The accuracy of ratings depends on the underlying data generating process (DGP). The author basically made an assumption: the DGP of movie ratings is normally distributed. And based on such an assumption, if a website’s movie rating is closest to normal distribution, then it has the best rating.

However, how can you justify your assumption? Let’s say, the “objective” quality of all the movies are left-skewed, which is possible. Then your result will be totally wrong.

I am just an economist doing economic theory.