Well, 'zero' is what I actually did. For regret, you'd need to select a counterfactual of what I could have done. (And at least one of them needs to differ from zero to give you any results. Or more generally, two scenarios need to differ in at lest one aspect to produce a meaningful comparison. Identical scenarios in back-testing give identical outcomes here.)
Fair point on the shorting - I definitely prioritized retro drama over actual market mechanics there!