Two points. One, my understanding of Paul's basic premise is that it's not the class quality that matters, it's the unicorns that matter. As in, in this game, you're not chasing good classes. You want the one or two superstars, even if they're in a crap class. Secondly, the analysis I'd find useful is one that looks at unicorns using hindsight. How early can you tell them apart? They are the ones we came here for, remember? It's a very hard task, like telling the gender of day-old chicks.