Why A Single Number Can Never Truly Capture The Real Threat Of Disaster

Risk is not just a single number, but a complete curve that captures every possible outcome and likelihood, according to foundational concepts revisited by researchers. Decades ago, as modern industries and complex technologies expanded into everyday life, scientists faced a growing dilemma about how to measure danger objectively.

How a 1981 Paper Shaped Modern Risk Analysis

When the journal Risk Analysis launched its inaugural issue in March 1981, it opened with a landmark contribution that would permanently change how scientists, engineers, and policymakers think about danger. Published in the journal, a seminal paper authored by Stanley Kaplan and B. John Garrick established a uniform conceptual and linguistic framework for quantifying risk that is available in Six Decades of Risk Research. Before this work, safety considerations often fell short when applied to novel hazards like nuclear power plants, space exploration, and large-scale chemical pipelines.

Kaplan and Garrick argued that conventional safety assessments failed to communicate the true scope of potential disasters. They rejected the simplistic multiplication of probability and consequences as an inadequate summary of danger. Instead, they insisted that risk requires a comprehensive evaluation of all possible scenarios, each paired with its own likelihood and severity.

The Power of the Set of Triplets and Three Basic Questions

At the heart of the 1981 framework lies the set of triplets, a mathematical and conceptual pairing designed to structure any risk analysis. To operationalize this approach, analysts must answer three fundamental questions: What can go wrong? How likely is it to happen? And if it happens, what are the exact consequences? By organizing answers into scenarios denoted as a combination of scenario descriptions, probabilities, and consequences, experts can map out a wide spectrum of potential futures.

This structure helps decision-makers compare diverse threats, ranging from natural hazards like floods and earthquakes to human-made technological failures. Crucially, the authors proposed plotting these scenarios on a graph of cumulative frequencies against potential fatalities or damages, creating what is known as a risk curve. They famously stated that a single number is simply not a big enough concept to communicate the idea of risk, as it takes an entire curve to show the true picture.

Why Single Numbers Fall Short in Disaster Planning

The reliance on a single average metric, such as expected annual damage, remains common in cost-benefit analyses and insurance sectors today. However, critics and the original 1981 framework point out that averages mask low-probability, high-consequence events that can devastate societies. When communities focus only on average yearly losses, they often remain dangerously underprepared for rare, extreme catastrophes.

Another farsighted contribution of the Kaplan and Garrick paper was the inclusion of an unknown category to account for unobserved scenarios. By explicitly forcing analysts to assign a probability to things they might have overlooked, the framework encourages preparedness for unexpected shocks. Furthermore, the authors distinguished between frequentist probabilities based on repeated experimental trials and Bayesian probabilities reflecting a state of knowledge, paving the way for modern treatments of epistemic uncertainty.

Enduring Lessons for Modern Decision Makers

As modern societies grapple with compounding climate events, systemic vulnerabilities, and emerging technologies, the challenge of measuring risk has only grown more complex. While newer concepts like exposure, vulnerability, and resilience have been integrated into disaster risk reduction frameworks over the decades, the core architecture laid out in 1981 retains its utility. Researchers note that a more rigorous application of these foundational principles could have mitigated substantial harm over the past forty years.

The research was published in Six Decades of Risk Research on January 1, 2027.