So I made my own model

There's been a lot of mention of computer models concerning Covid-19.  The models predict total deaths in various countries as well as total cases.  These models have varied widely.  And by "varied widely," I mean they have fluctuated to the point of being essentially useless.  Also, there hasn't been full disclosure.  If I can't see all the code and all the input data, then I can't judge a model on its merits.  This isn't science.  It's digital card tricks in the dark.

So I made my own model.

Now, it's admittedly a simplistic model.  Also, I'm not going to claim my model is any better than anyone else's.  That remains to be seen.  But I will claim this: all is revealed.  No secrets.  I'm going to lay out the whole methodology so that the logic and data can be properly judged.  I'm not going to demand that people put their faith in it as if it's the programming version of the Gospel or something.

So let's begin...

My focus is American deaths only.  My model predicts about 102,000 of them.  That's total over the course of the plague.

I didn't arrive at this figure by extrapolating from other countries' data.  That's apples-to-oranges stuff.  Different cultures, different gene pool, different climates, different laws, etc.  If I'm going to extrapolate, then I'm going to do it America-to-America.  I'm going to compare the current outbreak to the Spanish Flu of 1918.

Of course, the country was a very different place back then.  Differences in sanitation, nutrition, transportation, medicine, etc.  Lots of variables.

In addition, severity of the current infection varies by ethnicity, age, location, underlying health factors, and so on and so forth.  Every new variable introduces an additional margin of error.  Many variables means many margins of error, and those quickly add up to the point where the whole exercise is rendered meaningless. 

What we need is a single conversion factor that acts as a reasonable proxy for all those other variables.

I've chosen infant mortality as my conversion factor.

Infants don't smoke or drink or engage in reckless behavior, so their mortality rate is somewhat standardized, or at least it's as standardized as humans get.  It encompasses nutrition, sanitation, medicine, etc.  Comparing the infant mortality rate in 1917--the year before the Spanish Flu hit--to the rate in 2017--the latest for which I could find numbers--should act as a reasonable proxy for all the differences between 1917 America and 2017 America.

The infant mortality rate in 1917 was somewhere between 100 and 109, depending on which source you use.  I decided to split the difference and go with 105.  That's 105 deaths per thousand live births.

The rate in 2017 was 5.8.  Dividing the latter by the former gives us a conversion factor of 0.05523810.  That means an infant was almost twenty times more likely to die in 1917 than in 2017.  Obviously, we've come a long way.  People who compare the current virus to the 1918 one without adjusting for the last century's worth of improvements in general health are doing it all wrong.

The Spanish Flu, according to the Wikipedia page, killed somewhere between 0.48% and 0.64% of the population.  Again, I'll split the difference and go with 0.56%.

We now multiply that percentage by our conversion factor in order to estimate what percentage of the people the Spanish Flu would kill if it happened today.  That gives us 0.030933%.

We'll assume that Covid-19 is just as powerful as the Spanish Flu.  This may or may not be a good assumption, but we'll go with it as a "worst case" scenario.  That means we can expect Covid-19 to kill off 0.030933% of the population.  In a nation of 330 million, that means 102,080 deaths.

So there you go.  It's a simplistic model, and I've probably missed some obvious-in-hindsight logical error, but I doubt it's any worse than any of the other models out there.  And I've fully disclosed my inputs, assumptions, and methodology, so I'm not just engaging in digital magic tricks.  All my cards are on the table.

Time will tell if this model is worth a hill of beans or not.

In the meantime, it's something else to discuss.  If it makes someone pause and think a little, then that's enough for me.  Because this thing has become a mania, and we could stand to do a lot more thinking and a lot less knee-jerk reacting.

Stay safe, folks, and take reasonable precautions, but stay calm, too.  Panicking doesn't help anyone, and the odds of survival are highly in your favor.  We'll make it through this so long as we don't murder each other and burn the whole economy down along the way.