference lists. First, remove pairs that interfere with each other — akin to tossing from a suitcase items that don’t go together. Then delete pairs that prevent improving the situation for many at once. After this double cleanup, what remains is a trimmed list — a veritable collective mirror of desires. It turns out that within this trimmed list lies hidden the club of chosen ones.

The main takeaway: the club of chosen ones is exactly those pairs that do not lose to any other pair in the trimmed list. Any union outside the club necessarily loses to someone inside the club, and to see this, no more than three steps are needed. That is, instead of sifting through millions of possibilities, it’s enough to apply two simple cleanup procedures. This approach erases the boundary between different views of stability and yields a practical algorithm for optimization in a wide variety of areas: from network markets to dynamical systems, where agents themselves drift toward consensus. Eight hundred participants can rest easy — their club is predictable.