It has been repeatedly emphasized that proton transfers "involving" three or four membered ring transition states are not favorable and should not be drawn as the primary mechanism for any reaction.
However, with regard to a recent discussion involving $\ce{H3CSH}$ it was noted that the rate of proton tunneling in the anionic form $\ce{H3CS-}$ was likely extremely high.
Carbon-Sulfur Bond Lengths; Resonance Effects (Or Lack Thereof)
So:
a) Is this a valid rule of thumb: avoid intramolecular proton transfers unless the transition state has 5 or 6 members in its ring?
b) What's more significant with regard to proton transfer - proton tunneling or physical proton transfer? Or is the answer an "it depends"?