Confidence Is Often the Last Stage Before Failure.
Please see footnote about video after the story.
Confidence feels like control, that’s why it’s dangerous. In most environments, confidence is treated as a sign of competence. The person who sounds certain is trusted. The processes that have been carried out before are often repeated. A situation that looks familiar is handled the same way. We become efficient, because it feels right, until it isn’t.
Our confidence rarely appears all at once, it builds over time. We make a decision, it works, by definition, we use it again. Through repetition, the same issue results in a streamline, maybe some shortcuts, same result, so it now becomes routine. A situation is handled successfully, so it’s assumed to be understood. Over time, repetition becomes trust, and trust reduces scrutiny. That’s where the shift happens.
The question stops being: “Is this correct?”, and becomes: “We’ve seen this before, haven’t we?”. That assumption is rarely challenged. Because familiarity feels like knowledge. In operational environments, especially in emergency services, this pattern shows up constantly. A call that sounds routine gets treated as routine. Situations that resemble previous incidents are approached the same way. A fire alarm went off at a local store for the seventh time this month, you pile in the Engine and respond to be canceled seconds before you arrive. The seventh alarm off do you still put on your SCBA? Do you go through a plan “if this is the real thing”? Decisions are made quickly, because they’ve been made before. But on that seventh alarm, what happens?
On the eighth alarm, they found her. Third floor, back bedroom, smoke inhalation. Alive, barely. The crew that responded had run that address seven times. They knew the building, knew the layout, knew the owner’s habit of burning dinner. They knew everything except that on the eighth call it wasn’t dinner. Confidence didn’t kill their awareness all at once. It borrowed it slowly, one uneventful call at a time. Until the one that wasn’t.
Most of the time, that works. But when it doesn’t, it fails fast, because the moment something is treated as known, it stops being examined. Details get filtered instead of evaluated, deviations are overlooked instead of questioned. Signals are interpreted to fit expectation, not reality, that’s the cost of confidence. Overconfidence doesn’t look like arrogance most of the time. It looks like familiarity, experience and may sound like: “We’ve handled this before.”, and that’s the problem.
No two situations are identical; they may share characteristics or resemble something you’ve seen before; but the differences are where the risk lives. Those differences don’t announce themselves. They sit just outside what you expect to see. If you’re operating from confidence instead of awareness, you won’t see them.
This is how blind spots form, not from lack of knowledge, but from assumed knowledge. The process becomes trusted, the outcome becomes expected, and questioning becomes unnecessary. Until something breaks that shouldn’t have. In hindsight, the signals are obvious, they always are. A detail that didn’t quite fit, a condition that was slightly different, or a response that didn’t match the situation as cleanly as expected. But those signals weren’t missed because they were invisible. They were missing because they were dismissed. Confidence didn’t eliminate risk, it reframed it.
The goal isn’t to remove confidence, it’s to keep it from replacing attention. The most dangerous point in any process isn’t uncertainty, it’s certainty that hasn’t been tested.
Get Kevin Schell’s stories in your inbox
Join Medium for free to get updates from this writer.
Confidence doesn’t fail loudly. It fails in the three seconds between recognition and response, when your brain is already running the playbook for a situation that turned out to be something else entirely. By the time you realize it isn’t routine, you’re already behind it. That gap between what you expected and what was actually there, that’s where people get hurt. That’s where careers end. That’s where after-action reports begin.
Nobody walks into a situation thinking they’re overconfident. They walk in thinking they’re prepared. That’s the distinction that never shows up in the debrief. Preparation and assumption feel identical from the inside, right up until they don’t. The only reliable difference between them is whether you kept asking questions after you thought you already knew the answers.
Which is the problem with this line of work.
When you find answers, someone rewrites the questions.
About the Author
Kevin Schell is an intelligence analyst specializing in behavioral threat assessment, protective intelligence, and open-source investigation. His work focuses on identifying escalation trajectories, synthesizing fragmented indicators across organizational domains, and producing structured, defensible risk assessments under conditions of uncertainty.
*Footnote — Lets talk about the video for a moment. AI in its most infinite of glory. I must have made 20 attempts with different verbige to make 3 firefighters, full turnout, non chalantly responding to a false alarm. The call turns out to be a fully engulfed building so they need to hurriedly don their masks. Kind of the premise of the article. Of all the attempts, this was the most classic, and I had to use it. It seems SORA2 wanted to not only put three firefighters in the mix, but felt adding a civilian right in the middle was necessary. I mean he dons the SCBA better than anyone in any other video! But, I guess SORA believes he has been riding Atlanta fire engines for YEARS at this point, nobody questions him, he shows up, he sits in the middle seat, he minds his business, and the moment things get real he reaches into his purple shirt and produces a face mask that he apparently carries everywhere.
This dude IS the behavioral threat assessment case study. Consistent presence, no obvious red flags, trusted by the crew through pure repetition, completely normalized. Nobody tracked the pattern. Nobody questioned the deviation from expected personnel.
He was there the whole time.
And for 10 credits he will continue to be there. “The Most Dangerous Person in the Room Was Riding in the Jump Seat the Whole Time and Nobody Asked Any Questions” — should be my next article.
Behavioral opacity. Baseline established through repetition. Zero disruption until the moment of crisis.
He was Part 1 this whole time and I didn’t see it coming.
