The similarity of seismograms recorded by two seismic stations demonstrate that the 31 October 2007 moment magnitude M5.4 Alum Rock earthquake is a repeat of a 1955 ML5.5 earthquake. Both occurred on Oppenheimer et al.'s (1990) Zone V “stuck patch” on the central Calaveras fault, providing new support for their model of Calaveras fault earthquake activity. We suggest that Zone V fails only in a family of recurring M ∼ 5.4–5.5 earthquakes. The 1955 and 2007 earthquakes are the penultimate and ultimate Zone V events. Earthquakes in 1891 and 1864 are possible earlier Zone V events. The next Zone V event is not expected in the next few decades, assuming a time‐dependent recurrence model: the mean forecast date is 2064 (2035–2104, 95% confidence range). We further suggest that Zones I, II, III, and IV fail in recurring M ∼ 5.1–5.3, M ∼ 5.6–5.8, M ∼ 6.1–6.3, and M ∼ 4.9–5.0 earthquakes, respectively. If our earthquake recurrence model is correct, the next Zone I event is overdue and could occur anytime, and M5–6 earthquakes should not occur on Zones II, III, and IV before 2014, 2012, and 2026, respectively. We cannot rule out the possibility that Zone VI, which lies at the southern end of the Mission Seismic Trend, where the southern Hayward and central Calaveras faults appear to connect at depth, fails aseismically or in large events on the southern Hayward fault, such as last occurred in 1868, or in large events on the adjoining northern Calaveras fault segment.