Catastrophic Failure Root-Cause Investigation

Definition

Catastrophic failure root-cause investigation is the practice of using independent, overlapping inquiries to determine how a high-consequence failure occurred before designing corrections and prevention.

Current Synthesis

Bill Gurley’s presentation draws one pattern across Surfside, the Boeing 737 Max, Katrina, Fukushima, and Challenger: physical mechanisms and institutional failures are often discovered together, while operators, regulators, or other implicated bodies may resist scrutiny. The Challenger episode supplies a deeper worked case in which cold-sensitive sealing, budget pressure, mixed goals, cadence claims, contractor communication, and launch authority form one causal chain. Gurley’s COVID application still does not establish which origin hypothesis is correct.

Key Claims

• High-consequence failures warrant inquiry proportionate to their human and economic cost.
• Overlapping investigations can compensate for blind spots, capture, missing authority, and institutional self-protection.
• Root cause includes both the immediate technical mechanism and the decisions that allowed the vulnerability to persist.
• Credible inquiry needs evidence access, independence from implicated institutions, and findings legible enough to support action.
• The COVID comparison is methodological: the absence of an established root cause supports investigation, not certainty about a laboratory accident.

Evidence

• Multiple-searcher pattern: Bill Gurley’s presentation cites engineering, legal, legislative, journalistic, whistleblower, family, and commission inquiries across the comparison disasters.
• Technical plus institutional cause: Bill Gurley’s presentation pairs concrete mechanisms such as O-rings, levees, seawalls, generators, and MCAS with ignored warnings, regulatory capture, or organizational decisions.
• Independent investigator models: Bill Gurley’s presentation uses Richard Feynman and Kiyoshi Kurokawa to show how credibility and independence can make technical findings actionable.
• COVID application: Bill Gurley’s presentation argues that the pandemic’s scale should have produced a major independent inquiry, while still declining to select an origin hypothesis.
• Challenger decision chain: 502 宇宙超级工程是如何垮掉的：刘怡谈「挑战者号」事故四十周年 connects O-ring behavior and ice concerns to program design, maintenance burden, prior anomalies, public commitments, and Go Fever Organizational Risk / 硬干狂热.

Counterevidence & Qualifications

The comparison cases differ in evidence access, jurisdiction, technical complexity, international politics, and elapsed time. Gurley’s presentation is advocacy and does not provide a balanced review of all origin evidence or responses from criticized institutions. The Challenger episode is also a secondary structured summary rather than a primary record. A missing consensus is not itself evidence for one hypothesis, and an organizational label does not replace reconstruction of the actual decision chain.

What Changed

• Created the cross-disaster investigation framework.
• Preserved the distinction between demanding inquiry and proving a laboratory origin.
• Added Challenger as a worked example of physical mechanism, warning escalation, and program incentives operating together.

Related Concepts

• CAPA / Corrective and Preventive Action - process that turns root cause into correction and prevention.
• Investigation Conflict of Interest - reason implicated institutions cannot be the only investigators.
• COVID Origins Investigation Uncertainty - unresolved application at the center of the source.
• Scientific Self-Correction - broader norm of revising claims when evidence changes.
• Public Service Journalism - journalism as one independent accountability channel.
• Mandatory AI Incident Investigation - adjacent domain-specific requirement for investigating high-consequence AI incidents.
• Space Shuttle Challenger / 挑战者号 - worked accident case joining component and institutional causes.
• Go Fever Organizational Risk / 硬干狂热 - escalation pattern that root-cause inquiry must test rather than assume.