A theoretical study of collective escape proposes that animals can improve both speed and accuracy by treating an undisturbed neighbour as negative evidence about a possible threat.
In the model, flight by another animal raises perceived danger while continued stillness discounts it. This prevents false alarms from cascading uncontrollably as groups grow. The model also fits observations of wild sulphur molly shoals responding to bird attacks, where inferred discounting is strong enough to keep false-alarm rates roughly stable across group sizes.
