Fig. 1

Relationship between foraging success and behavioral unpredictability in patch residence time. Each boxplot summarizes simulation results from 1000 foragers. A forager’s patch search times (i.e., durations needed to find a patch) were generated from a gamma distribution with mean \(\lambda ^{-1}\) and variance \(\sigma _{s}^2\), and patch residence times were generated from a gamma distribution with mean \(t^*=\sqrt{\lambda b}/\lambda\) and standard deviation \(\sigma _{t}\) that represents behavioral unpredictability. When foraging success is evaluated in a short duration (300 time units, right figure), behavioral unpredictability increases variability in behavioral outcome while having a minor effect on the expected behavioral outcome. This pattern disappears when foraging success is evaluated in a long duration (100,000 time units, left figure). \(\lambda =0.01, a=100, b=25, \sigma _{s}^2=5\)