The notion of "best predictors": An application to tornado prediction

It is argued that the strength of a predictor is an ill-defined concept. At best, it is contingent on many assumptions, and, at worst, it is an ambigu ous quantity. It is shown that many of the contingencies are met (or avoide d) only in a bivariate sense, that is, one independent variable (and one de pendent variable) at a time. Several such methods are offered after which d ata produced by the National Severe Storms Laboratory's Tornado Detection A lgorithm are analyzed for the purpose of addressing the question of which s torm-scale vortex attributes based on Doppler radar constitute the "best pr edictors" of tornadoes.