If I type a common sequence like "er" or "th," I do it with a single flick of the hand, not four separate ones.

You do "th" with one hand? I suggest that is less efficient than coordinating two shorter moves by the respective nearest fingers. "rt", of course, is a hand flick. Perhaps my vim navigation has biased me. "h" totally belongs to my right trigger finger and moving my left middle finger all the way over to the 't' so that a left hand flick can pull of a 'th' rapidly sounds like far too far out of the way.