A better Way to Apportion Startup Equity?

I'd like to buy into Seth's proposed approach to apportioning equity in a startup.  I could even see building some sort of sliding mechanism based on realization of specific goals. 

But the idea of putting 90% of a firm's equity in escrow and assigning it to the appropriate partner over time based on a set of mutually agreeable metrics seems ridiculous to me.  Are there real-world examples of this actually working?

Seth says that assigning equity percentages based on a "moment in time" will yield the "wrong" result.  I see a couple problems with this logic.  First, even if you wait a year, or two years, any assignment, at the time it's made, is still based on a moment in time.  Sure, there's more history to work of off, but the future is still unknown. 

Secondly, and more importantly, this discussion leaves out the notion of risk entirely.  A startup is fundamentally like other investments, involving a risk upon entry and a reward upon exit.  Is there any reason why a startup should be treated differently than other investments when determining the "reward" side of the equation?  Who is burdening the risk up front, and what share?  Without addressing this question, one would not likely see necessary investors stepping up to the plate in the first place.

This discussion hits home for me, and I'd love to know how successful startups are doing this.