The piston really only approximately goes straight up and down.
Engines are not perfect machines. If you looked at the general
dynamic alignment of everything on the microinch level etc, there is
probably a slight amount of rotation inherent in the structure of an
engine including the crankshaft and rod deformations. When combined
with the varying pressure fluctuations on each ring, the ring
gradually rotates
in one direction or another. Remember you are dealiung with something
that is cycling back and forth 40 times per second, so that it doesn't
take very much rotation each cycle to give a lot of eventual motion.

Put it another way - why shouldn't it rotate?