Y Combinator Demo Day took place earlier this week, and as usual, the companies generally look really impressive. It’s become this funny ritual in the tech scene twice a year at Demo Day, when we all watch the crazy scramble between YC companies and investors and wonder, “Gee, companies these days are consistently high-caliber enough that a lot of this all looks kinda random. How well could you do as an angel investor at Demo Day if you simply walked around and offered SAFE notes to the fourth, eighth, and each subsequent multiple of four-th founder you met?” Or one more advanced version I saw on Twitter (I apologize for not remembering who wrote this, because it was great): “Go to demo day, point a camera at the crowd instead of on stage, figure out which pitches are generating the most FOMO by watching the crowd, and then just offer term sheets to all those companies, sight unseen."
This got me wondering. In a market for term sheets like YC Demo Day, which isn’t “efficient” per se but certainly feels like it’s found some sort of maturity to it, how much of the ultimate success of any given term sheet offer is due to skill on the part of the angel investor, and how much is due to luck? Clearly, investing in startups is an exercise that requires skill. And the future inevitably won’t be kind to the majority of these startups; that’s just the math. But I feel like if you went to demo day and offered a term sheet to a completely random startup, or even to your pick for the least impressive startup there, there would be a non-zero chance that you’d end up having picked the best of the batch.
You might be inclined to assign this quasi-randomness of outcomes as a sign of YC or startups’ relative immaturity; but in fact it’s the opposite. In the public stock market, the most mature market we have for this kind of thing, stock picking is an activity that trends towards complete randomness unless you can prove the exception. Michael Mauboussin, who has a few fantastic books on the subject including The Success Equation: Untangling Skill and Luck in Business, Sports and Investing, has a great thought experiment question on Skill versus Luck that I think about all the time. It’s a really simple question:
Can you lose on purpose?
The question is almost deceptively simple: if you can lose a game on purpose, then there must be some skill involved; if you can’t, then it might be an exercise in pure luck. Yes, of course you can lose a footrace, a game of chess, or a business venture on purpose. No, you can’t lose the card game War on purpose. But what about the stock market? If I gave you a thousand dollars with a specific ask: Invest this into public companies with the goal of losing all of the money as fast as possible; can you do it?
Now, I’m sure many of your reactions were something like, “Well that’s easy. I’ll simply ____ (and then fill in the blank here with “Make 200 trades at a $5 commission” or “Mark my trades so that the IRS takes all my money” or “Buy some options or other derivative contracts that are guaranteed to not work out”) And yes, you’re correct in that all of those stock market-adjacent activities, like market making, diversification, tax efficiency, or creative finance, legitimately require skill. But just the pure act of picking? It’s harder to lose on purpose trading public stocks than you might like to admit. If you were genuinely skilled at picking stocks that underperform the market, then you could simply make the reverse trades and clean house. Most people are not able to do either; not on command, and especially not repeatably.
The “Can you lose” thought experiment is interesting for two reasons; first, because it’s an interesting perspective on luck versus skill; second, because if the answer is Yes, then you might get some non-obvious insight out of asking, “okay, well then if I can identify a skillful strategy for losing on purpose, then what is the opposite of that?”
Mauboussin’s thought experiment was restricted to public companies, though, where we could make two key assumptions: one, that the market is deep enough that all prices represent a genuine market, and one motivated trader’s win is another’s loss; two, that betting on a company doesn’t materially affect their chances of success in any direction. (Yes, the stock price going up or down will affect the company’s cost of capital, their executive tenure, and other things. But this is a thought experiment! Come on.) But with startups, and especially early stage ones like in YC, we can’t make either assumption. Raising money has a direct impact on a startup’s chances of success; and is there are no short sellers betting against startups. (Except maybe on Twitter.)
But in a strange way, the phenomenon still could hold true. Public market companies are all companies that meet a minimum set of standards, and whose price is set more or less reasonably by the investing public, who cannot know the future but does their best on average. YC companies, similarly, are all startups that meet a minimum set of standards, and whose price is set reasonably at “Who knows?!?”, and where the caps on SAFE notes if we’re really being honest are more a reflection on the price of early stage money than they are on the particular valuation merit of a startup.
Furthermore, the self-fulfilling nature of early-stage investing whereby offering a company a term sheet has some real impact on other investors also wanting to offer them a term sheet, increasing the odds that the round will work out, that the company will make it to Series A, and so on. Sure, offering a term sheet to a dud startup with dud founders probably won’t increase the odds of success if the company was a stinker anyway. But if basically all the founders in a YC cohort are impressive, and all of the startups are plausible, and all of them have followed the right steps to make sure their companies are a genuine shot on goal at scalable success, then it’s worth revisiting our question: can you lose on purpose? What would it take?
I think it’s an interesting thought experiment to imagine what losing on purpose at Demo Day would actually look like. You could only offer term sheets with punitive valuations and deal structures, but then you’d simply get ignored; you wouldn’t lose, you’d simply stay at baseline. You could offer uncapped SAFE notes, throwing caution to the wind and saying “Come at me, dilution”; that would certainly be a great way to ensure you’ll be wiped off a cap table in short order. But then maybe some founders take your offer, and opportunistically leverage your vote of confidence into extra momentum that helps make a difference on their path to success. (Or you could simply pivot into being a media company that monetizes impressively with venture investments made along the way.) Maybe it’ll take some creativity. You could act like a huge jerk to founders and get banned from Demo Day; but again that’s not losing; that’s just forcibly not playing.
At the end of the day, I can only think of one strategy for effective self-sabotage: by asking the cursed question: “Who else is investing?” Think about it. Suppose you went to demo day and adopted a strategy: you talk to every founder you can, and offer them a term sheet, but only if someone else has already offered them one. This should, in optimal conditions, be a legitimately “skillful” way to gain the worst possible outcome, and find yourself with the highest amount of the worst-quality investments. You’re minimizing your chances of making an offer to any genuinely unique companies; nor will your advances at the genuine rockstar companies be likely accepted. You’ll successfully select for the most average startups, whose founders (who have the most information and leverage in the room) are consciously selecting investors who are offering a strategy deliberately designed to not make high-conviction investments. That ought to work, I think.
So, if an optimal strategy to lose on purpose is to deliberately select for companies that already have other term sheets, it follows that an optimal strategy to win on purpose would be to do the opposite of that: either try to make the first offer a company ever sees, the highest conviction offer a company ever sees, or simply the only offer a company ever sees. The math checks out on all of those, I think. And all three of those activities require real skill. Being first requires skill; it means having the best deal flow, and that takes work and skill. Being highest-conviction (and communicating it successfully) requires skill; it means being a good salesman. So it may turn out that Demo Day, just like the stock market, is tending towards a mature enough market that all of the skill and all of the hard work that matter are actually in all of the activities you have to do peripheral to the market itself; in earning the right to pick effectively; not the picking itself.
One absolutely surprising read: as you likely heard, David Koch passed away this week. One thing you probably didn’t know about him: he averaged twenty points a game in college basketball. Another thing you probably didn’t know: he survived a commercial airplane crash. Koch was on a 1991 US Air flight from Columbus to Los Angeles in a 737 that collided on the LAX runway with another flight, killing 34 people. Koch survived, and wrote a detailed account of the event that’s remarkable:
Passenger’s account of escape from burning 737 highlights cabin safety issues | David H Koch, Cabin Crew Safety.
Thanks to M.G. Siegler on Twitter for the link.
Some other links this week:
Google is tightening its grip on your website with new AMP updates | Owen Williams
Less than half of Google Searches now result in a click | Rand Fishkin
Welcome to Cleveland, where the Browns are contenders | Ben Baskin, Sports Illustrated
Have a great week,
Alex