A paper in JGR on measuring El Ninos says that the 1997-1998 El Nino was a one-in-70 year event (± 20 years).
An El Nino just 25% larger is predicted to occur every 700 ± 200 years. (Read the paper, by noted climate skeptic David Douglass, to find out exactly what "larger" means -- basically, it's sea surface temperature.) Douglass finds that ENSO magnitudes follow a Gaussian ("bell curve") distribution. He also finds no increase in the frequency of ENSOs, going back to 1865.