You're basing your curve off faulty data. We haven't been doing testing so we don't really know. Until a week ago no one was doing anything about it and staying home like they are now. Again that will vastly throw off your curve. It's still too early. Give it a week and re-run your model.