I think you need to look at this from Fred Wilson's perspective as a VC. Exit is the only the measure of success. It is how he and every VC is ultimately incentivized by their LPs. The more exits and the larger those exits, the more money the VC makes and the more likely they'll have LPs invest in their next fund.
Of course, one can argue that startups should be not be in it for the exit, but when you take VC money, that is what you are signing up for.
As a VC the startups that don't exit are crucial to the analysis. For their purposes they should be included with an exit value of zero.
Say the startups with the highest (or lowest) funding all die before exiting -- that's a really important thing to know. That'd affect the regression outcomes (not to mention the decision-making process of everyone involved.)
Of course, one can argue that startups should be not be in it for the exit, but when you take VC money, that is what you are signing up for.
Fred talks about he defines success here - http://www.avc.com/a_vc/2010/06/how-we-measure-success.html
"We are financial investors and we do want to see our portfolio companies become valuable."