Current earthquake prediction methods were so far mostly based on recorded historical earthquakes. However, our knowledge of long-term earthquake behaviour is strongly restricted by the rarity of large earthquakes and the short historical and instrumental records of earthquakes, and hence our understanding could be misled by incomplete and unreliable data as well as uncertain analysis. These facts can certainly affect the progress of scientific research on seismic hazard assessment and earthquake forecast. Therefore, for a seismogenic fault, it is important to determine a complete and reliable long-term history of earthquake ruptures on all segments of the fault if possible. Most studies on this topic attempted to determine rupture histories from distributions of damage, surface ruptures, and aftershocks of historical and modern earthquakes.