Extrapolating from two samples to "once-in-a-century" does not strike me as rational.