If you're using a telescope, say with a modest focal length of 1000mm, then if it is held still on a tripod, then the stars will move across the field of view a

If you're using a telescope, say with a modest focal length of 1000mm, then if it is held still on a tripod, then the stars will move across the field of view at a rate of a few pixels per second (depending on the exact sensor, and what part of the sky you're pointing at). Therefore, each individual picture will need to be less than a second in duration, before stacking in software. It does work. I have produced acceptable images of the Andromeda galaxy by using a 150-600mm lens at 150mm on a tripod taking 1.3s exposures. I took 1151 images, which is a total of 24 minutes of light collection, and the image was rather mediocre. I then make a rudimentary tracker, and took 22 exposures of 6s each, at 600mm, and the results were way better. At 600mm, I would have had to take 0.3s exposures untracked, and it would have been awful. I then bought a proper tracker, and did 254 exposures of 30s each at 600mm, and the result was just so much better. The problem is that a lot of the pretty stuff we can make pictures of is very dim, so the main thing driving image quality (assuming the optics are OK) is the noise level of those dark pixels. There are several reasons why it is better to have a physical tracker. One is that the noise from the sensor consists of shot/Poisson noise (from the photons coming in), dark current, and readout noise. If you have very short exposures, then the readout noise dominates. The shot noise is a physical limit - you can't get lower noise than the shot noise for the number of photons that you collect; you can only improve that by collecting more photons. The readout noise is entirely added by the equipment, and you can minimise that by taking longer exposures. The dark current is improved by cooling the sensor down. Other reasons are the data quantity (if you're processing it later or keeping the raw data, as most astrophotographers do), processing time, and wear on the shutter mechanism (if it is mechanical). Also, if you have a reasonably long telescope, if you don't have a tracker then the subject will keep moving out of the frame in less time than you would expect, and you'll have to keep re-pointing the telescope anyway. Lucky imaging is the opposite - then you do deliberately want to take lots of short exposures to get sharper images, but the trade-off is that the noise is higher, and you really want to have a sensor with an electronic shutter. Short exposures are only suitable when there is something in the frame that is bright enough to register as a reasonably bright few pixels, otherwise there is nothing to align the frames with, and lucky imaging can't work out how sharp each frame is.