The only thing that graph tells me is your prediction model was crap from the git-go. I told you that extrapolation was not proper modeling two months ago, but now you say we were saved by lockdown. What a tool.
And Sweden was saved by not having a lockdown...
My "model" utilized actual case data and the mathematical formula of case growth derived from the case data. The mathematical formula closely matched case growth until April 2. After April 2, the growth of new cases slowed, so the mathematical formula tells me what the cases numbers would look like without the quarantines/lockdowns/etc. The quarantines, etc., were put into place during the last two weeks of March. We see a slowing of the growth of new cases after an incubation period passed.
Covid-19 spread at the rate of each person infected 2.5 others within a period of 9 days. The mathematics of this is pretty straightforward.