Free tool
How many startups should an angel invest in?
How many companies, and how much capital, give you a good chance of at least one large winner.
Result
Chance of at least one 20x winner among 25 companies
72.3%
Companies needed for 80% confidence
32
Capital each year
$312,500.00
Total capital needed, including reserves
$937,500.00
A planning figure, not a forecast. It treats outcomes as independent and the chance per company as an assumption.
Illustrative only. This is not investment, tax or legal advice. Check the result against your legal documents.
Chance of at least one winner
By number of companies, at a 5.0% chance per company.The dashed vertical line marks your own portfolio size; the horizontal line marks your target confidence.
How this is calculated
The question. If each company in your portfolio has the same chance p of becoming a large winner, how many companies give you a good chance of at least one?
chance of at least one winner = 1 − (1 − p)ⁿ
where n is the number of companies. (1 − p)ⁿ is the chance that every one of them misses.
companies needed = ln(1 − target confidence) ÷ ln(1 − p), rounded up
capital needed = companies × first cheque × (1 + reserves)
capital each year = capital needed ÷ years to deploy
The default of 5% is illustrative. It is an assumption for planning, not a measured rate for any investor or market. Research on angel returns finds that most angel investments return nothing and a few outliers produce most of the returns, but it does not give a reliable rate for a 20x-or-better outcome. Set p to your own view.
The winner threshold is a label. Choosing 10x, 20x or 50x names what you mean by a large winner. It does not change the calculation; your choice of p does that.
What this leaves out. Outcomes are treated as independent, so one company's result says nothing about another's. In practice companies in the same sector, stage or year tend to rise and fall together, which makes a winner less likely than this calculation suggests. This is a planning tool, not a forecast.
Guide
What the tool answers: how many companies, and how much capital, you need for a reasonable chance of one large winner.
Why size is the first decision
Angel returns are not distributed evenly. Most investments return nothing, and a small number of exceptional outcomes produce most of the gains. The published evidence on this is consistent and uncomfortable: in one widely cited sample, more than half of angel investments returned less than the capital invested.
That shape has a direct consequence. If your returns depend on catching a rare outcome, the number of attempts matters as much as the quality of each one. A portfolio of five good companies is not a smaller version of a portfolio of thirty good companies. It is a different bet, with a meaningfully worse chance of containing the outcome that pays for the rest.
The arithmetic
If every company has the same chance p of being a large winner, then the chance of at least one is:
1 − (1 − p)ⁿ
where n is the number of companies. The term (1 − p)ⁿ is the chance that every single one misses.
With p at 5%:
| Companies | Chance of at least one winner |
|---|---|
| 10 | 40.1% |
| 25 | 72.3% |
| 50 | 92.3% |
To reach 80% confidence you need 32 companies. For 90%, you need 45.
What it costs
Twenty-five companies at a $25,000 first cheque, with 50% held back for follow-ons, is $937,500. Deployed over three years, that is $312,500 a year.
This is the number most people have not done. The portfolio size that gives a reasonable chance of a winner implies a capital commitment that is often several times what someone expected to spend, and it has to be money that can sit illiquid for the best part of a decade.
If that total is more than you want to commit, the honest responses are to invest smaller cheques across more companies, to accept a lower confidence, or to invest through a fund or syndicate where someone else carries the breadth.
What the 5% means
It is an assumption, and the tool labels it illustrative.
The research establishes the shape of angel returns clearly. It does not give a reliable rate for any individual investor's chance of a 20x outcome, because that depends on deal flow, sector, stage, and judgement, none of which are captured by a single number.
Set p to your own view. If you think your access is better than average, raise it and see how much less portfolio you need. The sensitivity is instructive: at 10%, sixteen companies reach the same 80% confidence that took thirty-two at 5%.
The assumption to be aware of
The calculation treats outcomes as independent: one company's result tells you nothing about another's.
That is not true. Companies in the same sector, stage or vintage tend to rise and fall together, and a correlated portfolio has a worse chance of a winner than this maths suggests. The number the tool gives you is therefore optimistic. Treat it as a floor on portfolio size rather than a target.
Questions
Does the winner threshold change the answer?
No. Choosing 10x, 20x or 50x names what you mean by a large winner. The calculation uses p, your chance per company, and that is where your choice of threshold should show up: a 50x outcome is rarer than a 10x one, so it deserves a lower p.
Should reserves be 50%?
It is a common starting point and it is contested. The evidence on following on is genuinely mixed: one large dataset found higher mean returns from always exercising pro rata, and higher median returns from never doing so. What matters more than the number is deciding your policy in advance rather than deal by deal.
I cannot invest in 32 companies. Is angel investing not for me?
It means direct angel investing at that cheque size is not for you. Smaller cheques across more companies, or a fund or syndicate, gets you the breadth without the capital. What does not work is a handful of direct investments held with the expectation of venture-style returns.
Why does the tool assume every company has the same chance?
Because any other assumption needs information you do not have. You cannot know in advance which of your companies is the outlier; if you could, you would only invest in that one. Treating them as equal is honest about that.
Other calculators
Track every position like this, automatically
Add your real positions once and TheAngel keeps every figure on this page current, across your whole portfolio.
Start tracking your portfolio