Venture investors use “power law” to describe outcomes dominated by a small number of very large winners. Strictly, a power law is a particular statistical distribution; a skewed set of returns does not by itself prove that exact model fits.
For portfolio design, the practical issue is concentration of outcomes. If a few investments produce most of the proceeds, how many opportunities should an investor take, and how much should each receive?
Two calculations help. One concerns the chance of encountering an exceptional winner. The other concerns whether the position is large enough for that winner to matter.
More deals increase the chance of at least one winner
Assume, purely for illustration, that each investment independently has a 2% probability of a specified exceptional outcome. The chance of getting at least one such outcome in n investments is:
1 − (1 − 0.02)ⁿ
| Number of investments | Chance of at least one exceptional outcome |
|---|---|
| 10 | 18.3% |
| 25 | 39.7% |
| 50 | 63.6% |
| 100 | 86.7% |
The calculation supports a limited conclusion: under these assumptions, more independent opportunities reduce the chance of missing every exceptional winner.
It does not establish that 100 deals are suitable for a particular investor. The 2% input is invented for the example, not a measured venture base rate. Real opportunities differ in quality and can share exposure to the same customers, technologies or financing conditions.
An exceptional company must still produce enough proceeds
Suppose an investor deploys £1m equally across 20 companies, at £50,000 each. If 19 return nothing, the remaining investment must return £1m, or 20 times its cost, just to recover the invested capital before expenses and tax.
Now spread the same £1m across 100 companies at £10,000 each. If 99 return nothing, the survivor must return 100 times its cost to achieve the same break-even result.
Broader exposure increases the number of opportunities, but smaller positions require larger individual multiples when the rest of the portfolio contributes nothing. A probability table that omits position size does not describe portfolio economics.
Actual portfolios can also receive useful proceeds from several moderate outcomes. Include those rather than forcing every model into a single-winner story.
Connect the required multiple to ownership
Take the £50,000 investment and assume it initially buys 5%, later diluted to 2.5%. Returning £1m requires a £40m equity distribution from the company on a simple proportional basis.
For the £10,000 investment, assume the same entry valuation buys 1%, later diluted to 0.5%. Returning the entire £1m portfolio requires a £200m equity distribution.
These examples ignore preferences, fees and tax. They illustrate why an investor cannot assess a potential “fund returner” from the company's market opportunity alone. The retained ownership connects the company outcome to the portfolio result.
Independence is an assumption worth challenging
Twenty companies selling similar products to the same customer group may face a common budget shock. A hundred companies selected through one narrow source may share an exposure the headline deal count conceals.
Conversely, a specialist investor may understand a concentrated market unusually well. The relevant trade-off is between the benefits of that knowledge and exposure to common failure conditions.
Group proposed investments by the events that could damage them together. Sector labels alone are insufficient: apparently different businesses may all depend on cheap growth capital or one distribution platform.
Deal count also changes the work
Adding investments requires more selection, documentation and ownership decisions. If the additional volume reduces diligence quality or leaves no time to respond to existing holdings, the assumed probability of success may change.
The first table holds that probability constant. An actual investment programme may not be able to do so as it grows. It must also account for fees, minimum cheques and follow-on decisions.
Use the probability calculation to expose sensitivity, then build a feasible capital and time budget. Our portfolio-construction guide works through those constraints. There is no universal deal count hidden inside the phrase “power law”; the useful answer depends on access, position size, common risks and the work the investor can sustain.
FAQ
How many startups should I invest in?
The power-law idea does not supply a universal number. Work through capital, minimum cheques, ownership, costs, common risks and the time available for diligence. More opportunities can reduce the chance of missing every exceptional outcome, but only under assumptions that your actual investment programme can support.
Is a broad venture portfolio the same as a public-market index fund?
No. Private opportunities are not uniformly accessible, and entry terms, selection and fees can differ materially. A broad strategy still needs a defined opportunity set and admission criteria. Simply writing many cheques does not create representative exposure to all startups.
Does one huge winner guarantee a profitable portfolio?
Only if the proceeds attributable to your retained stake exceed losses and relevant costs elsewhere. The company's success and your investment multiple are different measures. Dilution and a small initial position can leave an exceptional business contributing less to the portfolio than its headline exit suggests.
Is the 2% winner probability in this article a measured VC success rate?
No. It is an invented input used to demonstrate the calculation. The example also assumes independent outcomes and a constant probability across investments. Actual portfolios can differ on all three points, so the table should not be used as a forecast.
