I help run a YC alumni-led fund, and one of the questions that always comes up for our group is: What's the optimal number of YC companies to back per batch? We built an app that uses two models for approaching this problem - we call it the Unicorn Ratio. Live version: https://orangecollective.vc/unicorn-ratio Repo: https://github.com/orangecollective/unicorn-ratio The two models it uses are: Kelly Criterion—A mathematical model that optimizes portfolio construction. https://www.investopedia.com/articles/trading/04/091504.asp Binomial Distribution—A probability model that predicts your chances of backing a unicorn based on portfolio size and historical success rates https://www.investopedia.com/terms/b/binomialdistribution.as... We actually learned of the Kelly model last week during a call that YC organized for a group of demo day investors (thanks YC). We built and open-sourced this in the same spirit of collaboration. Related, on the My First Million podcast, Garry revealed some data for investors at YC Demo Day—specifically the returns for investors who consistently backed 3+ startups each batch over a two-year span. Here’s the 1 min clip. https://youtube.com/clip/UgkxVnlzcslmeMCTDFc9G2R--W7pigxVe8N... So, what’s the takeaway here? Conventional VC wisdom is all about making a few, highly concentrated bets. But this suggests a broader, more systematic approach to YC investing could also be a viable strategy (and likely put you amongst the top returning funds). LMK if there are other models or approaches worth considering too. Disclaimer: don’t use these models as investment advice. |