Covid Pause – Minus 32 Days and Counting (Part III) (128,221)

This will be short because I am kind of worn out after I created it.

Also, the IHME model has been updated. Son tomorrow I’ll look at their new data and see how it is different.

So what you see there is the minimum, maximum, and average predictions for cumulative mortality, for each state, on this day, according to the IHME model that was published on 3/26.

I looked at the actual cumulative mortality and noted where it fit in their spread, and gave it a score:

• Negative numbers under the average model line (with -1 at the minimum line and larger negative numbers under the minimum
• Positive numbers over the average model line (with +1 at the maximum line and larger positive numbers over the maximum

I then took a weighted average of the most recent 4 scores (40% for today’s, 30% for yesterday’s, 20% for 2 days ago, 10% for 3 days ago.

Based on that weighted average, I projected the number of deaths in the future. A negative average means fewer deaths in the future and and earlier and smaller apex (maximum deaths per day). A positive average means more deaths and a later and more severe apex.

Here are 3 hotspot states I am looking at now:

According to the 3/26 model, New York State was going to hit peak death on 4/8 with 547, on average. The minimum was 4/5 and 317; the maximum 4/13 and 1,357.

Right now they are near the maximum extreme of the 3/26 model. I have the peak at 4/12, with 1,229 lost that day.

In the end they are projected to have 24.666 deaths, close to the 26,983 maximum in the 3/26 model. Their average was 10,243 and minimum was 4,911.

From the 3/30 model that I have not yet downloaded the data for:

Looks like a tighter prediction.

Again, it’s tighter. An 8,000 to 22,000 spread. So my estimate right now is above their maximum in the 3/30 model. That’s good news—or less bad, anyway.

Here’s the Michigan trend lines:

Michigan slightly above the 3/26 average.

I’m overestimating based on the 3/30 model. 😀

Louisiana:

Also slightly above average.

And also overestimating, based on the 3/30 model.

I expect the prediction for the US to come down a lot after I fit my scheme to the new model data.

But this all depends on massive social distancing for the next month. The whole county is expected to pause and shelter in place, so that the bugs get resolved in their current hosts and don’t get to find new homes.

We are supposed to get updated federal guidelines tomorrow. We need to eventually have some rules in place that are easily followed, and will do as much as possible to prevent the bugs from multiplying our of control before we can quarantine anyone who has them.

The 6’ spacing needs to go eventually. Too many businesses can not operate under conditions like that. This is why I feel it is a good investment to wait until the end of April to end this—in return, we can be allowed to get slightly closer, and maybe just wave to each other for the summer.

The Abbott tests seem promising—if they work, they can isolate any positives quickly before a community spread situation occurs again. That was the best new of the day today (other than the Comfort arriving in NYC to care for non-Covid patients).