Is the predictive model wrong?

[Copper]

Impeach him!

[jon-nyc]

The more I look into this model the less I like it. It strikes me as more of a ‘technical analysis’ than a ‘fundamental analysis’ in that it’s basically a curve fitting exercise heavily relying on the Wuhan experience.

It seems to be predicting way too early of a peak and is overall too optimistic about our social-distancing measures.

I would love to see the details behind the McKinsey models.

[George K]

@jon-nyc yes.

And the simple fact that its accuracy, in historical perspective, is pretty poor.

[LuFins Dad]

@jon-nyc Which model? The covid19.healthdata.org model? I’ve been calling that model crap for the opposite reason. It undervalues decisions made by the general public and overvalues decisions handed down by the government.

[jon-nyc]

Yes that model. That’s not really the opposite of my view, I also think it overstates the effect of our lockdown.

[LuFins Dad]

@jon-nyc do you have a link to the McKinsey models?

[jon-nyc]

They’re not public to my knowledge.

[George K]

[LuFins Dad]

I’d like to see the ventilator numbers. And Florida is way too early to take anything from a 1 day aberration

[George K]

@LuFins-Dad said in Is the predictive model wrong?:

Florida is way too early to take anything from a 1 day aberration

Absolutely. We've seen 1 day dips before.