#Sweden what explains decline in deaths #SocialDistancing or #herdimmunity. In all regions that had typical curve, daily death peaked when total deaths = 0.02-0.04% of pop (yellow Bars) @MLevitt_NP2013 @AlistairHaimes @FatEmperor @EffectsFacts @gummibear737 @GunnarBj 
I've omitted last 2w to remove reporting lag. Region by region lets look. First - Stockholm
Sodermanland
Ostergotland
Gotland
Dalarna
Jamtland
Vastmanland
Orebro
Vastragotaland
Kalmar
For comparison here's London showing same pattern (but of course only lockdown explains decline!)
Some regions do not fit the pattern well, but these are regions with no typical curve and which had lots of negative excess death during March-Jun (which I've omitted to try and isolate covid deaths) e.g. So it's not clear if covid really took off in these places.
Skane never reached 0.02% total pop deaths
Vasterbotten
Blekinge splutters along but has a big crash once 0.02-0.04% region is reached. Least two weeks omitted so decline is not reporting lag
Jonkoping messy but still peaks in 0.02%-0.04% region
Uppsala doesn't quite fit, it starts long decline earlier with lower peaks later. But lots of negative excess death in mid march (covid peak?)
Kronoberg peaks just after 0.04% reached, but this region had lots of negative excess death during period and earlier in year
Norbotten mess,y peaks earlier then nada, then small spike. but again raw data messy
Vastanorrland another one, peak just after 0.04%, but again a messy region with no curve, lots of negative excess death
Just to show, places which do fit 0.02-0.04% pop death saturation pattern, raw data not noisey
Gavleborg, another one no curve, messy raw data
Halland not sure if covid ever took off proper
Ah finally forgot Varmland a semi fit - again messy data no proper curve
