Daniel's been carrying a pack of LED road flares and a high-vis vest for years, and he's noticed something that most people never think about until they're standing on a dark shoulder at two in the morning. The gear is good. The instructions are essentially nonexistent. You get told how to switch the flare on, and that's it. Nobody tells you where to put it, what pattern to lay out, what to do differently in fog, or what you can actually expect a driver to see from half a mile out. His question is really about the physics underneath all of that. Why does fog make a bright light worse instead of better? Where do flares actually belong around a stopped vehicle? Which patterns attract attention and which ones mislead or blind people? And if you're the one pulled over in the worst possible spot, what visibility can you honestly count on?
And the thing he's circling, whether he said it in exactly these words or not, is that illumination is a signaling problem. It's not a brightness problem. More light is not automatically safer light. You can have a flare that's visible from a mile away in clear conditions and actively dangerous in fog because it's washing out the very thing it's supposed to mark.
So let's start with what light actually does when you throw it into fog.
The first thing to understand is that fog isn't just a dimmer switch. It changes the physics of how light travels. When you've got a point source, a flare or a headlight or a flashlight, the light leaves in a tight beam or a fairly tight pattern. In clear air, most of those photons travel in a straight line to your eye and you see a bright point. You can tell where it is, how far away it is, roughly what it's doing. Fog is made of water droplets, and those droplets are mostly in the range of a few microns to maybe twenty or thirty microns across. Visible light has wavelengths in the range of roughly four tenths to seven tenths of a micron. So the droplets are bigger than the wavelength, but not enormously bigger. That puts you in a scattering regime called Mie scattering.
Mie scattering being the thing that makes fog look like a glowing wall instead of a collection of droplets.
The light hits the droplet and scatters broadly rather than passing through. A lot of it bounces sideways, some of it bounces backward toward you, some of it keeps going but now it's been kicked off at a weird angle. The net effect is that a point source stops being a point. It becomes a diffuse glow. And here's the counterintuitive part. Past a certain intensity, adding more brightness doesn't buy you more visibility. It buys you more glare. The extra light scatters off the droplets between you and the flare, and that scattered light forms a bright haze in front of the source. Your eye tries to focus on the flare, but it's getting hit with light from every point in that haze. The contrast between the flare and the background drops, and the vehicle you were trying to mark disappears into the glow.
So the flare is now hiding the car instead of revealing it.
That's the failure mode. And it's worse with LED flares specifically because of color temperature. Most LED road flares are cool white or blue white. They're built to be eye-catching in clear conditions, and they are. But shorter wavelengths scatter more aggressively in fog. Blue light scatters something like several times more efficiently than red light in those conditions. That's physics. It's the same reason the sky is blue. The shorter wavelengths bounce around more. So a cool white LED flare in dense fog produces a bigger, brighter, more diffuse glare wall than a warmer source at the same output.
So the very thing that makes the LED flare great on a clear night is working against you in the one condition where you most need it to work.
And Daniel specifically raised the dazzle problem, so let's talk about what happens to the driver. When a bright source hits the eye, the pupil constricts. That's the fast response. Then the retina adapts. If the source is bright enough, you can get a temporary scotoma, a blind spot that persists after the source leaves your field of view. A driver passing a flare array at highway speed might get a direct hit from a cool white LED at full brightness, and then for a second or more they're functionally blind in that part of their visual field. At sixty miles an hour, a second is eighty eight feet. That's a long way to drive while you can't see the shoulder.
And that's before we get to the moth effect.
The moth effect is the single most counterintuitive fact in this whole topic. There's a documented tendency for some drivers, particularly intoxicated or fatigued ones, to steer toward a light source rather than away from it. The light you placed to warn people becomes a target. It's not universal, it's not every driver, but it's real enough that emergency responders have to think about it. If you put a bright light directly behind your vehicle, a tired driver who's locked onto that light may drift toward it. Which is exactly where your car is.
So the light is doing the opposite of what you wanted. It's pulling the hazard toward the driver.
And this is where it gets uncomfortable, because the moth effect isn't just a theoretical concern. There are studies of nighttime crashes where vehicles drifted off the road and struck parked emergency vehicles or tow trucks that had their lights running. The drivers weren't necessarily drunk. Some of them were just exhausted, and the light was the only thing their brain could anchor to. So they steered toward it. Not consciously. But the hands follow the eyes.
I remember reading about a case in Ohio, a state trooper who'd pulled someone over on I-71. He had his full light bar running, and a semi drifted off the road and clipped the back of his cruiser. The driver had been on the road for eleven hours. When they interviewed him afterward, he said he remembered seeing the lights, but he couldn't have told you whether they were in his lane or on the shoulder. His brain had just locked onto the brightest thing in his field of view and steered toward it.
That's exactly the mechanism. And it's why the placement question isn't just about being seen. It's about being seen in a way that tells the driver where not to go. You're not just marking your position. You're marking a path away from your position.
Which brings us to placement. The standard doctrine, the thing highway crews actually train on, is that you don't put flares directly behind the vehicle. You put them along the shoulder edge, in a line or a taper that guides traffic away from the vehicle and back into the lane. The geometry does the work. A driver sees a line of lights angling away from the shoulder, and the natural reading is that the road continues that way. The lights are saying move left, not stop here.
So the array is a sentence, not a pile of punctuation.
The distance scales with speed. The faster the road, the earlier the first flare needs to be. You want the driver to register the warning and start adjusting before they're committed to the lane position. On a highway, that first flare might need to be a couple hundred feet back. On a residential street, you're working with much less distance and much lower speeds.
The taper pattern. Flares angled to funnel traffic away from the shoulder, not parallel to it.
Right. And there are placement mistakes that are extremely common. Never in the traffic lane itself. That's obvious but people do it. Never directly behind the vehicle where they silhouette it. That turns your car into a dark shape behind a bright light, which is the worst possible contrast situation. Never in a cluster that reads as a single point source. If you put three flares in a tight group, a driver at distance sees one bright blob, not three warnings. And never so close to the vehicle that a driver only sees them after they've already committed to the lane. The whole point is to give the driver time to react before they're on top of you.
So we know where the light goes and why it behaves the way it does. The next question is what pattern you put it in, and what happens when everyone on the road is doing this at once.
Pattern is where it gets interesting, because you're trading off two different things. Steady burn versus flash. A steady light is easy to localize. Your brain can fix it as a stationary object, estimate its distance, understand what it's marking. A flashing light grabs attention faster. That's just how the visual system works. Motion and change trigger the alerting response. But a flashing light is harder to range. Your brain has to work harder to figure out where it is and whether it's moving. And a flashing light can be mistaken for a turn signal or an emergency vehicle.
So the tradeoff is attention versus legibility.
And flash frequency matters enormously. Very fast strobes, multiple hertz, are attention grabbing but disorienting. Your visual system gets hit with repeated high-contrast transitions and it can't settle. For photosensitive individuals, those frequencies are hazardous. Seizure risk. Slower, deliberate flash patterns read as warning rather than emergency vehicle approaching. There's a whole vocabulary of flash patterns that emergency services use, and civilians mostly don't know it exists.
What patterns should people actually avoid?
Alternating left right sequences. If you've got two flares alternating, that reads as a turn signal. A driver approaching at speed sees a blinking left right pattern and their brain says turn signal, not stopped vehicle. That's the wrong message. Patterns that mimic police or emergency vehicle light bars are illegal in many jurisdictions, and they're dangerous because drivers may react as if an officer is present. They slow down suddenly, they pull over, they do unpredictable things. You don't want to trigger that response. And any pattern that creates a false sense of motion. Some LED flares have chase sequences where the lights appear to move along the array. That's designed to direct traffic, but if it's set up wrong, it can look like the hazard itself is moving.
The pattern is a language, and most people are speaking it without knowing the grammar.
Then there's the visibility question. What can you actually count on? In clear night conditions, a quality LED flare is visible from a long distance. A mile or more isn't unreasonable for a bright unit on a dark road. But visible and identifiable as a warning are different thresholds. A driver might see a point of light from a mile out and not know what it is until they're three hundred feet away. And in fog, rain, or blowing snow, the usable warning distance collapses. Sometimes to a fraction of the clear condition figure. A flare that's visible from a mile in clear air might only be identifiable as a warning from a hundred feet in dense fog.
The person standing on the shoulder is making decisions based on a visibility number that doesn't exist in the conditions they're actually in.
That's the efficiency principle Daniel raised. Illumination should be placed and patterned to maximize the contrast between the hazard and its surroundings, not to maximize raw output. A dimmer, warmer, well placed light often outperforms a brighter, cooler, poorly placed one. The goal is contrast. You want the driver to see the vehicle as a dark shape against a lit background, or the flare as a bright point against a dark background. Not a bright haze with a car hidden inside it.
What happens when multiple drivers all deploy flares in the same stretch of road? A pileup, everyone's got their kit out, everyone's lighting up.
The array becomes visual noise. You've got overlapping flare patterns, different colors, different flash rates, different placements. The driver approaching that scene sees a field of lights with no coherent structure. The taper patterns don't read as tapers anymore because they're crossing each other. The moth effect risk compounds because there are more bright points to lock onto. And nobody's coordinating. Each driver is optimizing their own little array, and the result is a mess that's less safe than any single well placed array would have been.
There's a coordination problem that no civilian kit addresses.
Then there's the legal layer. In many jurisdictions, civilians deploying flares on a roadway are in a gray zone. Some places prohibit it entirely. Some places require it as part of an emergency kit. The liability if a driver is dazzled or misled by your flare array is largely unexamined. You could be doing what the packaging told you to do and still be creating a hazard that a court might hold you responsible for. The kits encourage deployment. The law hasn't caught up to what the kits actually do.
The person who buys the kit is getting a tool with no doctrine, no legal guidance, and no warning that the physics changes completely when the weather does.
That's the gap. The packaging tells you how to activate the flare. It doesn't tell you where to put it, what pattern to use, or what to do differently in fog. It's like selling someone a fire extinguisher and not telling them where to aim it.
A cool white LED flare at full brightness in dense fog is the perfect example. You've got a source that's bright enough to be seen from a mile away in clear air, and in fog it's producing a glare wall that obscures the vehicle it was meant to mark. The driver sees a bright glow and nothing else. The car is somewhere inside that glow, but the driver can't tell where, and now they're driving toward a light they can't localize.
The steady burn warm LED flare at the shoulder edge, that's the comparison case. A warmer source scatters less in fog. A steady burn is easier to localize. Placed at the shoulder edge rather than directly behind the vehicle, it marks the boundary of the road without creating a target. A driver registers that as a fixed point, understands it as a warning, and steers away from it. The dimmer light is doing more work.
We've been talking about this as a geometry problem. Where do the lights go, what angle, what distance. But there's a whole other layer here that we haven't touched.
Hilbert: It was a wind problem.
Say more.
Hilbert: I spent two winters working night shifts for a county road crew upstate. Late eighties. Part of the job was setting out flares and reflective markers on ice slicked two lane roads before the plows came through. You'd lay out what the book said, and then the wind would come through and the whole array would drift. The old fusee style flares, they're basically a tube with a spike on the bottom. You stick them in the snow, but if the snow's hard packed or it's ice, the spike doesn't bite. So they roll. Wind catches them and they roll into the lane. Now you've got a burning flare in the middle of the road, which is the one place you definitely don't want it.
The LED pucks don't roll.
Hilbert: They don't. That's the thing the new ones got right. But they also don't melt through ice. A fusee burns hot enough to sink down through a snowpack. An LED puck sits on top. Plow goes by, the wake buries it. You come back and your whole array is gone. Just little bumps in the snow where the lights used to be.
The geometry was right and the weather erased it.
Hilbert: The crew had an unofficial rule. One flare per ten miles an hour of speed limit, plus one. Fifty five mile an hour road, six flares. I never saw it written down anywhere. Don't know if it's still taught. But that was the number. You lay them out in a taper, first one way back, then each one a little further from the shoulder. And then you hope the wind doesn't rearrange it for you.
One per ten plus one. That gives you a spacing that scales with stopping distance. At fifty five, six flares gives you enough room to draw a line that a driver can read as a line. If you've only got three, the gaps are too big and it doesn't read as a pattern.
Hilbert: I still have a box of the old fusees in the garage. My wife's asked me three times to throw them out.
Of course she has.
Hilbert: They're probably no good anymore. But I can't bring myself to do it. You spend two winters relying on something, you don't just toss it.
The fusee is a different tool entirely. It's not just a light source, it's a heat source. That's why it melts through ice. And it's a flame, which means it works in conditions where batteries might fail. But it's also a fire hazard, and it burns out, and it's one time use. The LED puck is better in almost every way except that one thing. It doesn't melt through anything.
The newer tool solved the rolling problem and created a burial problem.
Hilbert: That's about the size of it.
The rule of thumb, one per ten miles an hour plus one, that's exactly the kind of thing that never makes it into the packaging.
Because the packaging is written by someone who's never stood on a dark road in the snow. They know the product, not the conditions. The conditions are where the doctrine lives. And the doctrine is mostly oral. It's passed between people who've done the job, not written down in manuals. That's why Daniel's question is so good. He's asking for the thing that everyone assumes exists and mostly doesn't.
What would a responsible kit even look like? If you were going to sell someone a pack of LED flares and actually tell them how to use it, what would you put in the box?
A single card. Not a manual, a card. With a diagram of the taper pattern. First flare at a distance that scales with the speed limit. A note that says fog changes everything, reduce brightness if you can, use warmer color if you have the option. A warning not to use flash patterns that mimic emergency vehicles. And a line that says if you're not sure, one steady light at the shoulder edge is better than six lights in a confusing pattern.
The same light that warns one driver can blind or attract another, and there's no way to know which driver is coming.
That's the deeper tension. You're making a decision in the dark, under stress, with incomplete information, and the physics is working against you in ways that are counterintuitive. Brighter isn't safer. More lights isn't safer. The best you can do is understand the principles and apply them as well as the conditions allow.
The principle is contrast, not brightness. You're trying to make the hazard visible against its background, not to outshine the darkness.
Which is a good place to land. The next time you're on a dark road and you see a line of lights angling away from the shoulder, you'll know what you're looking at. Someone thought about the geometry.
If you've got a question like Daniel's, something you've noticed about the world that doesn't quite add up, send it to us. We'd love to pick it apart.
This has been My Weird Prompts. The human AI collaboration podcast. Thanks to our producer Hilbert Flumingtop for keeping us on the air.
Email us at show at my weird prompts dot com. We'll be back soon.
