In a Fred Wilson wrote about seed investing: recent post “When you lose on 60–80% of your investments, you really need the ability to make 10–20x on your winners.” Does that make sense? Money multiple fund Assume that a seed invests out of a fund. investor And that he wants to make a 12–15% IRR per year on this fund. And that he exits a after 7–10 years. startup With 13.5% IRR per year and 8.5 years till exit, the seed investor needs to make (1 + 13.5%) ^ 8.5 = 2.9x on his fund. Loss Assume, like Fred, that the seed investor loses on 60–80% of his investments. With a 2.9x money multiple on his fund and a 70% loss, the seed investor needs to price each startup that he invests in at a 2.9 / (1 - 70%) = 9.7x expected money multiple. The seed investor needs to make 9.7x on each startup that he invests in. Not just on his winners. Dilution Assume that each startup raises 1–3 follow-up rounds. And that each round buys 15–25% of the company. With 2 follow-up rounds and 20% new shares per round, the seed investor will dilute 1 - (1 - 20%) ^ 2 = 36%. With a 9.7x expected money multiple and 36% dilution, the seed investor needs to price each startup that he invests in at a 9.7 / (1 - 36%) = 15.2x money multiple. So If you believe these 3 assumptions then Fred’s “when you lose on 60–80% of your investments, you really need the ability to make 10–20x on your winners” makes sense. “When you want to make 2.9x on your fund, lose on 70% of your investments and dilute 36%, you really need to price each startup that you invest in at 15.2x” makes even more sense. Joachim Blazer is the author of The #1 Guide to Startup Valuation .