Why it works
Hold one stock and your outcome is that company’s outcome. Hold thirty across different industries and no single failure can undo you, while the ones that succeed still count in full.
That asymmetry is the point. A stock can only fall 100%, but it can rise many multiples of that. Spreading across many holdings caps what any one disaster costs while leaving the upside of the winners intact.
Correlation does the work
The benefit comes from holdings that do not move together, not from the count. Two assets, each with 20% volatility, split evenly:
| Correlation | Portfolio volatility | Benefit |
|---|---|---|
| +1.0 (move identically) | 20.00% | None |
| +0.5 (loosely related) | 17.32% | Modest |
| 0.0 (unrelated) | 14.14% | Substantial |
| −1.0 (perfect opposites) | 0.00% | Complete |
Perfectly correlated holdings give nothing. Uncorrelated ones cut volatility by 29% with no change in expected return. That is the whole mechanism, and it is why owning twenty banks is not diversification while owning a bank, a utility, and a software company is.
Use the screener to spread across sectors rather than accumulating variations of the same bet, and compare two holdings to see how similar their businesses really are.
How many holdings
With genuinely uncorrelated assets, portfolio volatility falls with the square root of the number held. Starting from 20% for a single position:
| Holdings | Portfolio volatility |
|---|---|
| 1 | 20.0% |
| 5 | 8.9% |
| 10 | 6.3% |
| 20 | 4.5% |
| 50 | 2.8% |
Notice the shape. Going from one holding to five removes more than half the volatility. Going from twenty to fifty barely moves it. Real assets are positively correlated so the true floor is higher, but the pattern holds: the first handful of decisions matter enormously and the rest are refinements.
Beyond a point, more holdings mean more to monitor and no measurable benefit. If you cannot say why you own something, that is a reason to index rather than to add another name.
Allocation matters more
Which stocks you pick attracts nearly all the attention. How much you hold in stocks versus bonds versus cash determines far more of your result, because that split drives both the return you can expect and the drawdown you will have to sit through.
Set it from your horizon rather than your mood. Money needed within a few years does not belong in equities regardless of how attractive they look; money that will not be touched for thirty years is badly served by cash, which loses purchasing power with near certainty.
Then rebalance occasionally. Selling what has run and buying what has lagged is uncomfortable in exactly the way that makes it work, and it keeps the risk profile you originally chose rather than the one the market has drifted you into.
Where it fails
Correlations converge in a crisis. Forced selling hits everything at once, so the protection is weakest precisely when it is most wanted. Diversification softens crashes; it does not prevent them.
Owning many funds is not diversification. Five large-cap US funds hold substantially the same companies. Look through to the holdings rather than counting tickers.
Employer stock concentrates everything. Salary, benefits, and savings all riding on one company is the largest correlated bet most people ever make, and it is usually invisible because it accumulated gradually.