Hi. An old statistics joke goes something like this: “One out of every 20 people is crazy. So, think of your 19 closest friends. Are any of them crazy? If not, then it must be you!” This joke exemplifies the same type of error made in the sentence “In fact, if the model is reliable 1 of the 10 times must result in no El Niño (we don’t know when it will occur).” The model could be reliable even if, in all 10 of the first 10 times, El Nino did occur. Or even if it didn’t occur even once. Statisticians frame the difference between theory & observation as one of “populations” vs “samples” or “hypotheses” vs “data.” True, any observed proportion of El Ninos occurring other than 90% does support best an alternative hypothesis, rather than the hypothesis that the theoretical “population” proportion is 90%. It does not imply, though, that the model is unreliable. The most we can do is collect large numbers of observations, calculate the observed proportion of El Ninos, & then measure the strength of evidence for or against the model’s reliability.