this is an excellent study based on swedish data.

herd immunity is reached LONG before the predicted 1-(1/R0) threshold.

in sweden, it looks to have been around 17%, not 60%.

this is because the model that generates that immunity threshold has issues. #more-26133">https://judithcurry.com/2020/05/10/why-herd-immunity-to-covid-19-is-reached-much-earlier-than-thought/ #more-26133">https://judithcurry.com/2020/05/1...

it presumes 2 things: homogeneity of social graph and homogeneity of individual susceptibility.

neither are true.

social graphs, like twitter, vary greatly. getting retweeted by one 200k follower account is worth 300 typical retweets.

infection is the same.

one "super-spreader" can do the work of hundreds. but, these people tend to get exposed first so, temporally, their effect slants early. they then become resistant. they then become "super-suppressors" and drive the infection curve slope way down.

it& #39;s just graph theory.

such a system spikes & collapses much faster than the homogeneous approximation model predicts

it means that working backward from early infection leads to a higher than idealized R0. they then plug that into a model that underestimates collapse

that& #39;s how you predict 60 vs 17

you get the same effect from those highly susceptible vs those who are not and that variance is multiplicative with social graph.

if you are highly susceptible and are a highly linked social node, you& #39;re worth 1000& #39;s of regular people in terms of percolation model weighting.

all in all, this is REALLY good news.

it means a lot of major urban areas in the dev world are already past this threshold.

it also means lockdowns are not needed (and were never going to be effective anyhow)