100 subs was massive overkill, your images should be effectively noise-free with anything over 50 subs. I had one imaging "expert" berating me for not knowing anything about signal to noise ratio as I didn't take into account the gain in signal to noise with exposure time. Problem with signal to noise increase with exposure time is that the increase is not all that big, and also it flattens off rather quickly especially if your skies are not too good. So in my location the exposure time signal to noise ratio increase flattens off in about 5-minutes meaning that I am getting no (further) S/N ratio increase in the 20-minute exposures I need to take to get faint stuff and the only way I can get the noise down in galaxy or faint nebula images is to take LOTS of subs. The reason why I have often said you need at least 8-hours of good data to get a first class image is that if you take a typical sub length as 10-minutes and then have 50 of them, you're talking just over 8-hours. It only took me about 10 years to work this stuff out
P.S. Oh - and by the way - all your 50 subs have to be GOOD. You have to go through every sub and check they are o.k. as I now from experience you can actually see the degradation in the stacked image with just ONE bad sub in 100.
P.P.S. And that is also the reason why I run multiple imagers in parallel - I increase the number of subs in a given time by the number of imagers.
P.P.P.S. Dithering is essential. So is good image calibration with good flats and bias frames.