It's highly improbable, like p(<0.01), that military strikes on June 21st could directly cause a "No" resolution for Polymarket prediction markets that already closed on June 19th or 20th.
My prior is that market resolutions are time-bound; once the closing bell rings, events afterward can't change what already happened.
Think of it: if your flight was supposed to leave Monday and got cancelled, a new flight on Wednesday doesn't make you arrive on time on Monday.
We'd need truly extraordinary evidence to update this view, as the base rate for such backward causality in markets is effectively zero.