This is why statistics are important
How Panic-Buying Revealed the Problem With the Modern World
The false "efficiency" they talk about is due to misuse of statics. they are planning for "average" usage.
Trouble with that is that it means half the time they have enough to cover usage. and half the time they *don't*.
What they need to do is use z scores. Using those you can take the data and determine what levels will cover 50%, 80%, 90% etc.
For stores 80-90% ought to be ok. for medical and other ctritical services, 99% or higher would be better.
Incidentally, this is the kind out stuff that's done to come up with "10-year", "100-year" etc values for storms and floods.
Which brings up another factor. You have to *keep collecting* data and revise values as usage changes. That's why the estimates for storms and floods aren't doing so good these days.
This entry was originally posted at https://kengr.dreamwidth.org/1112016.html. Please comment there using OpenID.