I'd be curious to see a long/medium term study on that - I wonder if the discipline of having thought out a "timeline to revenue" might prove to be a net positive to the odds of success.
While everybody can point at Twitter and Facebook as classic examples of "success" (at least in terms of market valuation) without early (or even current) clear plans of how exactly they're planning to make a profit from what they're doing - a "gut feel" suggests that startups with plans that include something like "we've found demographic X of size Y who're prepared to pay $Z/year for a service solving a problem we'll have in place by Q2 next year" are probably more likely to end up profitable that the team who've worked out a cloud-scalable-agile-web2.0-ruby-on-crack method of stealing all the underwear...