An epidemic with a longer incubation period gives individuals more time between infection and symptoms. More time should mean more opportunity to prepare — to reduce contacts, to warn others, to change behavior. The intuition is that a longer warning produces a stronger response.
Wang (2026) models epidemic dynamics on heterogeneous networks using a mean-field game framework where individuals rationally weigh infection risk against the cost of social distancing. The counterintuitive result: longer incubation periods produce weaker behavioral responses and larger outbreaks. Exposed individuals — infected but not yet symptomatic or contagious — maintain full contact rates because they face no immediate cost from distancing. The incubation period is not a grace period for preparation. It is a delay during which the signal to act has not yet arrived, because the signal is symptoms, not exposure.
The mechanism is temporal discounting applied to behavior rather than money. A disease that takes two weeks to show symptoms gives each infected person two weeks during which rational self-interest says “keep going — you feel fine.” By the time enough people are symptomatic to trigger population-level behavioral change, the epidemic has established a larger base than a disease with shorter incubation would have allowed.
The general principle: a warning's effectiveness depends not on how much advance time it provides but on whether the advance time is accompanied by a reason to act. Time without signal is not warning — it is delay with the appearance of warning. The distinction matters because systems designed to “detect early” can produce worse outcomes if early detection reaches actors who have no incentive to respond until later signals arrive.