284 followers
@rdrighetto A long, long time ago, G. Telles and I started with the hypothesis that the data we were analyzing was too dirty. But ultimately found that the "usual" (classical? Standard?) parameters of the field were actually throwing away ~25% of all usabl