> doubling the size of the hash table whenever
  > there are too many entries in it.

And then re-hashing and dealing with all that implies. Surely all of that is much harder than chasing a pair of pointers down the list.

Also, it doesn't work in the case of, say, Pollard's Rho method of factoring. I really need to change the original article to make it clear that that's a case I'm interested in.