The link that Don posted is basically accurate except for details (some of that came from my papers). The main detail is the recurrence interval of 2000 years. Pinter et al did infer that the last earthquake on it was 5000 years ago (it could not have been more recent from their work, but it could have been older). If it is slipping at 0.75 mm/yr, in 2000 years you would build up 1.5 meters of slip. The Patterson thesis (1979), confirmed by our own observations, suggest an event or closely-spaced in time events that slipped 3 or 4 meters. So, the recurrence interval might be 4 or 5 ka, but with bigger earthquakes. The 0.75 mm/yr is based on reasonable logic (Pinter et al GSA Bull 1998), but is based also on certain assumptions. Given how clear the stream left offsets are, it was surprising that the rate was this slow. My work is hinting at some of the left-lateral motion being offshore south of the Island. And, yes, Patterson said that the fault wasa capable of M7, but that was not including the part of the fault north of Santa Rosa and San Miguel Island, or a cascade of multiple ruptures going to the Malibu Coast Santa Monica Dume-Hollywood-Raymond. Although perhaps unlikely, this worst case would rupture right down Santa Monica Blvd (e.g., Dolan et al. 2000 GSA Bull) to Pasadena. The whole system is 200 km + long, and breaking the whole thing would be a larger earthquake than Landers.
Chris