I suggest that there's selection bias in stories like this. It's well know in finance that average pay off grows with variance, both because of the "risk premi

I suggest that there's selection bias in stories like this. It's well know in finance that average pay off grows with variance, both because of the "risk premium" and because your losses are capped at "going out of business". If you've got a dollar and engage in a $1 bet with 50/50 chance, your average pay off is $0. If you've got a dollar and engage in a $4 bet, your average payoff is $3, because in half the cases you win $4, and in half the cases you win $-1 and then go bankrupt. In a period when all bets are higher variance, and using a retrospective study that implicitly discards all the firms that go bankrupt, then we certainly conclude "high risk bets rule!".