[js1138]

If mutations are random, then there may be no mutations in a given child, or 1 mutation, or thousands. A deck might not be shuffled at all...it might have a single card moved...or it might be shuffled entirely.

Point mutations would be considered copy errors. Errors are edited, so any that get through would tend to be small. Large errors would likely be fatal, so they would immediately be lost.

So there is a bias towards small errors. This has nothing to do with intelligence in the sense of anticipating what is needed or what would be useful.

[Southack]

"Errors are edited, so any that get through would tend to be small. ... So there is a bias towards small errors."

I agree fully.