In statistics, Pyrrho's lemma is the result that if one adds just one extra, but specially formulated, variable as a regressor to a linear regression model, one can get any desired outcome in terms of the sizes and signs of one's coefficients, the overall goodness of fit and the in-sample forecast performance. This argument was advanced by Herman Wold and Lars Juréen but named and explained more fully by Theo Dijkstra. Dijkstra named it after the sceptic philosopher Pyrrho and concludes his article by noting that this lemma provides "some ground for a wide-spread scepticism concerning products of extensive datamining".
- Wold, Herman and L. Juréen (1953) Demand Analysis: A Study in Econometrics, John Wiley & Sons (2nd Ed)
- Dijkstra, Theo K. (1995) "Pyrrho's lemma, or have it your way", Metrika, Volume 42, Number 1, 119–125 doi:10.1007/BF01894292
- (Dijkstra, p. 122)
- Hendry, David F. (1995) Dynamic Econometrics, Oxford University Press