In a note on description versus prediction Dijkstra [1] demonstrated the lack of predictive power of purely descriptive models. In this comment it is argued that the main difficulties of predictive models are not mathematical but substantive problems concerning explanation versus prediction rather than description versus prediction.