$4.(10$ points$)$ Students invariably ask me how much data I want them to take and I have to tell

them over and over that I have no stake in how much data they take. But I DO explain the pros

and cons of more or less data. So explain why showing a line fit to two $(2)$ data points was

never enough, why three $(3)$ data points is probably not enough, and why five $(5)$ data points is

sort of a practical minimum.