Margin of safety is the gap between the price you pay and your conservative estimate of the intrinsic value of the business. Buffett's bridge metaphor is the canonical illustration: when you build a bridge designed for thirty-thousand-pound trucks, you only drive ten-thousand-pound trucks across it. The engineering buffer protects you when your assumptions are wrong, when the world is harsher than you expected, or when the price falls further before recovering. The margin is not a precision number; it is a protection against the limits of your own analytical skill.
| Quality of business | Stable / tangible-heavy / predictable | Cyclical / capital-intensive | Intangible-heavy / regulated / fast-changing |
|---|---|---|---|
| Examples | Coca-Cola, See's Candies, regulated utilities | Steel mills, autos, mining | Pharma without patents, single-product tech, financial firms in tail risk |
| Suggested margin requirement | 20–30% below central estimate | 33–50% below central estimate | 50%+ below central estimate, or skip entirely |
| Why | Predictable cash flows; smaller analytical error band | Wider cash-flow range; bigger margin needed for the trough scenario | Tail outcomes can drive intrinsic value to zero; even a large margin may not cover the binary risks |
| What an owner watches | Quality persistence, capital-allocation discipline | Cycle position, balance-sheet endurance through trough | Whether the franchise is genuinely durable or whether the next disruption ends it |
Margin of Safety = (Worst-Reasonable-Case Intrinsic Value − Price) / Worst-Reasonable-Case Intrinsic Value
Worked example — Westmoor Optical at $32. The owner builds three intrinsic-value scenarios over a careful month of work: bull case $52 (premium-frame strategy compounds with the demographic tailwind), base case $44 (current trajectory continues), bear case $24 (recession compresses optional-spending categories and the strategy stalls). The margin-of-safety calculation is run against the bear case, not the base. At $32, the margin against the bear case is negative — the owner is paying more than their own pessimistic scenario suggests is appropriate. The disciplined response is to set a watch price. The owner decides: 'I will not size this position above 1% until the price drops to $26 — that gives me a 7–8% buffer against my own bear case plus a cushion for analytical error. If the price falls to $24, I will size to 3% of the portfolio.' This is the loss-avoidance-first / return-second hierarchy that Klarman, Graham, and Buffett all teach. The patience itself is the strategy.
Sit with the ideas.
Tirebridge Materials has an estimated intrinsic value of $50 per share, with a 95% confidence interval of $42 to $58. Tirebridge currently trades at $42. From a long-term owner's perspective using Graham's margin-of-safety discipline, which framing is most defensible?