SignalFrogDiversification is the one measure in investing you get almost for free: spreading capital across many holdings lowers volatility without sacrificing expected return. It rightly counts as a foundation.
Except that it does not deliver what many expect of it.
Two kinds of risk
Single-name risk hits one company or sector — a product failure, a lost lawsuit, a management error. Spreading works excellently against this. A few dozen holdings across different sectors remove almost all of its effect.
Market risk hits everything at once — a rate turn, a recession, a geopolitical break. Against this, spreading within the asset class does nothing at all. A hundred stocks fall together in a crash.
The uncomfortable part
In calm times asset classes move differently and diversification looks good. Under stress they converge: whoever needs liquidity sells what can be sold — and that is often whatever is still showing a gain.
What actually helps
Against single-name risk: broad spreading, cheaply available through index funds.
Against market risk: a portion that is not tied to the market — cash, short-dated government bonds, plainly money. Not because it earns a return, but because it keeps you from having to sell at the worst possible moment.
The practical question is therefore not "how broadly am I spread" but "how long can I go without selling anything".