Why should the "prime-based" hashcode implementation be used instead of the "naive" one?

I have seen that a prime number implementation of the GetHashCode function is being recommend, for example here. However using the following code (in VB, sorry), it seems as if that implementation giv