A long‐standing debate in active tectonics addresses how slip is accumulated through space and time along a given fault or fault section. This debate is in part still ongoing because of the lack of sufficiently long instrumental data that may constrain the recurrence characteristics of surface‐rupturing earthquakes along individual faults. Geomorphic and stratigraphic records are used instead to constrain this behavior. Although geomorphic data frequently indicate slip accumulation via quasicharacteristic same‐size offset increments, stratigraphic data indicate that earthquake timing observes a quasirandom distribution. Assuming that both observations are valid within their respective frameworks, I want to address here which recurrence model is able to reproduce this seemingly contradictory behavior. I further want to address how aleatory offset variability and epistemic measurement uncertainty affect our ability to resolve single‐earthquake surface slip and along‐fault slip‐accumulation patterns.

I use a statistical model that samples probability density functions (PDFs) for geomorphic marker formation (storm events), marker displacement (surface‐rupturing earthquakes), and offset measurement, generating tectono‐geomorphic catalogs to investigate which PDF combination consistently reproduces the above‐mentioned field observations.

Doing so, I find that neither a purely characteristic earthquake (CE) nor a Gutenberg–Richter (GR) earthquake recurrence model is able to consistently reproduce those field observations. A combination of both however, with moderate‐size earthquakes following the GR model and large earthquakes following the CE model, is able to reproduce quasirandom earthquake recurrence times while simultaneously generating quasicharacteristic geomorphic offset increments. Along‐fault slip accumulation is dominated by, but not exclusively linked to, the occurrence of similar‐size large earthquakes. Further, the resolution potential of tectono‐geomorphic records is dominantly affected by measurement uncertainty, that is, by the ability to correctly infer a geomorphic marker’s pre‐earthquake(s) morphology. Typically reported values for measurement uncertainty suggest that it is improbable to resolve more than five large earthquakes from tectono‐geomorphic records.