How Analysis of Competing Hypotheses (ACH) Helps Make Sense of Complex Accidents

When an accident is serious, tragic, and technically complex, the hardest task is often not collecting information, but making sense of competing explanations. Different witnesses see different things. Technical systems interact in ways that are not obvious. Early narratives can harden before the evidence is fully examined.

NOTE: The MRH-90 incident is a real-world example of a tragic event. This ACH example is broadly accurate but is no substitute for the actual report findings. I have simplified and modified the details of the event to produce an easy to understand example. No disrespect is intended to any person or the Inquiry.

What is Analysis of Competing Hypotheses?

ACH is a structured analytical method originally developed in the intelligence community to deal with complex, ambiguous problems where multiple explanations are plausible and the evidence is incomplete or contested.

Unlike conventional analysis, ACH does not ask: Which explanation has the most support?

Instead, it asks: Which explanation is least contradicted by the evidence?

Introductory context: the MRH-90 accident

The MRH-90 accident involved a military helicopter conducting a night training flight on 28 July 2023. The helicopter entered a rapid nose-down descent and impacted the water, resulting in the loss of the aircraft and crew. The flight was being conducted at low altitude, in formation, and using night-vision devices.

In response, a formal inquiry was established to examine the circumstances of the accident, including the sequence of events, actions of personnel, aircraft systems, environmental conditions, and organisational or training factors. The inquiry was not limited to a single cause and was not bound by rules of evidence; its purpose was to make independent findings based on the available material and to consider both primary and contributory factors.

A simple illustration using the MRH-90 accident inquiry

One structured way investigators deal with this problem is Analysis of Competing Hypotheses (ACH). Rather than starting with a preferred explanation and looking for supporting evidence, ACH asks a different question: Which explanations best survive contact with the evidence, and which explanations struggle to account for key facts?

This approach is particularly relevant in aviation and coronial contexts, where inquiries are explicitly tasked with examining human actions, equipment, environmental conditions, and organisational or systemic factors, rather than searching for a single, simple cause.

What ACH does differently

ACH does not try to prove one story is “right.” Instead, it tests multiple plausible explanations at the same time, using the same body of observable information.

The method focuses on inconsistencies, not volume. A fact that contradicts an explanation matters more than many facts that merely fit it. This helps counter well-known human tendencies such as confirmation bias and early narrative anchoring.

The inquiry materials show an explicit intention to consider multiple causal pathways and contributory factors, rather than attributing the accident to a single cause  .

An Inquiry-room illustration: four competing explanations

Based on the scope and terms of reference described in the inquiry materials, at least four broad explanations are logically possible. These are hypotheses, not findings.

• H1: Pilot action or decision-making error

• H2: Aircraft system or equipment malfunction

• H3: Environmental or operational conditions

• H4: Organisational, training, or systemic factors

The inquiry documents explicitly state that matters such as acts or omissions of persons, equipment issues, environmental conditions, and training or procedural systems fall within scope.

Testing the hypotheses against observable facts

ACH proceeds by identifying observable inputs — things that were recorded, seen, or documented — and asking whether each input is consistent, inconsistent, or neutral with respect to each hypothesis.

Below is a simplified illustration, using only high-level facts described in the inquiry documents. It is not exhaustive and does not represent conclusions.

Simplified ACH summary table

Observable input (from documents)	H1 Pilot action	H2 System issue	H3 Environment	H4 Systemic factors
Aircraft climbed and then pitched nose-down before impact	Consistent	Neutral	Neutral	Neutral
Radio call “pull up, pull up, pull up”	Consistent	Neutral	Neutral	Neutral
Night operations using night-vision devices	Neutral	Neutral	Consistent	Neutral
Formation flying over water at low altitude	Neutral	Neutral	Consistent	Neutral
Inquiry explicitly considers equipment malfunction	Neutral	Consistent	Neutral	Neutral
Inquiry examines training and procedural adequacy	Neutral	Neutral	Neutral	Consistent
Inquiry seeks independent findings, not bound by rules of evidence	Neutral	Neutral	Neutral	Neutral

What this approach clarifies

This type of structured comparison shows several important things clearly:

• No single fact automatically proves a particular explanation.

• Some facts affect multiple explanations at once.

• Certain explanations begin to accumulate tension with the evidence, while others remain broadly consistent.

ACH does not determine truth. It narrows the field by identifying which explanations are least contradicted by what is known, and which explanations struggle to accommodate key observations.

Why this matters in intelligence, inquiries, or court rooms

In formal proceedings, decision-makers are not persuaded by confidence alone. They are persuaded by reasoning that can be traced, tested, and explained.

ACH produces an explicit reasoning trail. Anyone can see:

• which explanations were considered,

• which facts were tested, and

• why some explanations weaken under scrutiny.

This transparency is particularly valuable in environments where findings may later be examined, challenged, or revisited.

Closing observation

In complex accidents, the most dangerous assumption is that the explanation is obvious. Structured methods like ACH exist to slow reasoning down, expose hidden assumptions, and prevent early narratives from substituting for analysis.

ACH does not make decisions easier. It makes them more defensible.