We present a new method to use directly observable surface‐slip measurements in seismic hazard estimates. We present measures of scaling‐relation fits to slip‐length data. These fits show sublinear scaling, a slowing in the rate of slip increase for the longest ruptures, so that L scaling—scaling with the length of the rupture—does not hold out to very large aspect ratio events. We find the best fitting for a constant stress‐drop model, followed next by a square root of length model. The constant stress‐drop model, newly introduced here, provides a geometrical explanation for a long‐standing puzzle of why slip only begins to saturate at large aspect ratios. The good fit of the constant stress‐drop model to the slip‐length data lends further support to the observations of constant stress‐drop scaling across the whole range of magnitudes of earthquakes, from small to great earthquakes. The good fit of the constant stress‐drop model is also reflected by the low variability about the mean, with an average of less than a factor‐of‐2 variability in stress drop about the mean observed. Converting magnitude‐area scaling into implied slip‐length scaling, we determine qualitative consistency in the functional forms, but a quantitative difference of, on average, ∼30% more slip estimated from magnitude area compared with slip length.
Online Material: Tables of magnitude‐length‐width, magnitude‐area, and surface slip‐length relations.