-
Well that conclusion is a mathematical modelling so its not set in stone i guess.
But i think its somewhere along the lines of: take the median date of the tests that has now been analysed. Now go back say 2 weeks (the approx time for antibodies to be developed) Now you have an approximate time frame of when 7.3% of the population were infected. From here look at the rate of the spread from this date forward and calculate what this should amount to today accounting for lowering of the spread in recent weeks and whatnots.. This is my understanding of how those numbers are produced.
Stockholm only at 7.3% with antibodies according to The preliminary results which would indicate that around 20% of the ppl of Stockholm would have had the virus today. A substantially lower figure that one would have hoped for unfortunately.