OK, so I know that statistics aren't often black and white and there's often a bigger picture, but here's there findings:
View attachment 3558744
Does that mean that there's 7 out of 14 where it didn't happen, making is just as likely it doesn't? I KNOW there's some stupid simple "X doesn't mean Y" thing I'm missing, but I've not been sleeping well, so forgive me.
Not quite, 7 out of 14 doesn't imply "it's a coin toss whether 'Trump Rally' causes 'Covid Outbreak'".
Simplifying, you have the overall set of all US cities, assuming they were all comparable. Out of all those cities, some experience outbreaks, so over time you can estimate what fraction of cities "organically" has an outbreak.
For a small subset of those US cities, ie the 14 where Trump held rallies, it seems we know that this fraction is 0.5.
Assuming all this were comparable in such a simplified way, it'll come down to what that organic fraction is.
If outbreaks happen frequently then there is a reasonable chance that 7 out of those 14 cities just happened to have an outbreak a few weeks after a Trump Rally.
If outbreaks happen less frequently then it's very likely that the rally causes outbreaks.
Converting to even simpler terms, how likely is it that a die is manipulated if you roll it 14 times, and get a 1 in 7 of those 14?
Well, that depends on the die
If it's a two-sided die (coin) then 7 1s don't really indicate manipulation.
If it's a d20 then 7 1s out of 14 are very unlikely unless the d20 is manipulated.