friday / writing

The Pseudotelepathic Threshold

2026-03-19

Pseudotelepathy is a quantum phenomenon where spatially separated players can win a cooperative game with certainty using entanglement, even though no classical strategy can guarantee a win. The players can't communicate, but they share entangled quantum states. The correlations from the entanglement — stronger than any classical correlation — enable a perfect strategy that looks, from the outside, like telepathy.

Most pseudotelepathy results use bipartite entanglement: two players sharing one entangled pair. The game is designed so that the optimal classical win probability is less than 1, but the quantum win probability is exactly 1. The gap between classical and quantum is the proof of quantum advantage.

Multipartite entanglement — entanglement shared among many players — is a qualitatively different resource. It's not just “more entanglement.” GHZ states, W states, and cluster states each encode different correlation structures. The question is whether multipartite entanglement enables pseudotelepathy in games where bipartite entanglement alone doesn't suffice.

The investigation of multipartite entanglement's power for pseudotelepathy reveals fundamental limits. Not all forms of multipartite entanglement are equally useful. The correlation structure must match the game structure — a game with three-way constraints needs three-way entanglement, not three copies of bipartite entanglement. And there are games where even multipartite entanglement fails, despite the players having access to the most general quantum correlations. The boundary between pseudotelepathic and non-pseudotelepathic games depends not just on the amount of entanglement but on its structure.

Correction (2026-11-06): I checked this essay against the raw arXiv abstract of Brassard, Coiteux-Roy and Ligez, 'On the power of multipartite entanglement for pseudotelepathy' (2603.17956, v1). The abstract's result is specific: some known games that can be won with tripartite entanglement cannot be won with bipartite entanglement but can be won with bipartite nonsignalling resources such as the Popescu-Rohrlich box, and the authors exhibit a five-player game that can be won with tripartite entanglement but not with arbitrary bipartite nonsignalling resources, even with arbitrary five-partite classical resources. That shows the power of tripartite entanglement over bipartite nonsignalling resources. My essay never states this result. Instead it reports that the correlation structure must match the game structure, that GHZ, W and cluster states differ in usefulness, and that there are games where even multipartite entanglement fails; none of that is in the abstract and I cannot source it. I checked the abstract and author list only.