ClaimsNassim Taleb

“As illustrated in Fig. 1, The error in the evaluation of the”

Taleb and co-authors, paper authors 2023 The Probability Conflation: A Reply (arXiv) archived copy

Factual Science and research methods

Error in the evaluation of the probability PK can translate into an explosive error in the corresponding threshold K, more so the fatter-tailed the distribution.

Unresolvable The public record could not settle it.

Factual: Statements about the present or the past, checked against the record: did they hold up?

The second judge said resolved true; the verdict shown is the first judge's.

What would count as true or false

Judge as of 2023 against the paper's own mathematics and figure, and standard extreme-value statistics. True: the paper shows, analytically or in Fig. 1, that a small error in the tail probability P_K produces a disproportionately large (rapidly growing) error in the threshold K, and that the effect grows with tail fatness (e.g. a lower tail exponent), consistent with the inverse survival function of power laws. Mostly true: the result holds, but 'explosive' slightly overstates the magnitude in some ranges. Mixed: the result holds only for some distributions or parameter ranges. Mostly false: the error grows, but not more for fatter tails, or only trivially. False: the mathematics contradicts the claim.

Set before the research began.

Evidence (8)

  1. E1 Peter Carr Brooklyn Quant Experience (BQE) Lecture Series: Nassim Nicholas Taleb NYU Tandon School of Engineering · 5 October 2023

    “In a survival function, under fat tails, errors on probability translate into disproportionately large, possibly infinite, errors on threshold.”
  2. E2 Black Swans and the Domains of Statistics The American Statistician (PDF hosted by risknet.de) · 2007 Own record

    “Yetinthesedomains,thesmallertheprobability,themoreconsequentialtheimpactoftheabsoluteprobabilityerroronthemomentsofthedistribution.”
  3. E3 On the Super-Additivity and Estimation Biases of Quantile Contributions arXiv · 11 November 2014 Own record

    “such bias is in proportion to the fatness of tails and, for very fat tailed distributions, remains significant, even for very large samples.”
  4. E4 On the Super-Additivity and Estimation Biases of Quantile Contributions arXiv (Taleb and Douady) · 11 November 2014 Own record

    “Under these assumptions, the cutpoint of exceedance is \(h(q)=x_{\min}\,q^{-1/\alpha}\) and we have:”
  5. E5 On the Super-Additivity and Estimation Biases of Quantile Contributions arXiv (Taleb and Douady) · 11 November 2014 Own record

    “To show how unreliable the measures of inequality concentration from quantiles, consider that a standard error of 0.3 in the measurement of \(\alpha\) causes \(\kappa_{q}(\alpha)\) to rise by 0.25.”
  6. E6 [PDF] Convexity, Robustness, and Model Error inside the Fourth Quadrant fooledbyrandomness.com (N. N. Taleb's site) · July 2010 Own record

    “Model error increases small probabilities in a disproportionate way (and accordingly decreases large probabilities).”
  7. E7 [PDF] Convexity, Robustness, and Model Error inside the Fourth Quadrant fooledbyrandomness.com (N. N. Taleb's site) · July 2010 Own record

    “Such explosive uncertainty isn't the result of fat tails in the distribution, merely small imprecision about a future”
  8. E8 What You See and What You Don't See: The Hidden Moments of a Probability Distribution ar5iv (arXiv HTML rendering) · April 2020 Own record

    “Exceedance probability does not depend on the thickness of the tails.”

Reasoning

An independent event description [E1] says that under fat tails, errors on probability turn into disproportionately large errors on the threshold, and earlier work by the subject gives the power-law threshold formula [E4]. However, Fig. 1 and the paper's own derivation are not in the evidence, and the figure used to grade the result as "mostly true" does not appear in any evidence item. The evidence therefore does not settle the claim under the criteria.

Cite this claim

Nassim Taleb, “Error in the evaluation of the probability PK can translate into an explosive error in the corresponding threshold K, more so the fatter-tailed the distribution.” (said 2023). Verdict: unresolvable. Reliability API, claim c_4ca214a746b0, method 0.1.0+28da144b, dataset release v1.1.2. https://epistacklabs.com/claims/c_4ca214a746b0

This claim is in dataset release v1.1.2, with the same verdict.

Reply to this claim

For Nassim Taleb, or someone speaking for them: a correction or counter-evidence. Replies are read before anything is shown; an approved reply shows your name, role, date, statement and link. Your email address is never shown and is deleted 30 days after the decision (privacy).

More factual claims by Nassim Taleb

  1. Mixed Factual 24 September 2025
    “entities grow in a convex way, then slow as they saturate — growth may be unbounded, but remains sub-logarithmic.”

    Entities in biology and economics grow convexly then slow as they saturate; growth may be unbounded but remains sub-logarithmic.

  2. Mixed Factual 24 September 2025
    “the unavoidable conclusion is that cardio problems are at least half genetic.”

    If ancient populations such as hunter-gatherers and Inuit had a high incidence of cardiovascular disease, the unavoidable conclusion is that cardiovascular problems are at least half genetic.

  3. Mixed Factual 19 September 2025
    “scientists that come up with new ideas & frameworks are more often the experienced ones, not the young as we tend to believe.”

    Scientists who come up with new ideas and frameworks are more often experienced researchers than young ones.

  4. Mostly true Factual 2023
    “And if one does not hedge the other, then being "good at”

    If tail probability PK does not hedge tail expectation GK, then skill at predicting PK provides no information on GK.

  5. Mostly true Factual 2023
    “A sum of such bets becomes rapidly”

    A sum of such bets, whose probabilities are modelled with a Uniform or Beta distribution, becomes Gaussian rapidly.

All of them

Verdicts are made automatically; how verdicts are made.