Abstract Motivated by the Holocene history of great Cascadia earthquakes proposed by Goldfinger et al. (2012), we developed a simple model by which we can examine this further. Combining their observation that the plate boundary ruptures in four characteristic segment types with an empirical area versus seismic moment relation and an assumption of constant stress drop, we obtain a model for the along‐strike slip distribution for each earthquake type. Summing the slips in each segment over the number of earthquake of all types produces estimates of Holocene plate convergence that quantitatively agree with independent estimates obtained from geodetic and geologic data. Because the width of the fully coupled subduction interface decreases to the south, the assumption of constant average stress drop in earthquakes requires that their slip also decreases to the south. To satisfy the plate convergence constraints, this requires shorter and increasingly numerous earthquakes to finger in toward the south, just as observed in the Goldfinger et al. (2012) paleoseismic record. We conclude that the Goldfinger et al. (2012) history is quantitatively consistent with plate tectonic constraints, and that the Cascadia subduction zone is fully coupled seismically. The seismic flux release rate averaged over the Holocene is estimated as 2.7×10 9 m 3 /yr.