In the mid-1980s a research psychologist named Gary Klein went out with fireground commanders to study how they made decisions under time pressure. He expected to find some fast version of the textbook model: list the options, weigh them, pick the best. That's not what he found. The commanders he interviewed mostly didn't compare options at all. They looked at the fire, recognized it as a kind of fire they'd seen before, and a course of action came to them already attached. They'd run it through in their head for a second or two, checking it against what they were seeing, and if it held up they gave the order. If it didn't, they modified it or reached for the next one. One option at a time, not a list.
He called it recognition-primed decision making, and wrote it up in a book called Sources of Power. It's the most honest description I've read of what actually happens in the front seat when the smoke's showing. Nobody's weighing a matrix. You see it, you know it, you act.
The catch
Recognition only works if you've got something to recognize. The commanders Klein studied had twenty-plus years of fires behind them. Their gut was reliable because it had been trained by a lot of real feedback: they'd made the call, seen what happened, and adjusted, hundreds of times.
Klein later co-wrote a paper with Daniel Kahneman, who spent his career documenting all the ways intuition gets it wrong, and the two of them agreed on the conditions under which a gut feeling is worth trusting. Two things: the environment has to be regular enough that there are patterns to learn, and you have to have had enough practice, with feedback, to learn them. Firefighting passes the first test. The second one is the problem. There are fewer fires than there were, so the reps that used to build the pattern library aren't coming by themselves.
That leaves the new officer in a bad spot. He's being asked to make recognition-primed decisions with a thin library. His gut will still produce an answer, because that's what guts do. It just won't be a trained one.
Where the reps can come from
This is the argument for simulation, and I want to make it carefully, because it's easy to oversell.
A scenario on a screen is not a fire. It doesn't build the physical part of the pattern: what heavy smoke feels like on your neck, how the floor sounds, how long a line takes to charge when you're the one waiting. Hands-on training does that and nothing else does. But a lot of what the commanders in Klein's study were recognizing wasn't physical. It was a shape: this kind of building, this kind of smoke, this time of day, this many people, therefore this. That part can be practiced without a fire, and the fire service has practiced it for a long time, with tabletops, with sand tables, with the senior guy telling stories in the kitchen.
What a tool can add to those is volume and feedback. Volume, because you can run twenty arrivals in an evening instead of one. Feedback, because it can tell you what it graded and why, cite the SOP, and show you the version of the call where you'd made the other choice. Klein's commanders got their feedback from the fire. A new officer can get a cheaper, weaker version of it from a screen, and cheaper and weaker beats none.
The other thing a tool can do is make the recognition visible. When you've made the same call three times on three different scenarios and gotten it wrong the same way each time, that's a pattern in you, not in the fire. The after-action shows it. That's hard to see from inside your own head, and it's the thing a good training officer has always done for people. A screen can do a rough version of it at 11 at night when the training officer's home.
Keep it in its place
None of this makes the gut trustworthy on its own. It makes it a little better trained. The point of building recognition is so that, on the real fireground, the first thing that comes to you is more likely to be right, and you still check it against the building in front of you. That check is the part Klein's commanders never skipped. Neither should you.