Scores by kind of claim
Each kind of claim is scored on its own, out of 100: the share that held up, with the range it likely falls in. Fewer resolved claims give a wider range.
-
Factual claims
Statements about the present or the past, checked against the record: did they hold up?
73 67 79 -
Predictions
Statements about events outside the person's control, checked once they fell due: did they come true?
Insufficient record: 0 predictions resolved, 5 needed for a score.
-
Promises
Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?
Insufficient record: 0 promises resolved, 5 needed for a score.
All types combined: 73 (likely 67 to 79, high confidence), from 94 claims: 94 factual claims. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
Nassim Taleb's claims
26 claims
“PK corresponds to the zeroeth moment and is always”
Taleb and co-authors, paper authors 2023 The Probability Conflation: A Reply (arXiv)
The exceedance probability PK corresponds to the zeroeth moment and is always thin-tailed.
Resolved true The record shows it happened or was so.
The second judge would not resolve this claim; the verdict shown is the first judge's.
What would count as true or false
This is a mathematical claim. Judge it by standard probability theory. TRUE: two things hold. First, the exceedance probability P(X>K) equals the expectation of the indicator 1{X>K}, which is a zeroth (partial) moment. Second, because that indicator is bounded between 0 and 1, the indicator and its sample estimator have finite moments of all orders and thin-tailed (for example binomial or Gaussian-limit) behaviour, whatever the distribution of X. MOSTLY TRUE: the statement holds under the usual reading, with only a minor terminological imprecision, such as calling it a "partial" moment versus a "zeroeth" moment. MOSTLY FALSE: the statement holds only under extra conditions on X that the claim leaves out. FALSE: the exceedance probability or its estimator can be fat-tailed or have infinite moments in general, or it does not correspond to a zeroth moment.
Evidence
- T H E T E C HN I C A L I N C E RTO COL L EC TION arXiv (PDF of Nassim Nicholas Taleb, Statistical Consequences of Fat Tails) · 2020 Own record
“this distribution-independent convergence concerns probabilities of course, not moments -a result this author has worked on and generalized for the "hidden moment" above the maximum.”
- T H E T E C HN I C A L I N C E RTO COL L EC TION arXiv (PDF of Nassim Nicholas Taleb, Statistical Consequences of Fat Tails) · 2020 Own record
“By Glivenko-Cantelli the distribution of $\mu_{K,0}$ should be independent of the initial distribution of X, but higher moments do not, hence there is a bit of a problem with Kolmogorov-Smirnoff-style tests.”
- On the Super-Additivity and Estimation Biases of Quantile Contributions arXiv · 11 November 2014 Own record
“and such bias is in proportion to the fatness of tails and, for very fat tailed distributions, remains significant, even for very large samples.”
- Reducing bias and MSE in estimation of heavy tails:a Bayesian approach ar5iv (arXiv HTML rendering)
“Hence the asymptotic MSE behaviour for the tail probability estimator has the same characteristics as the tail index estimator.”
Reasoning
The criteria say to judge this by standard probability theory. Under it, P(X>K) = E[1{X>K}], which is a zeroth partial moment. The indicator is bounded between 0 and 1, so it and its sample mean have finite moments of every order and binomial or Gaussian-limit behaviour, whatever the distribution of X. E2 and E1 agree: they describe the zeroth-order part as distribution-independent (Glivenko-Cantelli) while higher moments are not. E3 and E4 do not contradict this, because they concern different estimators: concentration ratios and parametric tail-index-based tail probabilities, not the empirical exceedance frequency.
Cite this claim
Nassim Taleb, “The exceedance probability PK corresponds to the zeroeth moment and is always thin-tailed.” (said 2023). Verdict: resolved true. Reliability API, claim c_18973addda9e, method 0.1.0+28da144b. https://epistacklabs.com/claims/c_18973addda9e
Claims without a verdict
Is a verdict here wrong, or do you speak for Nassim Taleb? Reply on the claim's own page, with the form at its foot (how replies work). Verdicts are made automatically; how verdicts and scores are made.