Revising models due to changing reality or just bad models?

I keep harping on the bad math within the COVID-19 reporting. But I guess it is the world in which we live. I do not blame the Mathematicians. In back offices, they are trying to build models to predict outcomes. The models are only as good as the data they are fed, and the current data is lacking. Those of us in the industry understand that. But we then share that model to our bosses, who then play a game of telephone with the data, then hand a talking point to the media [not known for their math skills] and here we have it.


This is a phrase used a lot in the data modeling , data mining industry; it fits these models today. The models are improving, but what does it really mean when the numbers are “revised down”..again and again?

Does it mean that the reality on the ground is improving? Many will make this conclusion. Jumping to: “social distancing is working.” The situation on the ground is, no doubt, helping flatten the curve which helps ensure that our hospitals are not overrun. But they were yet to be overrun, people were not dying in the streets, laying on gurneys like a MASH unit. The fact is that we are using social distancing to stabilize the system. However, social distancing does not (and is not intended to) change the outcomes, it does not stop “everyone” from eventually getting the disease. Note, if it is able to delay us enough to break the disease by seasonality, a vaccine and/or treatment options then we will have fundamentally altered the mortality, but we are not there.

So if “social distancing” is not the answer to dropping mortality… I offer an alternative and math based answer. The GARBAGE IN that we are feeding the model is simply improving. More and more data, more and more cleaned and refined data from a larger population is improving the model(s). Now my pet peeve…the mortality rate is likely NOT CHANGING at all. Only our ability to model the reality of mortality is changing. Put another way, our confidence in the model to predict reality is improving. We do not yet have a reality on which to test this model and prove accuracy. What I have always loved about math and science is that there is a right answer. 2+2 IS 4. I can prove it. Math without reality is simply opinion and conjecture.

Update 4/13: [Bill Bennett – RealClear Politics]
Our officials and media have warned us of 2 million deaths in the United States.  Then 200,000 deaths.  Then 100,000 to 240,000.  This needs to stop. There have been a total of 68,000 coronavirus deaths worldwide. 


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: