Fig 1. Framework for creating sampled datasets.

Each row represents a single sampled dataset. The left column contains all of the observed activity trajectories in the motivating data, with the red segments representing the selection of days J which were selected for a given sample. The second column shows the estimated reliability ${\widehat{R}}_1$ based on the previous selection of days. The final column shows the subsequent prophecy estimate of the number of replicates $\widehat{J}$ required to achieve a reliability of 0.8. The histograms at the bottom of the second and third columns represent the distributions of ${\widehat{R}}_1$ and $\widehat{J}$ across all 200 samples generated in this way.