In the actual experiment I cited, every point mutated. It really makes no difference in general. No organism knows in advance what mutation will enhance its reproductive potential.
Yes, 1953 points.
But every point didn't mutate in every possible combination (i.e. "all possible permutations") as you claimed...because that would require 4 to the 1953rd power for the number of mutations...a mathematical result vastly greater than the mere 20 million total mutations observed in the experiment that you cited.
Now, are you still going to avoid answering if mutations are random or biased?