Models are maps. Not absolute predictions of the future.

We are heading out on a very long trip.
We’re not sure what the weather is going to be exactly.
We don’t know if the roads will be in good condition.
We don’t know what the other drivers on the road might do.
1/4

We don’t know how tired we are going to be and/or get distracted and frustrated with the kids in the car.
We are hoping that our plans are well founded.
We’ve only done this trip or something like it once, and it was a bit different.
So we will get some things wrong.

2/4

We use probabilities to make some educated guesses about what may happen.
We can be reasonably sure about tomorrow, but two months away is a bit more complex.
We run through a whole bunch of scenarios, but we aren’t completely sure which one it will be.
#COVIDVIC19
3/4

But we (incredibly flawed humans) might be distracted by patterns that suit our particular biases, by hopes that things will go better than expected, and by limited access to detailed data.

So models are maps. Not absolute predictions of the future.

4/4