I calculated and plotted new deaths and the respective "growth rates" starting on March 14 and ending March 28. I also used the linear trend function. As you can see, the trend in growth while erratic, is UP. If it should continue, by April 21 it will be at 1.45.
And next day prediction is pretty useless - the variance is too great.
the "absolute worst case" being as unlikely as the sun burning out on the same day.
Yes, sure, but the point you CONTINUE to miss is, if the death growth rate (an epidemiological rather than only mathematical concept) continues to go up like this, it DOES mean that by April 21st, if the proportionality to the infected cohort holds, we WILL run out of new people to infect which is the MAIN point I'm making.
The growth rate going up to 1.45 only reinforces MY point: the peak in this case will happen even sooner because it would mean that a highly contagious virus would keep pushing up and up even faster until it reached the whole population (which coronaviruses CAN do, like I said - one of the representatives of this family of viruses is the common cold and it gets pretty much every American except the most secluded ones - it wouldn't get a bubble boy - two to five times per year).
"The next day prediction is pretty useless" - again, it is routinely used in epidemiology. But yes, it varies up and down especially regarding reporting... like I said, on Sundays you'll see a smaller number simply because reporting organizations are closed... you may catch up again on Mondays, and there are other intervening factors, like for example the virus gets to a retirement home complex and ravages through it getting a bunch of senior citizens, or gets a peak like in New Orleans given the 1.4 million tourists that visited on Mardi Gras and infected the locals, so 5 weeks later which is about how long it takes for the virus to kill, you may see a cluster of deaths.
But that's why in another thread / post I said if the death rate growth average for two consecutive 5-day periods remains 1.0, then you may realize that it's likely that you got to the inflection point, which is the mid-point of the curve and the peak of the epidemic; from that point on, the trend flattens and reverses and the number of new deaths starts to shrink day to day.
It is quite incredible, sorry for shouting but it's not just shouting but just highlighting: YOU CONTINUE TO MISINTERPRET ME THAT I'M MAKING A PREDICTION THAT WE'LL HAVE 4 MILLION DEATHS ON APRIL 21ST WHEN IT'S NOT WHAT I'M SAYING. I'm not saying that it is likely that we'll have the worst case scenario. I'm saying that WE CAN'T HAVE ANYTHING WORSE THAN THE WORST CASE SCENARIO, BY DEFINITION, AND WITH THIS PROGRESSION IF UNCHANGED, THE WORST CASE SCENARIO WOULD MANIFEST BY APRIL 21ST.
Anything that makes it less than the worst case scenario comes with the curve FLATTENING EARLIER which still makes my main point hold: that we will necessarily attain peak BY April 21st, that means, maybe earlier (which would mean we'd have fewer deaths than that), maybe then, but NO LATER than April 21st. Jesus!!!
This is due to the simplest of facts... because there are only three possibilities: either the progression doesn't slow down (due to the aggressive virological characteristics of this virus that is highly contagious and entered America by multiple entry points in all 50 states), or the virus is SO aggressive that the progression actually accelerates, or it does slow down given the caveats I've mentioned in post #2 (and also in post #1, the herd immunity phenomenon).
So, let's regroup and think.
Hypothesis # 1: the progression doesn't slow down. In this case we do reach saturation of the American population with 100% infected by April 21st therefore the peak happens by then because the virus has no one else left to infect. QED.
Hypothesis # 2: the progression even accelerates. In this case we do reach saturation even sooner than April 21st therefore my prediction that it will happen BY then still holds. QED
Hypothesis # 3 (the most likely one): the progression does decelerate, like is usually the case for all pandemics so that we never get to the absolute huge numbers of April 21st. That means by definition that we will have passed the peak, because epidemiologically speaking, a deceleration of the death rate growth below 1.0 *IS* the inflection point that marks the peak. Therefore, my hypothesis that the peak will come before April 21st still holds. QED.
I hope that NOW you will get it.
The thing is, like I said, let's just wait for April 21st. That day, we'll come back to this thread, you'll see that the Death Growth Rate will be necessarily less than 1.0, therefore you'll see that I was right, at which point I'll hope for an apology for having called my thinking "insane" and "whack-a-doodle."
Continues below in two posts