Thread by @doctorow

In #XKCD 2295, "Garbage Math," Randall Munroe does a stellar job elucidating the uncertainty plaguing our ability to estimate when lockdown will end.

xkcd.com/2295/

1/

The thing is, several key variables are unknown and attempts to estimate them produce "garbage numbers." These include:

* What percentage of people have covid-19 antibodies

* What percentage of people who have covid-19 antibodies are immune to the disease, and for how long

2/

* When we will develop a vaccine for covid-19

* How long production of the vaccine will take after it is developed

* When we will discover effective therapeutics for covid-19

* How long it will take to ramp up their production

3/

These variables produce a combinatorial explosion of scenarios for when the lockdown will end, and how long after that the crisis will pass. These scenarios are the reason we have a range of "estimates" - they're not estimates, they're scenarios.

4/

If a and b, then x more weeks. If a, b and c, then y more weeks. If a and c, etc, etc.

These kinds of scenarios are a lot more common than predictions, but scenarios are usually presented (by governments, press and science communicators) as predictions.

5/

In part, that's because our public discourse has little room for scenarios - but also, that's partly WHY our public discourse has little room for scenarios.

eof/