arXiv:2610.06799
Bell nonlocality does not imply violation of nonlocal games
Ricardo Faleiro, Lucas E. A. Porto, Marco Túlio Quintino, Flavien Hirsch
Abstract
Nonlocal games offer a conceptually clear way to reason about Bell nonlocality. However, we show that as originally defined, and most commonly used, i.e. with a deterministic predicate, they cannot be complete witnesses of nonlocality. Concretely, we prove the existence of nonlocal behaviours which do not violate any such nonlocal game. Furthermore, we also show that quantum mechanics can produce these game-classical nonlocal behaviours. We prove that this phenomenon exists in every bipartite scenario $(X,Y,A,B)$ with $X,Y\geq2$ and $A,B\geq4$.
AI Summary
The paper asks whether every Bell-nonlocal behaviour violates a nonlocal game with a deterministic win/lose predicate, when the input and output alphabets are fixed. Allowing probabilistic predicates is already known to recover arbitrary Bell inequalities; the question concerns the more restrictive deterministic convention. It defines the game-classical set $\mathcal G$ of no-signalling behaviours satisfying every deterministic-predicate game and proves that $\mathcal G$ is a polytope whose facets admit game representations. It then establishes $\mathcal L\subsetneq\mathcal G$ and $\mathcal Q\cap(\mathcal G\setminus\mathcal L)\neq\varnothing$ in scenario $(4,4,2,2)$ and every bipartite scenario with $X,Y\geq2$ and $A,B\geq4$. The proof identifies local-polytope facets that cannot represent games, even after scaling, constant shifts and no-signalling transformations. Mixed-integer linear programming finds game representations for 173 of 174 nontrivial facet classes in $(4,4,2,2)$; the remaining class is analytically game-inequivalent. Two of 33 known facet classes in $(2,2,4,4)$ have single-block obstructions that persist under standard input/output lifting. Explicit quantum violations, combined with a geometric mixing argument near these facets, establish quantum game-classical nonlocality. A numerical example brackets the game-violation threshold between mixture weights $0.041$ and $0.042$, subject to solver precision. The separation is scenario-dependent: the exceptional binary-output facet becomes game-equivalent in $(5,5,2,2)$. The $(2,2,4,4)$ facet catalogue is not known to be complete. The relationship between $\mathcal G$ and $\mathcal Q$, and whether local operations with shared randomness can activate game violations for every nonlocal behaviour, remain unresolved.
OpenAI's ChatGPT-6 Astra and Anthropic's Claude Opus 5.5 were used for text proofreading and producing the tikz code from a hand drawn sketch.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 17 pages, 1 figure