We are going to have to agree to disagree on bias images. As an amateur photometrist I often take Light exposures which vary between a second or few to a minute or few — roughly two orders of magnitude — for each sub. Bright stars saturate in less than five seconds; faint ones are barely measurable after stacking an hour’s worth of subs.
Taking bias frames is a matter of a few minutes work in my experience. Each image has essentially zero exposure time (almost by definition) and at most a few seconds download time. Also in my experience, processing bias frames is essentially cost-free in most (all?) software. After co-adding the bias images to create a master bias, flush their subs to recover the storage.
Very strongly agree with the advice about taking a good number of dark, flat (and bias!) frames to average out the noise. I generally use somewhere between 10 and 100 depending on my patience and the brightness in each filter of the fluorescent panel used for the flats.
You may have gathered that there are almost as many opinions as photometrists…