Forecasting with fat tailed distro

Random variables in the power law class with tail exponent α ≤ 1 are, simply, not forecastable. They do not obey the LLN. But we can still understand their properties.”

“If one claims fitness or nonfitness of a forecasting ability based on a single observation = 1, she or he would be deemed to be making an unscientific claim. For fat tailed variables that “= 1″ error can be made with n = 106. In the case of pandemics, n = ∞ is still anecdotal. (..) It takes 1014 observations for a “Pareto 80/20” (the most commonly referred to probability distribution, that is withα ≈ 1.13) for the average thus obtained to emulate the significance of a Gaussian with only 30 observations.”

Taleb justifying the precautionary principle on statistcis ground, in a challenge with Ioannidis who called for reopening the economy based on poor data


Taleb and coronavirus

the precautionary principle applies when traditional statistical averages are invalid because risks are not ergodic”

“Practically all the uncertainty tends to make the problem potentially worse, not better, as these processes are convex to uncertainty.”

The Only Man Who Has A Clue

From Trolley to Mechanical Ventilator

how “problems” change over time, one day everybody is musing on a robot car who (that) must choose who to kill, then a small rna virus arrives and everybody turns to the problem of mechanical ventilators, except that not, nobody wants to acknowledge that in the Covid19 epidemics thousand of critical patients are subject to rationing life/death decisions based of a simple rule or a more complex clinical evalution algorithm? And the dilemma goes: you save a 60-yearld old man in perfect shape of a 35-year old pregnant woman with breast cancer ?

Facing Covid-19 in Italy — Ethics, Logistics, and Therapeutics on the Epidemic’s Front Line

New England Journal of Medicine has a retrospective on the ICU crisis in Italy where decisions around scarce mechanical ventilators are actually done, at time of writing over 4,000 people dead and this charts show the problem in stricking clarity