So now that we know true infection rates were 30x higher than detected in Wuhan, we can somewhat predict the course in NYC, though they are presumably testing more widely there. Assuming these things:
) New Yorkers won't wash their hands or self-quarantine. Fuggedaboutit!
) therefore it will take ~40% infection rate for R to drop below 1
) they have ~20x higher infection rate than tested
Then with 15,000 confirmed cases, the real number is 300,000, or already 3.75% of their 8 million population. This should already be enough to reduce the growth rate by reducing R, which could show up within 5 days or so in the testing numbers if they correlate well with the trends in actual infections. To reach the ~3 million cases needed for R < 1 and cause an actual reduction in new cases, infection rates need to increase by a factor of 10. For average growth rate scenarios this would take
40%: 7 days
30%: 9 days
20%: 13 days
15%: 17 days
10%: 25 days
Add a few days for the R <1 to play out in the real world, and that's about how long NYC has before their charts inexorably trend very quickly to 0.
This illustrates how slowing the growth rate drastically prolongs the time for the epidemic to run its course. Presumably the current 125 deaths per 300,000 cases will hold steady for a total of 1125 deaths. I'm sure NYC can handle that within 25 days, but I'm also pretty sure they could handle it within 9 days.
This shows the inherent foolishness of the panicky Flatteners: by assuming the worst case in every instance, they're preventing faster resolution of the problem in most cases. Economically, the longer it takes, the worse things get, and that costs lives too. The strategy should have been, right from the start, to quarantine the vulnerable, which has little economic impact. Then allow the virus to spread as quickly as it could among the rest of the population. This could have been achieved very quickly simply by not closing the schools. It's been handled wrong from day 1.