The arithmetic
Expected returns
How to build an honest expectation for an angel portfolio, and why the average outcome and the likely outcome are very different numbers.
A power-law distribution has a property that makes intuition unreliable: the mean and the median are far apart. The average outcome across many angel portfolios can be respectable while the outcome of any individual portfolio is more likely than not to be disappointing, because the average is carried by a small number of exceptional results.
This means two true statements coexist uncomfortably. Angel investing as an asset class can produce good returns. Your particular portfolio has a meaningful chance of returning less than you put in. Both follow from the same distribution, and understanding that is the difference between an informed angel and a disappointed one.
What follows is a way of building an expectation you can hold honestly, using arithmetic you can check, rather than a number borrowed from somebody else's marketing.
Building the expectation
Take a portfolio of twenty-five investments. Assign a plausible distribution of outcomes — how many go to zero, how many return capital, how many return a few times, how many are large. Multiply through. The result is not a prediction; it is a statement of what would have to be true for the portfolio to work.
The value of the exercise is that it forces the question backwards. If your assumed distribution needs one twenty-times outcome in twenty-five investments, you can ask whether the companies you are actually seeing could produce one. Frequently the honest answer is that your deal flow does not contain that kind of company, which is a more useful finding than any projected return.
Do it twice: once with a distribution you would call reasonable, and once with one you would call disappointing. The gap between them is the range you are actually underwriting.
| Outcome | Reasonable case | Disappointing case |
|---|---|---|
| Total loss | 13 investments | 17 investments |
| Return of capital (1×) | 6 | 5 |
| Modest (3×) | 4 | 3 |
| Strong (10×) | 1 | 0 |
| Exceptional (30×) | 1 | 0 |
| Total invested | £150,000 | £150,000 |
| Total returned | £348,000 | £84,000 |
| Portfolio multiple | 2.32× | 0.56× |
Constructed illustrations to show how sensitive the result is to the top of the distribution. Not observed data.
Why the top of the distribution dominates
In the illustration above, the difference between a 2.2× portfolio and a 0.56× portfolio is two investments. Everything else is broadly similar. That sensitivity is the whole character of the asset class.
It means that effort spent improving the middle of your portfolio — turning failures into break-evens, break-evens into modest wins — barely moves the result. Effort spent increasing the chance of holding one exceptional outcome moves everything.
That reframes several decisions. It argues for backing companies that could plausibly become very large rather than ones that will probably become moderately successful. It argues for portfolio size. And it argues strongly against selling winners early.
Time, and what it does to the number
A 3× return over ten years is a compound annual rate of about 12%. The same 3× over five years is about 25%. Multiples ignore time entirely, and angel investing has long holding periods, so a multiple quoted without a period is not a return.
Convert your expectation into an annual rate before comparing it with anything else. A portfolio returning 2.5× over twelve years compounds at roughly 8% a year, which is a sobering comparison against public markets — and the honest one, given that angel money is illiquid, undiversified and at risk of total loss.
This is not an argument against angel investing. It is an argument for being clear about what return would actually justify the risk, and for noticing that the justification depends almost entirely on the top of the distribution.
In short
What to take away
- Build your own expectation from an assumed distribution rather than borrowing a headline number.
- The result is dominated by the top one or two outcomes — everything else is noise.
- Convert multiples into annual rates before comparing with any other investment.
- A meaningful chance of returning less than you invested is the normal, expected case.
From the decoder