So, a cautious assumption until data shows it's not accurate that "holidays lead to increased transmission" seems to me to be warranted.
Have you seen data that shows holidays don't lead to increased transmission?
It is evidence that other things can also lead to increased transmission.
If my hypothesis were "ONLY holidays can lead to increase transmission", then that evidence would falsify it.
But my hypothesis is that "holidays can lead to increased transmission", so that evidence doesn't falsify it.
The upward curve can be a result of school openings and then the sudden steep upward trajectory in the beginning of November being due to Halloween followed by continued Holiday visitation throughout Nov/Dec. There's also the messaging from the Trump administration that encouraged masses of people to ignore rules and treat the virus as if it's "no worse than the flu".
Similarly the reduction in cases can - speculatively - be attributed to better protocols being implemented nationwide; e.g. states that had lax rules having an "oh shit" moment when hospital capacities were at maximum which in turn resulted in stricter measures. Better mitigation strategies in schools across the nation is another reason. Trump's election drama diverted attention away from the "covid = hoax", so people were more willing to accept stricter rules.
Lastly - and this is something that can't really be proven so take it with a grain of salt - people are avoiding testing in order to drive down rates so the schools can fully re-open. I think we might be able to see hints of this last one in hospitalization data if those numbers remain the same while the case numbers drop.
EDIT: https://covidtracking.com/data shows hospitalizations are also dropping (although deaths only started dropping recently), so I would say that last reason for cases dropping is bullshit concocted on social media, but I'm going to leave it on this post so others can call it out as BS when they see it elsewhere
I feel like the most likely candidate is weather. Even hot places like Florida and California get cooler, shorter days, with less humidity.
Not necessarily. They also removed one big set of gatherings: Work.
Workplaces and the commute to them are massive contributors. Many people are in (loose) contact with many more people at work than at a large Thanksgiving dinner, and contact tracing is a lot more effective for the Thanksgiving dinner (everyone there is a close contact) vs. work ("oh, but they were more than 1.5 meters apart, so it doesn't count" because you can't quarantine everyone in a 100+ people open plan office that you've been saturating with infectious aerosols for 8 hours).
To be clear I don't mean "a pretty reasonable assumption" to mean it was exceptionally likely to be true (with, say 95% confidence) in my mind in this it's more like a 65% chance of being true.
I guess I was using it as shorthand for "my prior is >50%"
Anyway, I wouldn't be shocked if that assumption was wrong and if it is, it makes sense that it'd be for the reasons you outline
And yes, I get that it's considered more morally justified by many than watching football in a stadium or playing hockey. But at some point, the virus transmission in a given scenario needs to be dispassionately assessed whether you're robbing a bank or feeding orphans in Uganda.