Why it is a question
THEORY-265 holds Tissot et al.’s result (LIT-989, read in NOTE-773) in the setting they ran: one game of thirty clean, non-overlapping, consistent modules of ten, couplings 1 within and 0.01 across, one parameter setting. The authors expect natural interactions to be graded and frustrated, and they do not claim the mechanism survives that. The phase rule is a gradient on cycle-averaged payoff, “loosely analogous” to Kuramoto entrainment, and no run couples units through their outputs.
The sixteenth status review’s re-run (THEORY-265, note of 2026-10-29) covers part of the scope. With within-module couplings graded, the result held at across-coupling 0.01. At across-coupling 0.1 or more, random phases already reach the optimum, so there is no lock-in for timing to undo. Overlapping modules, frustrated constraints and output coupling were not run.
What would answer it
THEORY-265’s former promote_when, unchanged:
A replication outside the authors’ clean game: modules that overlap, couplings of graded strength, or frustrated constraints, together with a sweep of the coupling ratio and the phase rate, showing that payoff-driven synchrony still forms along the strong couplings and still turns those groups into units that switch as one. A second route would settle more: the same result with entrainment through the units’ own outputs (a Kuramoto or HKB coupling) instead of a payoff gradient on phase. It fails if the effect disappears once modules overlap or the ratio of within to across coupling falls toward one. What cannot settle it: more runs of the same 30-by-10 game, or biological synchrony cited by analogy.