The Probability Distribution of P-Values

On his facebook page, Nassim recently posted links to a new short technical paper on the probability distribution of p-values and a video commentary. He wrote:

I was able to pull out the exact meta-distribution of p-values (i.e. p-values as random variables).
The point is that the same phenomenon will produce p-values all over the map.  A true p-value of .12 will produce p-values <.05 more than half the time, so people may never replicate and get the same result.
One Hundred Years of P-Value Bullshit!

Here is the text of the paper, which was originally posted on his website, Fooled By Randomness.

Tail Risk Measurement Heuristics

Nassim kicks off The Bank of England’s One Bank Flagship Seminar, the first such seminar offered by the bank in an effort at greater transparency:

The first part of this talk – The Law of Large Numbers in the Real World – presents fat tails, defines them, and shows how the conventional statistics fail to operate in the real world, particularly with econometric variables, for two main reasons: 1) we need a lot, a lot more data for fat tails; and 2) we are going about estimators the wrong way. The second part – Detecting Fragility – presents heuristics to detect fragility in portfolios. Fragility is shown to be ‘anything that is harmed by volatility’. The good news is that while (tail) risk is not measurable, fragility is.