Conversation

What would failure-rate adjusted valuations of startups look like? VC-firm-level returns are meaningful, but it’s weird that individual valuations are a vanity metric. Valuation: price at which you sell a % to a private group that is expecting median investment to be worthless.
3
10
I’m beginning to think it’s almost malpractice of some sort to use valuations for anything other than IRR computations to allow VC firms to have a basis for raising the next fund. A purely technical use case since big returns take the longest to yield and failures are quick.
1
7
If you used real returns, VC investing would be too slow and conservative, so it’s ok to use valuations for that *in a portfolio* but using isolated valuations to valorize pre-exit founders is shady af. Makes failures look like going from billion to zero. No it’s zero to zero.
1
9
Ah got it. Valuations are how startups do virtue signaling 😐 It’s basically BIRGing *average* returns in a power-law sector where *median* returns are more meaningful but far less flattering.
1
7
Replying to
Why is the median more meaningful? It's more likely, but the average also carries information. Using actual results means using one shot from a distribution that has tautologically perfect accuracy but very low precision. Estimates are high precision, but potentially inaccurate..
2
Replying to
Right! Much as with estimating how good a player in a tournament is. If you wanna know how good the player is, you'll do much better looking at their expected average winnings rather than actual winnings if the estimators are savvy. The expectation is based on way more data.
2
Replying to
You’re mapping “player” to “VC”, I’m mapping it to “startup” Nobody but VCs and LPs gives a shit about how well they do. Startupball as a grand narrative is about the startups and their hero’s journeys.
Replying to
The VCs are the people on the sidelines estimating the chops of the player. The player is the founder of the startup, or the startup more broadly. We can look at the results to estimate their merit, or we can look at the expert-derived expected average. Tradeoffs seem analogous.
1