This is part of a long and excellent thread that is worth reading in its own right. I am almost entirely in agreement. With the exception that I want to write more about the risk of airborne transmission in superspreading 1/n https://twitter.com/DFisman/status/1292287203045933057

Superspreading likely follows a "Series of Unfortunate Events" (great show) however we also think that the transmission of this virus is driven by superspreading events, so they must be common enough to make up for the many that do not transmit 2/n https://wellcomeopenresearch.org/articles/5-67 

And those superspreading events are hard to explain without some recourse to airborne/aerosols. Look at this from @kakape quoting @ChristoPhraser https://www.sciencemag.org/news/2020/05/why-do-some-covid-19-patients-infect-many-others-whereas-most-don-t-spread-virus-all 3/n

And yet there's very good evidence that airborne transmission is rare among well studied transmission events in for eg healthcare. I will note that I hear and respect the likes of @j_g_allen but transmission is different from presence of droplet nuclei. Bear with me though... 4/n

To my mind it is impossible to simultaneously believe that the over distributed R (superspreading) is important, without also accepting that the mode of transmission that enables it is also important 5/n

The observed data can be explained by a *short* period in which some people are capable of infecting many more. And then they become less infectious. Nobody says you have to be equally infectious over the whole period you are infected 6/n

What does this mean for interventions? Actually much of the same. You limit gathering sizes, you limit the size of superspreading events. It remains about limiting contacts 7/n

And yet, if you are caught in a superspreading event you are surely exposed to a higher risk of infection. And the persistence of the virus in the population suggests such events must be important 8/n

The risk of transmission, and the risk by different routes varies over the course of infection. For most of the time I think it is probably pretty low 9/n

However the existence of superspreading events show that for *enough* it has to be super high. And I find it hard to explain these without a short-lived capacity for airborne transmission 10/n

And it is widely agreed that such superspreading events are essential components of the transmission dynamics so... ? 11/n

I think that transmission risk is low other than at the point of peak viral loads, which usually happens earlier in infection. I *think* that airborne transmission may happen from at least a fraction of people at that stage. How many they infect depends on how many contacts 12/n

What does this mean? Well first I want to be clear that this is an opinion based on what I see in the data. But I cannot find a way to get "overdispersed R0 driving things" into my head without a role for the mode of transmission that drives that over dispersion 13/n

But for crying out loud, this indicates also the importance of freaking distancing. If you aint there, you're not gonna be the infected or infectee. So distance and find ways to maintain it! (sorry I truly appreciate this is hard but it is also not my wheel house) 14/n

in conclusion we are almost where we were, but hopefully recognizing the importance of an overdispersed reproductive number and what may well be driving it. Distancing, testing and tracing remain crucial. Take care of yourselves, and others x 15/end