If an earthquake forecast is to be considered successful, it should perform better than a reasonable "least astonishing" null hypothesis. We view the Cellular Seismology (CS) method as a least astonishing hypothesis model for forecasting locations of future earthquakes, and we propose that CS is a useful standard of comparison for other, more complex, forecast models. Recent forecast models based on analyses of earthquakes in California such as that of Ebel et al. (2007) and the Pattern Informatics method of Rundle et al. (2002, 2007) provide opportunities for comparing methods that incorporate information about rates of seismicity with a method (i.e., CS) that only assumes that future earthquakes will occur near epicenters of past earthquakes. The CS method is considered here as a least astonishing hypothesis model because it considers a location to be a potential source point of future earthquakes even if only one past earthquake occurred near that point. By contrast, the Ebel et. al. (2007) 5-year forecast model (E07) maps the spatial distribution of rates of seismicity, and the Pattern Informatics model (PI) not only considers rates of seismicity, but also incorporates changes in rates of seismicity as a measure of the potential for future earthquakes to occur at some location. Our comparison of success rates of the E07 method and the PI method with CS for earthquakes in California has yet to reveal any strong evidence that inclusion of rates or changes in rates in the forecast model improves the success rate of the forecast.