Risk Is a Triangle

Transfers

The fire triangle — heat, fuel, oxygen — is the canonical model: three conditions, each necessary, none sufficient, dangerous only in combination. Remove any one and fire cannot sustain. This structure has been independently discovered across multiple domains, always with the same insight: danger is combinatorial, and mitigation is subtraction.

Key structural parallels:

• Necessary conditions, not sufficient ones — each vertex of the triangle is harmless alone. Heat without fuel is just warmth. Fuel without oxygen is inert storage. The paradigm teaches practitioners to look not for a single cause but for the combination that enables harm. This is the core transfer: stop asking “what caused this?” and start asking “which conditions converged?”
• Mitigation by subtraction — fire suppression works by removing one element: water removes heat, a blanket removes oxygen, clearing brush removes fuel. The triangle paradigm imports this into every domain it touches. The fraud triangle (pressure, opportunity, rationalization) suggests removing opportunity through controls. The epidemiological triad (host, agent, environment) suggests modifying the environment. The strategy is always the same: find the cheapest side to eliminate.
• Visual pedagogy — the geometric shape makes abstract risk concrete. Fire safety instructors draw a triangle on a whiteboard. Fraud auditors draw a triangle. The shape itself becomes a thinking tool — a way to check completeness (“have I identified all three sides?”) and a communication device (“which side are we attacking?”).
• Cross-domain isomorphism — the fire triangle, fraud triangle (Albrecht/Cressey), epidemiological triad, and lethal trifecta (Willison) are structurally identical. Each maps three necessary conditions onto the same geometric shape. The paradigm is not metaphorical in the usual sense — it is a recurring structural pattern discovered independently in fire science, criminology, epidemiology, and information security.

Limits

• The number three is a pedagogical convenience, not a natural law — real systems often have more than three necessary conditions. The fire triangle became the fire tetrahedron when fire science recognized chemical chain reactions as a fourth element. The fraud triangle has been challenged by models with four or five factors. The geometric simplicity that makes the paradigm teachable also makes it reductive. Practitioners who internalize “three sides” may stop looking after they find the third condition.
• Equal weighting is an illusion — the triangle’s symmetry implies each condition contributes equally. In practice, conditions are rarely equal. In the fraud triangle, opportunity is far more controllable than pressure or rationalization. In the fire triangle, you cannot remove oxygen from an open environment. The geometric equality can mislead practitioners into treating all three sides as equally viable mitigation targets.
• Subtraction assumes elimination is possible — “remove one side” sounds decisive but may be infeasible. You cannot remove all untrusted content from an AI agent’s environment. You cannot eliminate all financial pressure from employees. The paradigm’s clean mitigation logic can create false confidence about the completeness of controls.
• Static structure, dynamic reality — the triangle is a snapshot. Real risk evolves: conditions strengthen, weaken, and interact over time. A condition that was absent can emerge. A condition that was controlled can escape controls. The paradigm’s fixed geometry does not naturally accommodate temporal dynamics or feedback loops.

Expressions

• “The fire triangle” — the original and most widely taught instance, standard in fire safety education worldwide
• “The fraud triangle” — Cressey/Albrecht’s criminological application: pressure, opportunity, rationalization
• “The epidemiological triad” — host, agent, environment in disease transmission
• “Remove one side of the triangle” — the standard mitigation advice, used across all domains where the paradigm appears
• “Which leg are we cutting?” — practitioner shorthand for choosing a mitigation strategy

Origin Story

The fire triangle emerged from fire safety pedagogy, likely in the early 20th century, as a teaching tool for firefighters and safety engineers. It has no single inventor — it codified what chemists already knew about combustion into a visual mnemonic. The triangle was extended to a tetrahedron in the 1960s when fire scientists added the chemical chain reaction as a fourth element, but the triangle remains the dominant pedagogical form.

Donald Cressey’s fraud triangle (1953) applied the same structure to white-collar crime, though the direct influence of the fire triangle is unclear — the structural isomorphism may be convergent rather than borrowed. The epidemiological triad predates both, tracing to the germ theory era. Simon Willison’s “lethal trifecta” (2025) is the most recent instance, explicitly named after the fire triangle analogy and applied to AI agent security.

References

• Cressey, Donald. Other People’s Money (1953) — the fraud triangle
• Albrecht, W. Steve et al. Fraud Examination (2011) — systematized fraud triangle for auditing practice
• Willison, Simon. “The Lethal Trifecta” (2025) — AI agent security application, explicitly citing the fire triangle
• National Fire Protection Association — fire triangle pedagogy