Angel maths: the numbers no one runs
Every angel hears about power-law returns. Few work out how many cheques it actually takes to land an outlier — or what that means for cheque size.
The pitch you hear at every angel breakfast is the same: most of your cheques will go to zero, one or two will return everything. Power law. Outlier-driven. So write small cheques and write a lot of them.
It’s correct. It’s also incomplete.
The maths nobody runs is the second-order question: how many cheques do I need to write before the maths actually works?
What the data says
Across multiple angel datasets — AngelList, the BANSEA cohort studies, my own portfolio — the realised distribution looks roughly like this for early-stage angel cheques held to outcome:
- ~50% return less than 1x (or zero)
- ~30% return 1-3x
- ~15% return 3-10x
- ~5% return 10x+
Sounds workable. It’s not, until you do the next step.
The fund-of-one problem
If the top 5% of cheques drives most of your return, then to capture the expected distribution you need a sample size large enough that landing 5% feels probable, not lucky.
Statistical rule of thumb: you want roughly 30-50 cheques before your portfolio outcome converges anywhere near the underlying distribution. Below 20, you’re a gambler. Below 10, you’re buying a single lottery ticket and calling it strategy.
This has direct cheque-size implications.
If your annual angel allocation is $200K and you write 4 cheques a year at $50K each, you’ll hit 20 positions in five years. Reasonable. If you write 2 cheques a year at $100K, you’ll hit 20 in ten years — and the world has changed twice over.
The carry-cost nobody mentions
Larger cheque sizes also concentrate risk in single positions. A $100K loss is fine when you’ve written 30 of them. It’s not fine when you’ve written 5. The psychology bleeds back into selection: you start hesitating, demanding more “proof”, missing the early conviction trades that drive returns.
The cheques that did best in my portfolio were almost always the ones I would have hesitated to write at twice the size.
The conclusion that follows
Cheque size should be set by how many you can write before you run out of capital, conviction, or time — not by what feels meaningful per deal.
For most angels building from zero, that means smaller than feels right, more often than feels comfortable, for longer than feels patient.
Boring conclusion. Right one.