I thought predicting winners in startups was searching for "Black Swan" events - the ones models often miss...