Ensembles vs. information theory: supporting science under uncertainty