Accurate predictions of extreme events do not necessarily indicate the scientific superiority of the forecaster or the forecast method from which they originate. The way forecast evaluation is conducted in the media can thus pose a dilemma, say researchers.