I had the theory page up for a long time, and sometimes people would try and poke holes in the theoretical analysis, saying that there's no way to compare dissimilar lights, or some such thing. So I decided to try an example setup, holding as many things constant and controlled as possible, using a 250 Watt Halogen bulb versus a 250 W-s monolight in the same conditions, to see how a strobe compares against an incandescent light in actual practice.
Here's the setup. A guitar to give a sense of scale, and a metronome beating at 60 beats per minute (one click per second) to give a controlled moving object that moves at a speed roughly comparable to a slightly fidgety person. This is lit with a single 500 W-s monolight set at 1/2 power (so it's 250 W-s). The light is bounced into a 60 inch umbrella, immediately at camera right. This monolight has an advertised flash duration of 1/1300s, which is more than fast enough to freeze the metronome's motion for all practical purposes. It's also good enough to easily shoot handheld, though for consistency's sake, I'm going to leave the camera on a tripod for the whole shootout.
Shooting data: ISO 100, f/8, 1/125 shutter speed but 1/1300 flash duration (motion freezing is primarily related to flash duration, not shutter speed).
Now, for a halogen light:
Shooting data: ISO 100, f/8, 2 seconds.
Yikes! I left the white balance on "daylight", and the picture looks all orange. If I were using daylight balanced film, this is about what I'd see, and I'd have an awful time adjusting this to make the colors look normal. But I'm using digital, so let's set the white balance to "tungsten".
Shooting data: ISO 100, f/8, 2 seconds.
Ah, that's better. The only changes between this picture and the original one are that I unplugged the sync cord from the monolight, thereby disabling the flash, and I turned the 250 Watt modelling light on full power, giving me 250 watts of tungsten light into the exact same lighting setup, and I set the white balance to "tungsten". Oh, and in order to capture enough light, I had to change the shutter speed to TWO SECONDS. Fortunately, I was using a tripod for all of these shots.
The theory says that 250 W-s of flash should be equivalent to a 250W incandescent light running for one second. But my light meter said I needed two seconds. That's consistent with the fact that strobe tubes are a bit more efficient at converting electrical energy into light than incandescent bulbs. Note that the two second exposure is still just a hair darker than the strobe picture due to minor metering inconsistencies, but they're within a fraction of a stop.
The "still life" portion of this photo is pretty much equivalent to the original strobe version. But note that the metronome is completely blurred, due to the 2 second exposure.
So let's try opening the lens up to fix that blurring. Fortunately, I'm using a relatively fast 35mm lens (non-zoom), so I can go all the way to f/2.
Shooting data: ISO 100, f/2, 1/10 sec.
There, that's better. But note how the depth of field is much shallower. 1/10 second exposure time isn't so bad, and may even be adequate for some human subjects, if the camera is on a tripod. But I guarantee a fidgety kid can move back and forth faster than that metronome, so let's see if we can freeze the motion by turning up the ISO sensitivity.
Shooting data: ISO 3200, f/2, 1/320 sec.
OK, now we've stopped the motion for all practical purposes, though we're still far short of the 1/1300 exposure time that the original shot used. However, in order to stop the motion, we had to open the lens all the way up and set the ISO sensitivity to the highest position available on the D200. The result isn't nearly as pretty as the strobe version. The D200 is arguably quite a bit better at handling high ISO's than a film of 3200 speed would be, but it's still pretty ugly compared to the first shot.
For still life with a tripod where exposure time can go to multiple seconds, the incandescent light is fine. But for a moving subject, a 250 W-s strobe beats a 250 Watt incandescent light with no contest.
Another way of looking at it is that, when using hot lights with a moving subject, chances are that all your settings will be compromised. You will choose an aperture a bit wider than you'd ideally prefer, for the sake of getting more light. You'll choose a shutter speed a bit slower than you'd ideally prefer, for the sake of getting more light. You'll pose your subject in a way that minimizes movement, for the sake of holding still during that longer-than-ideal shutter speed. You'll choose an ISO setting (or film speed) a bit faster and noisier than you'd prefer, for the sake of getting enough light. You'll choose light placement a bit closer than you'd ideally prefer, for the sake of getting enough light.
A tripod and completely still inanimate subject allows you to use a long shutter speed, thus eliminating the compromises of the other settings. But that doesn't work for a moving subject.
Take the same photo with sufficiently powered flash units, and you'll choose an aperture based on sharpness and depth of field considerations only. Your shutter speed will be based on synchronizing with the lights, and the flash duration will freeze motion with no problems. You'll shoot at your camera's lowest ISO rating for noise-free images. You'll choose light placement and modifiers for the effect you want without worrying so much about getting enough light. And you'll probably be running your lights at a fraction of their rated output.
All photos were taken with a Nikon D200, 35mm f/2 lens (approximately equivalent to a 50mm lens on 35mm film), using a single light into a 60 inch umbrella at camera left, with identical placement of the light and camera for all photos. The same reflector and umbrella were used for the incandescent light as the strobe, since it was the strobe's modelling light that was used. All photos are exactly as they came from the camera, except for cropping/resizing. No levels adjustment has been performed. You may click on any large photo to get the original version exactly as it came out of the camera, with no manipulation whatsoever performed. The originals are big, so only do this if you've got a high bandwidth connection.