Lumens and candela answer two different questions, and only one of them tells you how far a flashlight will actually reach. Lumens count every bit of light the source emits in every direction. Candela measures how densely that light is packed into one direction. A 500-lumen light can put a usable spot on a treeline 500 metres away while a 2,000-lumen light beside it fades out at 300, and both spec sheets can be perfectly honest.
The one rule worth carrying: lumens decide how much area gets lit, candela decides how far the light carries. Everything below is how to apply that to a spec sheet before you spend money.
Lumens and candela are not competing numbers
A lumen is luminous flux, the total visible light leaving the source summed over the entire sphere around it. A candela is luminous intensity, the light leaving in one particular direction per unit of solid angle. One candela equals one lumen per steradian.
That relationship explains the entire problem. A bare 1,000-lumen emitter with no optic radiates its output evenly over 4π steradians, so any single direction gets roughly 80 candela. Put the same emitter behind a deep reflector or a total-internal-reflection lens and you can squeeze the same 1,000 lumens into a ten-degree cone, at which point the candela figure lands in the tens of thousands. The LED never changed. The optic is the product.
Two lights with identical emitters and identical lumen ratings can therefore behave like different tools. That is why comparing boxes by the biggest number on the front is close to useless.
The one formula worth memorising
Illuminance falls off with the square of distance:
E = I / d²
where E is lux at the target, I is candela, and d is distance in metres. ANSI/NEMA FL 1 defines a flashlight's beam distance as the point where illuminance drops to 0.25 lux, roughly the light of a full moon on a grey surface. Solve the formula for that threshold and the industry's throw equation falls out:
d = √(I / 0.25) = 2√I
The beam distance printed on a box is just the peak candela number, square-rooted and doubled. Once you know that, you can check any spec sheet in your head, and you can spot the ones where the distance and the candela were not derived from each other.
Ten real lights, checked against the formula
These are lights in the EDCDeal catalog that publish both a lumen and a candela figure. The distance column is my arithmetic, not the manufacturer's claim, so the comparison shows whether the two published numbers agree.
| Light | Output | Peak candela | Beam distance (2√cd) | Price |
|---|---|---|---|---|
| Streamlight MicroStream USB | 250 lm | 1,150 cd | 68 m | $42.79 |
| Streamlight MacroStream USB | 500 lm | 2,000 cd | 89 m | $59.30 |
| Streamlight Wedge XT | 500 lm | 1,900 cd | 87 m | $97.89 |
| Streamlight Wedge | 300 lm / 1,000 lm burst | 1,200 cd | 69 m | $99.33 |
| Streamlight ProTac HL-X | 1,000 lm | 27,100 cd | 329 m | $96.52 |
| Streamlight ProTac HL-X USB | 1,300 lm | 35,000 cd | 374 m | $101.96 |
| Streamlight ProTac 2L-X USB | 500 lm | 68,000 cd | 521 m | $70.49 |
| Streamlight Stinger 2020 | 2,000 lm | 25,000 cd | 316 m | $170.00 |
| Olight Odin S | 1,500 lm | 15,600 cd | 250 m | $149.99 |
| Acebeam TAC AA 2.0 | 900 lm | ~32,400 cd (360 m claimed) | 360 m | $23.99 |
Ten of ten reproduce the published beam distance to within a metre or two, which tells you the numbers are internally consistent: one was derived from the other, the way the standard intends. When a listing quotes a beam distance that does not survive this check, at least one of the two figures is decoration.
The pair that makes the point
Take two Streamlights from the same catalog. The ProTac 2L-X USB produces 500 lumens at 68,000 candela, which is 521 metres of beam distance, for $70.49. The Stinger 2020 produces 2,000 lumens at 25,000 candela, which is 316 metres, for $170.
The more expensive light makes four times the output and throws 205 metres less. Nothing is broken. The Stinger is a duty light built to flood a room or a roadside with even, wide light, so its candela is deliberately spread across a broad hotspot. The 2L-X is a pencil beam meant to identify something at the far end of a field. Buy the wrong one for your task and you will conclude the light is bad, when the specification was simply answering a different question.
The same lesson runs at the cheap end. The 250-lumen MicroStream USB at 1,150 candela reaches 68 metres; the 500-lumen MacroStream at 2,000 candela reaches 89 metres. Doubling the lumen figure bought 21 metres and a wider, friendlier beam, not double the reach, because the extra output went into area rather than concentration.
"1,000 lumens" on a pocket light is usually a burst
The Wedge is the clearest example here: 300 lumens constant-on with a 69-metre beam and three hours of runtime, plus a 1,000-lumen burst for the moments when you need a few extra seconds of reach. The headline number is a mode, not a ceiling you can work under.
For a long time this was the biggest gap in flashlight marketing, one number on the box and a driver that throttled after forty seconds. The 2025 revision of the standard, ANSI/PLATO FL 1-2025, added standardised measurement and labelling for short-duration elevated brightness, so burst and sustained output can now be declared separately instead of quietly conflated. Until the sheet you are reading separates them, assume the largest figure is the burst.
Beam shape: candela only describes the hotspot
Candela is a peak value measured at the brightest point in the beam. It says nothing about the corona and spill around that hotspot, which is where most of a light's usability actually lives.
Two lights can both claim 2,000 candela and feel completely different. One puts a tight coin of light at 90 metres and leaves your feet dark; the other lights a five-metre circle from arm's length out to 30 metres. For finding a dropped screw in a garage, spill wins. For picking a shape out at the edge of a field, hotspot wins. A candela figure cannot tell you which one you are holding.
A practical test: ignore the peak number and look at beamshots at 5 m, 20 m and 50 m. If all three are useless, the peak figure was decoration.
How much throw you actually need
Running 2√I backwards tells you the candela a job requires.
| Distance | Candela at 0.25 lux | Typical job |
|---|---|---|
| 10 m | 25 cd | under a car bonnet, reading a meter |
| 25 m | 156 cd | checking a gate or a driveway apron |
| 50 m | 625 cd | yard, path, campsite perimeter |
| 100 m | 2,500 cd | treeline, trail junction |
| 200 m | 10,000 cd | field, water, open ground |
| 300 m | 22,500 cd | identifying a shape at the far edge |
| 500 m | 62,500 cd | search and signalling |
Two honest caveats. 0.25 lux is dim: enough to see that something is there, not enough to see what it is. If you need to read detail rather than detect a shape, aim for 1 to 5 lux at the target, which cuts usable distance to roughly half to two-thirds of the rated figure. A 329-metre light gives you perhaps 165 metres of real identification, which is still a great deal.
Second, candela is quoted at switch-on with a full cell. As the driver steps down thermally and the battery sags, both figures fall. The runtime on the box is measured to 10 % of initial output, so it describes the tail of a curve, not a plateau. The inverse-square law applies to a light that is dimming, too.
Reading a spec sheet in sixty seconds
- Separate sustained output from max or burst.
- Find candela or peak beam intensity, not just a beam distance.
- Check which mode the candela figure belongs to, usually the highest one.
- Read runtime to 10 %, and note the mode it applies to.
- Note the emitter and optic: the same emitter behind a deep reflector and behind a TIR lens are different products.
What this looks like in practice
For a pocket light you will actually carry, 500 to 1,000 lumens is plenty, and 2,000 to 27,000 candela covers most urban and suburban range. A MacroStream at 2,000 candela and a ProTac HL-X at 27,100 candela live in the same pocketable size class; the difference is how much of your field the beam covers when you use it.
If you want AA cells, the Acebeam TAC AA 2.0 shows what a tight optic does with 900 lumens: 360 metres on a single AA, because the head is deeper than the cell suggests. If you want a light for a keyring instead of a pocket, the trade-offs change completely and candela per gram becomes the only number that matters, covered in the keychain flashlight guide.
For the wider question of what an EDC light is expected to do, start with what actually matters in an EDC light. If your real concern is a multi-day outage rather than reach, runtime and battery survival is the more useful read. And if the numbers above have you re-checking the box the Wedge XT came in, the lumen definition is worth five minutes.
FAQ
Is candela more important than lumens? Neither is more important; they answer different questions. Lumens tell you how much total light the flashlight produces, which governs how large an area you can illuminate. Candela tells you how concentrated that light is, which governs how far the beam reaches. For close work and room lighting, lumens and spill matter more. For seeing across a field, candela matters more.
How do I convert candela to beam distance? Use the ANSI FL 1 formula: beam distance in metres equals two times the square root of peak candela. So 2,000 candela is about 89 metres, 27,100 candela is about 329 metres, and 68,000 candela is about 521 metres. Manufacturers derive their published beam distance from candela the same way, which is why the two figures should always agree.
Why can a 500-lumen flashlight throw farther than a 2,000-lumen one? Because candela depends on the optic, not just the emitter. A light with a narrow beam angle concentrates its output into a small hotspot and reaches far on modest lumens. A high-output light with a wide reflector spreads its output over a large area, which is better for lighting a room or a roadside but produces less peak intensity and therefore less throw.
Is a 1,000-lumen mode usable for a whole task? Usually not. On compact lights the top mode is a burst, limited by heat and battery current, and the driver steps down to a lower sustained level. Check whether the spec sheet separates maximum and sustained output, and read the runtime figure, which is measured to 10 % of initial output. The mode you can actually work under is the sustained one.
What beam distance do I need for walking a dog at night? Somewhere between 50 and 100 metres of rated beam distance covers most suburban walking, which corresponds to roughly 625 to 2,500 candela. That range is met by many 500-lumen pocket lights. The more useful spec at walking distance is spill width at 3 to 10 metres, since that is where you actually look.