Base Rates in Venture Capital

“Executives and investors commonly rely on their own experience and information in making forecasts (the “inside view”) and don’t place sufficient weight on the rates of past occurrences (the “outside view”).”

-Base Rates by Michael Mauboussin (PDF)

Michael Mauboussin and Annie Duke, two of the best minds on the science of decision making, frequently talk about the importance of “base rates”. Think of base rates as unbiased probabilities, or sources of truth: they are the probabilities that have historically proven to hold true.

Examples:

• 10% of people are left handed
• 50% of marriages end in divorce

As humans we are terrible at forecasting the future. We tend to be overconfident and overly optimistic with our predictions. Take the statistic that 50% of marriages will end in divorce: nobody enters a marriage thinking this static applies to them. If we acted rationally and created a strategy adhering to base rates, most people would start marriages with a prenup, but they don’t.

This same overconfidence and bias frequently plays out in venture and alternative investing:

“My friend was an early investor in Figma. I’m just as smart and connected as she is, therefore, I should also should start angel investing so I don’t miss out on the next unicorn”

“Everyone is making money trading NFTs. But unlike most of these idiots I actually know something about finance and art so I’m likely to crush it”

“Wow, Sequoia is investing in this deal on AngelList! Since they are a top 1% VC fund, this deal will likely be a winner, I’ll co-invest as well”

Sally previously invested in Plaid and now she’s suggesting I co-invest in a new fintech deal. She obviously knows how to pick ‘em so I’m investing double my typical check size!

We often ignore even well known probabilities and instead forecast with overly optimistic internal narratives, what behavioral economists refer to as the the “inside view”.

The best way to combat the trap of narrative bias is to hold yourself intellectually honest using base rates (a.k.a., the outside view). When I coach new angels, an exercise I take them through is starting from the assumption that the base rate applies to them (a.k.a., the assertion that most people are MUCH better off investing in stocks). I then put the onus on the individual to convince themselves they can outperform the base rate.

The technique of first looking at base rates (and then building a case for investment from there) applies to many areas of investing and strategy decisions.

Without a consideration of base rates, most investors are also prone to base rate fallacies.

As an example: the media loves to focus on stories of young technology founders. You’ve probably heard: “being a startup founder is a young man’s game”. This created a false narrative - or base rate fallacy - that the majority of successful tech founders are young. We might assume a base rate like 75% of first time VC-backed founders are under 25. In fact, the median age of a first time VC-backed startup founder is 31+ and the average tech founder is 40!

To my knowledge, no good collection of venture base rates published, until now.

Drumroll…*

Startups

• The odds a seed-stage, venture-backed company becomes a unicorn land in a ~0.5-2.5% range depending on the dataset: Ilya Strebulaev (Stanford) 0.5%, CB Insights 1.28%, AngelList 2.5%. (Note: from 2018-2021 the AngelList probability roughly doubled.)
• A unicorn label is not the same as a healthy business. Bain found fewer than 0.7% of unicorns (about 15 of ~2,500) clear $1B in both revenue and cash flow, the real bar for a self-sustaining company at scale. [link]
• Liquidity has dried up: the IPO share of unicorn exits fell from 83% in 2010 to just 11% in 2024, with acquisitions and secondary sales now the main exit routes. Recent unicorn cohorts (2020-2022) are exiting at roughly a third of the historical pace, so the unicorn label increasingly means “still private,” not “cashed out.” [link]
• Startup outcomes follow a power law: a small share of investments generates almost all the returns. Y Combinator’s own portfolio shows it starkly: roughly 6% of YC startups (the unicorns) drove about 90% of all valuation growth, and Airbnb and Stripe alone account for half the decacorn growth. Across venture broadly, about 6% of investments generate roughly 60% of returns. [link]

Founders

• Historically, founder age skews older. The median first-time VC-backed founder is 31+, the average founder of a top 0.1%-by-growth startup is ~45, and a 50-year-old is roughly twice as likely as a 30-year-old to build a top-tier company. Success rates climb with age well into the 40s and 50s. (MIT/US Census study of company formation, 2007-2014.) [link]
• AI may be flipping that at the frontier. In AI specifically, the average unicorn-founder age fell from ~40 in 2020 to ~29 in 2024, while the average across all unicorns barely moved (~30 to ~33). [link] Top VC Keith Rabois has also made this case (anecdotally) across his portfolios at Khosla Ventures and Founders Fund: AI founders are often 19 to 21 and more technically skilled than the past decade’s, because they are “AI-native.” [link]
• Domain expertise is not the edge people assume. In Ali Tamaseb’s study of 200+ billion-dollar startups, only ~30% of consumer founders and ~40% of enterprise founders had prior experience in their industry, and the rate was the same for non-unicorns, so industry experience did not separate the winners (life sciences is the exception, ~75%). The outsider is closer to the base rate than the insider. [link]
• Repeat founders carry a measurable edge. Founders who already had a success succeed again ~30% of the time, versus ~18% for first-timers; even previously-failed founders (~20%) beat first-timers. Prior success is one of the few durable predictors. [link]

Angel Investing

(Note: historical data sets for angels are small and sparse)

• The larger the portfolio, the more likely you are to make money [link]
  • 50 company angel portfolio had a median IRR of about 10%
• Even with 50 companies, 11% of investors lost money [link]
• Individual SPVs are usually a losing bet on their own: pick a random single-deal SPV on AngelList and there is roughly a 90% chance it either loses money or merely returns your capital after fees and time. The median SPV hovers around 1x after ~3 years, and essentially all the positive return comes from the top ~10%. [link]
• Note: Angels can crush it. Investors in Uber turned $5,000 into nearly $25M. [link]
• Large returns (over 4x) generally take 10 years before liquidity [link]
• Most operator-angels expect an IRR of 30%+, but the data runs lower [link]
  • Active, diversified angels have historically averaged closer to ~24-28%
  • Top quartile angels reach ~35-40%
• More due diligence leads to better outcomes [Link]

VC Funds

(Note: more historical data is available for venture funds than angels)

• VC performance is persistent: a manager whose prior fund was top-quartile repeats top-quartile 44.7% of the time and lands above median 68.9% of the time, while a prior bottom-quartile manager reaches top quartile only 9.3% (versus the 25% you would expect by chance) [link]
• Only 25% of funds will generate an annualized return greater than 25% [link]
• Fewer than 4% of venture investments return 10x or more, and about 37% lose money (return less than the capital invested) [link]
• The average venture fund underperforms public markets: in the Kauffman Foundation’s study of ~100 funds, 62 failed to beat a public-market equivalent after fees, and only 20 beat it by more than 3% a year, all while locking up capital for years [link]
• On what LPs actually get: top-decile funds now cluster around ~3x DPI, and a 3.0x gross fund compresses to roughly 2.1x net after fees and carry. True top-quartile DPI above 3.0x net is rare, and distributions have been slow post-2021. [link]
• What LPs underwrite in a lasting (“franchise”) fund, per Jessica Archibald of Top Tier Capital Partners: 85% of 3x+ funds had at least one company that could return the entire fund; returning 1x within 9 years is baseline (5-6 years is exceptional); and the strongest firms raise fund after fund without friction. [link]

Follow-on / graduation rates a professional emerging manager should expect:

Stage	Benchmark target	Recent reality (2022+ cohorts)
Seed to Series A	>35%	~15-20%
Series A to Series B	>50%	under pressure
Series B to Series C	>50%	under pressure
Series C to Series D+	>60%	under pressure

Note: Graduation rates fell sharply after 2021. Carta data shows 30.6% of the Q1 2018 seed cohort reached Series A within two years, versus just 15.4% of the Q1 2022 cohort; the informal Series A bar rose to ~$3M ARR, which only ~20% of seed startups clear. [link]

I will continue to update as I find new base rates.