Interactive graph
Heed Every Near-Miss
Table of Contents
Heed every near-miss = treat a close call as evidence the system nearly produced serious harm — not as proof the activity was safe because nobody died this time.
When nobody gets injured, organizations often relax. The opposite read is usually right: the system failed, and luck blocked the worst outcome. A rescue, a panic incident, or a loss of control should trigger investigation and redesign before anyone runs the same activity again.
Example
A previous rescue during the same exercise is a warning to stop — not proof that the rescue team can keep handling it. Luck isn't a control layer.
Note Relationships
| Relationship | Wikilink | Reason |
|---|---|---|
| alternative | Incident Investigation | Formal probe vs immediate stop-and-redesign when the near miss is obvious |
| contradicts | Normalization of Deviance | when “we've done this before without dying” becomes permission to repeat |
| extends | Anti-Fragile Systems | Near misses are stress signals — learn before the system breaks for real |
| extends | Eliminate Before Managing | Investigation should ask whether the hazard can be removed, not only managed better |
| extends | Failure as Feedback | Close calls are feedback before harm |
| extends | Incident Investigation | Every close call deserves a root-cause pass, not a shrug |
| extends | Learning Organizations | Organizations that learn treat near misses as data, not luck to forget |
| implements | Active Knowledge Curation | Capture the close call, redesign the activity, don't let the lesson rot in memory |
