He compares a valuation cap against potential exit value.
One thing I sometimes do in office hours is calculate the implied probability represented by the startup's valuation. E.g. if a startup would be worth 10b if things go as planned and their dday valuation cap is 25m, they're a good bet if they have a 1/400 chance of being right.
Many users praise the reframe of implied startup success odds from valuation caps as powerful insight into positive expected value, while some dismiss the probabilities as unreal or liken it to a rigged casino.
The point is to make founders more confident by reminding them that they're a better deal than they realize. They arrive somewhat shamefaced because they have maybe a 1/10 chance of succeeding, and leave realizing that this makes them a stupendously good bet.
To get an accurate estimate you have to multiply by 3 or 4 to account for dilution by the time the company is worth billions. But the implied probability is still usually ridiculously low. Yesterday I talked to a startup whose implied probability was 1/10,000.
@paulg What evidence do we have that anyone can reliably tell between P=1/200 (good bet) and P=1/1000 (bad bet) where P is "probability of 10b valuation in 10+ years"?
@paulg @threadreaderapp unroll
@AgustinLebron3 @paulg i feel like the ‘statistical rigor’ applied to valuations is all post hoc to justify vibes that are entirely subjective.
@AK2Pac @paulg @AK2Pac Hallo, here is your unroll: https://threadreaderapp.com/thread/2063644836981260638.html Enjoy :) 🤖
@jrysana @paulg @DanielleFong my heuristic would be rough OOM of the TAM x p(can the team win enough and sustain enough loss get over the valley of death to get initial traction+scale) (weighed higher early stage) + bonus if there is some unique non-obvious insight/advantages (product, team, differentiation)
Built a free calculator for different sectors to map implied val probabilities.
People way over/underestimate how much conviction they need to have a positive EV, if val is low/high and market small/big
*click calculator on link or scroll to bottom (select sector for comps)
https://www.shack.xyz/pitch/
@jamesrosst @paulg The null hypothesis you need to reject is that the track record is due to deal flow.
Getting the best/brightest/most hardworking to want to work with you is the hardest bit. It's a lot easier if you can start w/ that population.
@paulg Yes! This is the right way to think about all growth companies to be honest. It is also possible to do a blended DCF + probability tree too (I did something like this for SpaceX).
@paulg You’re doing it all wrong. A good bet is one that is not dependent on terminal value in the first place. I wrote a whole book about it, then fixed it by improving the instrument you pioneered.
@paulg Another common mistake is thinking the probabilities are declining per hurdle, when they‘re increasing in reality. Much higher probability to go from 100M -> 1B than to hit 100M in the first place (given venture startup)
@paulg Experienced investors implying you have a 1/400 chance of succeeding makes them feel better?
@menhguin @paulg @DanielleFong gut feeling
@paulg I think I'm a 1 in 3 probability solo honestly But that's bc I've gotten an ex NASA Planck satellite data analyst physics PhD to invest 700k while beating her at Mario kart bc my math about mega millions billion dollar prize wwyd was very enticing at 3 economic levels.
@AgustinLebron3 @paulg Prob yc’s track record
@menhguin @paulg @DanielleFong I mean yeah that's pretty solid but what if the TAM is unclear? Like early social.
@paulg You calculate market size. You cannot calculate the chance of if someone is right. Google did not monetise their market share for yeaaars and their unique ability (not idea) is to attract the best of the best people.
@aitization @paulg Or maybe investing is game theory + mathematics
@ajxbuild @paulg Sure the investor needs to earn a return but the market price should not be a 95% discount to your EV
He compares a valuation cap against potential exit value.
One thing I sometimes do in office hours is calculate the implied probability represented by the startup's valuation. E.g. if a startup would be worth 10b if things go as planned and their dday valuation cap is 25m, they're a good bet if they have a 1/400 chance of being right.
