has there ever been an actual experimental study of the entropy of keysmashes
my guess: higher frequency of hand-alternation than random, emphasis on keys near the index/middle fingers, avoidance of runs of the same letter directly next to each other
a very robotic single-user instance