Has anyone had luck projecting ROI on new domain purchases (reg or aftermarket)?  Buying based on someone's estimated traffic or because it sounds good seems like a good way to waste money.  I want to know what stats I can look at that directly correlate to parking revenue prior to purchasing a domain.  Otherwise it is just blind speculation.

That is the key to this whole game...I think (based on looking at top sales) that you need a mix of blind speculation along with stats to make good buying decisions.  Case in point - my top 3 earners do not contain any quantitative stats.  So bottom line, no, I don't think there is any clear cut formula (or at least none that I have figured out to date) that will tell you if a domain will make you money or not.

I generally agree, though I think the "blind speculation" actually starts out as blind, then starts to move closer and closer to 20-20 vision as I review the results of some experiments and tests. Right now I think I need a pair of Coke-bottle lenses to sort out my portfolio and "guess" what domains to acquire next. That's part lack of serious experience and time, and part because I don't have a decent portfolio management app.