Yes

"But same process will happen in a single long exposure anyway. Whether you artificially divide it into 5 shot exposures or not is irrelevant."

Yes. That is why my title was "same".

"Also, your theory (if it is true, which I doubt it) only apply to white noise. Under low light condition, the noise could be shot noise (the statistic fluctuation of number of photons hitting the pixel). Shot noise has an 1/f spectrum. You can not add them up to cancle it. "

The statistic fluctuations has a Gaussian distribution around the mean. The mean is the signal (which adds up linearly), and fluctuation is the noise, which partially cancel out (more precisely, the square adds up).