In order to have a portfolio sufficiently weighted towards the the right deal(s), it is necessary to say no to enough of the wrong deals. (And of course to have access to the right deals, which many of the most advantaged startups can conclude far more quickly and discreetly with VCs than by marketing their business plan to the public.)

Modern portfolio theory isn't adequate protection when the probability of the average retail investor picking a portfolio consisting entirely of losers is sufficiently high.