I see where you're coming from; and I do agree that forecasters should be very judicious about their use of extreme values in the probabilities. Remember, though, those zeros do not represent a "0% chance for tornadoes." Look more closely. They actually are for "2 or more tornadoes" and "1 or more strong (F2-F5) tornadoes." The singular, weak, rogue tornado touchdown is acceptable in any severe thunderstorm watch -- even in one carrying zero probabilities for those two tornado categories.
Zero or 100 should not appear very commonly; but there would be some solidly justifiable occasions. For example, I would risk using a zero probability for the hail categories (10 or more, or a two inch stone) for a tornado watch covering a tropical cyclone landfall, where neither such condition has ever been documented nor is supported by the wretchedly unfavorable thermodynamic situation. A 100 could be justified in a necessary replacement or "tack on" watch, covering a major, violent tornado outbreak, when multiple cyclic and tornadic supercells are ongoing and moving through the area of concern shortly before the time of prior watch expiration.
Ultimately, any forecast should be judged by the results; for the atmosphere always provides the answer key to test of forecasting. In this event, Jon Racy put a zero for the probability of F2 tornadoes and for two or more tornadoes, and guess what: there were none, and for sound reasoning which I won't expend bandwidth upon but which can be found in the watch and MCDs for that event. The overwhelming majority of severe events were firmly within Jon's watch too. This one worked out as well as could be expected, especially considering the strange ambient situation characterized by the presence of 35-45 degree preconvective surface dew points in the Carolinas. :thumbup:
Thanks for the great words on the mesoscale page. I was glad they could be made ready before the bulk of the convective season.