“There is no value investing in startups.” – Paul Graham

In startup investing, most returns are from a tiny fraction of companies. You might invest in a hundred companies, have 99 fail, but have one return 1,000x.

One might, therefore, conclude valuations don’t matter. Suppose you’re an average VC fund. You start with $1B, and you invest $10M in 100 companies. 99 fail, and one returns 100x after five years, so your net return is 0%. (Median VCs actually lose money, but we’ll be generous.)

Now, it doesn’t matter what valuation you gave that one big hit. If the valuation was double, you would only have invested $20M, and your return is now -0.1% – a negligible difference.

But the key is, there’s no way to tell in advance what that big hit is. If you could tell, you could only invest in big hits, make 150% returns, and be rich beyond your dreams. To invest at a higher valuation for that one company, you’d have to do it for every company. And your returns are now -12% – terrible! Alternatively, if you invest at half the valuation, your returns are 15%, which is superb.

Like any other investment, startup investing isn’t about identifying successful companies, but identifying differences between what median investors think and how successful a company really is. Any nincompoop can see Apple is successful, but there’s little return there – its stock is correspondingly expensive.