[RobertoG]

" In science, prediction comes before explanation"

I'm going to be a little picky about your choice of language here.

It seems to me that, at least that your are predicting randomly, you can't possible make predictions with a good model, that is, a good explanation.