Tag Archives: P-Values

P Values, Deadlift, Precautionary Principle, Trump, Virtue, Kapish?

@nntaleb In high frequ quant finance we use >10^6 data points & nobody knows pvalue. Psychologists trying to get away with 10https://twitter.com/nntaleb/status/717366620331618304   Permalink 7:02 PM – 6 Apr 2016

@nntaleb Normal humans can learn from Iraq and Lybia. Journalism became the last refuge for those with brain defects.https://twitter.com/maxabrahms/status/717880860066856960   Permalink 6:56 PM – 6 Apr 2016

@nntaleb My Skin in the Game WIPhttps://twitter.com/evonomicsmag/status/717556674559668224   Permalink 4:03 AM – 6 Apr 2016

@nntaleb Por primera vez, texto en español [Bogota]https://twitter.com/luiscano/status/717442396473135104   Permalink 2:54 PM – 5 Apr 2016

@nntaleb To paraphrase Yogi Bera, 90% of deadlifting is mental; the other half is physical. Permalink 8:13 AM – 5 Apr 2016

@nntaleb Meanwhile a 80% power for a test and 5% p-value aren’t science.Pick a large n & fughedaoudit http://fooledbyrandomness.com/pvalues.pdf  pic.twitter.com/SUMZ6zCBeA Permalink 8:01 AM – 5 Apr 2016

@nntaleb Instead of risking Max deadlift, easier to figure it out from max{lbs: 3 reps <=failure}. @Pete_Spence @georgepearkes Permalink 6:43 AM – 5 Apr 2016

@nntaleb in lbs of course. Permalink 6:32 AM – 5 Apr 2016

@nntaleb What is the max deadlift for a noncompetitive flaneur like me (adj. by age group) that would be considered decent? Permalink 6:29 AM – 5 Apr 2016

@esalazar_y There are only two people missing #Seneca and #FatTony @nntalebpic.twitter.com/aojYNa4QNk Permalink 6:37 PM – 4 Apr 2016

@nntaleb The Precautionary Principle Simplified: Started rewriting @yaneerbaryam @normonics @RupertReadpic.twitter.com/i12VkJn8WC Permalink 11:40 AM – 4 Apr 2016

@kupiako @nntaleb @DontFadeeMeBro And his father died in 1999 Permalink 7:28 AM – 4 Apr 2016

@nntaleb Main diff betw. Trump and Fat Tony: 1) His FatToniness doesn’t like publicity, 2) Fat Tony wouldn’t marry a model.https://twitter.com/UncleDario1/status/716995760181616641   Permalink 7:30 AM – 4 Apr 2016

@nntaleb Also Trump had 4 siblings. @DontFadeeMeBro Permalink 7:26 AM – 4 Apr 2016

@nntaleb I only trust people who don’t care about their reputation. Virtue ethics is about acts of virtue even if they come with negative reward. Permalink 7:25 AM – 4 Apr 2016

@nntaleb That is a myth about Trump’s inheritance. Work the numbers of cohort of others who have inherited.https://twitter.com/gweirydd/status/716993004096077824   Permalink 7:20 AM – 4 Apr 2016

@nntaleb Trump: The Art of the Deal: Rediscovers scaling laws, how fragility increases with size.pic.twitter.com/55uXlT5XwD Permalink 7:15 AM – 4 Apr 2016

@nntaleb Trump: The Art of the Deal (p. 305): Politicians sh**t scared of the press.pic.twitter.com/ZwM5grWlsV Permalink 7:11 AM – 4 Apr 2016

@nntaleb Trump: The Art of the Deal (pp. 282-283): How consultants & executives don’t hold a candle to someone w/ #SITGpic.twitter.com/24nofJ5AcN Permalink 7:08 AM – 4 Apr 2016

@nntaleb “To me, committees are what insecure people create in order to put off making hard decisions.” Trump: The Art of the Deal (p. 281) Permalink 7:05 AM – 4 Apr 2016

@nntaleb My hobby is busting bullshitters @iandavidmorris, a la Feynman: “Science is the belief in the ignorance of experts”https://twitter.com/Priapus_D/status/716972631925387264   Permalink 6:04 AM – 4 Apr 2016

@nntaleb The fellow did not break the truth: comments in prnthss which is NOT standard but NOT unethical. Kapish? @semperfidem2004 @iandavidmorris Permalink 3:04 PM – 3 Apr 2016

@nntaleb In real science it is not about whether the scientist has the right standards; it is about his THEORY. Kapish? @homophonous @iandavidmorris Permalink 2:19 PM – 3 Apr 2016

@nntaleb 2) which is why Lux.’s theory is a valid hunch; disproofs are unrigorous w/absence of evidence fallacies. @joebradfordnet @iandavidmorris Permalink 2:08 PM – 3 Apr 2016

@nntaleb 1/Every Arabic speaker who had to memorize poetry knows that pre-Islamic vocabulary is v. different from Q. @joebradfordnet @iandavidmorris Permalink 2:05 PM – 3 Apr 2016

@republicofmath MT @MathFour Engineers think equations approx reality. Physicists think reality approx equations. Mathematicians know equations are reality Permalink 11:31 AM – 3 Apr 2016

@nntaleb 2) So in spite of his solecism I find him closer to the SUBSTANCE of the subject matter than your elaborate nitpicking @iandavidmorris Permalink 12:40 PM – 3 Apr 2016

@nntaleb 1) As I said he knew enough Syriac to have detected the problem and confused Payne Smith for his daughter’s translation. @iandavidmorris Permalink 12:38 PM – 3 Apr 2016

@nntaleb His “sparkle” is in Payne. Check again. Attacking someone’s thesis on his etiquette is failing us: we care about the truth @iandavidmorris Permalink 12:20 PM – 3 Apr 2016

@nntaleb Finally, in my world (science, search for truth), scholarship is about SUBSTANCE, not trifles. @iandavidmorris Permalink 12:16 PM – 3 Apr 2016

@nntaleb What makes me doubt your scholarship is that the fact that you did not check if the language was from Payne, not his own. @iandavidmorris Permalink 12:15 PM – 3 Apr 2016

@nntaleb You are conflating breach of ethics & breach of the etiquette of a research milieu. By nitpicking, you reinforced his thesis.@iandavidmorris Permalink 11:56 AM – 3 Apr 2016

@nntaleb 2/2 If it weren’t for Obama, Shillary Clinton wd have transformed Damascus into an AlQaeda or ISIS Base. Keep her on the lecture circuit. Permalink 9:33 AM – 3 Apr 2016

@nntaleb We need to keep Shillary out of the White House. Whatever it takes, even if it means having a robot as president.https://twitter.com/MaxAbrahms/status/716644245977899008   Permalink 9:24 AM – 3 Apr 2016

@nntaleb Summary: Anything that has an infinite (or “long”) shelf life (i.e. “sustainable”) falls under Precautionary Pple.pic.twitter.com/55pwCgaJM0 Permalink 6:55 AM – 3 Apr 2016

@nntaleb I think I got the p-hacking distribution & given no outcry by psych trolls it isn’t a “rediscovery”. http://fooledbyrandomness.com/pvalues.pdf   @learnfromerror Permalink 3:13 PM – 2 Apr 2016

@nntaleb Nobody can explain why the neocons are buying that Saudi Barbaria is helping Syria to become democratic. Is neocon a result of brain damage? Permalink 2:45 PM – 2 Apr 2016

@nntaleb If all progress is innovation, all innovation is not progress. As a matter of fact, 99.9% of innovations fail. @sarah_cone @uistbhoy88 Permalink 9:39 AM – 2 Apr 2016

@nntaleb The problem with architecture is that hotshot architects try to impress other architects and we are left paying the price. #noskininthegame Permalink 7:52 AM – 2 Apr 2016

@nntaleb Rainy Saturday: density of the preimage of the power of test. Concavity of measure, mean is lower than orig powerpic.twitter.com/mKd3T2CyNR Permalink 7:49 AM – 2 Apr 2016

@nntaleb Sadly, when a gentle, low-testosterone looking person promotes peace, he is much less convincing than a warrior-looking bully. Permalink 6:45 AM – 2 Apr 2016

@nntaleb The problem with architecture is that, unlike with fashion, music or cinema, fads leave a permanent scar. Permalink 4:59 AM – 2 Apr 2016

@nntaleb This summer I want to learn a new language. Which sequence is preferable? (Urdu as proxy for Hindi for script). Permalink 2:49 PM – 1 Apr 2016

@nntaleb Metadistribution: The dist of the pre-image of a power of a test. In other words, Power is as BS as P-Value.pic.twitter.com/xAGvyuqmjk Permalink 1:08 PM – 1 Apr 2016

@MaxAbrahms You know you’re a sectarian fraud when you say Assad must go for humanitarian reasons while cheering the Saudi-led carpet-bombing in Yemen. Permalink 2:24 AM – 31 Mar 2016

@nntaleb P-value is a stochastic measure. Stochastic measures are not measures. http://fooledbyrandomness.com/pvalues.pdf   @dsquareddigest Permalink 11:11 AM – 1 Apr 2016

@nntaleb Fixed! @SandrewFinancepic.twitter.com/fWj8wc5z5F Permalink 11:01 AM – 1 Apr 2016

@nntaleb Connecting the Precautionary PPle to Peters and Gell-Mann(2015); something MISSED by the idiots in decision theory.pic.twitter.com/sOplMWr1we Permalink 8:37 AM – 1 Apr 2016

@nntaleb We are reframing the MINI CERTIFICATE IN RISK TAKING as risk management is back office, risk taking is front office. http://realworldrisk.com/inaugural_program_comments   Permalink 8:11 AM – 1 Apr 2016

@garyruskin Then: #Monsanto hid the truth about PCBs http://bit.ly/1C6Gupx  . Now: Monsanto says “Roundup Is Not A Carcinogen” http://wbur.fm/1MZz35R   Permalink 6:12 AM – 1 Apr 2016

@nntaleb Summary of philosophical approach to the inverse problem (in Silent Risk).pic.twitter.com/Sw3BQt1kvW Permalink 6:05 AM – 1 Apr 2016

@nntaleb The precautionary principle papers: No, there is no *scientific* evidence “GMO” are safe http://www.fooledbyrandomness.com/PrecautionaryPrinciple.html  pic.twitter.com/d08P8gYtLr Permalink 4:25 AM – 1 Apr 2016

P-Values, Growth | Facebook

For those interested in p-values, a short (very technical) paper and a video commentary. 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!
The paper: http://fooledbyrandomness.com/pvalues.pdf

This is a companion to the short paper at http://fooledbyrandomness.com/pvalues… It explains the problem with p-values giving different values… (3/23/16)


I was at a venue where someone wanted to increase economic growth: “funding” and “investments” were mentioned every 240 characters.

Now some empiricism. Consider that almost all tech companies “in the tails” were not started by “funding”. Take companies you are familiar with: Microsoft, Apple, Google, Facebook. These companies started with risk-taking. Funding came in small amounts, way later.

Now let us take the idea to its next logical step. The minute something called a “business plan” is created, the idea is corrupt and the BS start dominating. People start thinking in terms of other people’s perception of the business, for “raising cash. Like prostitutes.

We can generalize to economic growth. The problem is that these discussions of “growth” are made by people who have never taken risks.

What is needed? 1) determination, 2) sophistication, 3) authenticity, 4) convexity and above all, 5) risk loving.

PS- Someone wanted to start an academic research center; his aim was to have a structure, a business plan, get research grants, etc. Then do research. I bailed out and did more for a total cost of $1.50, with http://www.fooledbyrandomness.com/FatTails.html  (3/21/16)

Belgium, P-Values, Nocera, Saudi PR, DARK Act, Monsanto | Twitter

@nntaleb The Saudi origins of Belgium’s Islamist threat http://wpo.st/NMlO1   Permalink 4:59 PM – 23 Mar 2016

@nntaleb There is a video see feed https://www.youtube.com/watch?v=8qrfSh07rT0   @NathanBiddle Permalink 6:47 AM – 23 Mar 2016

@nntaleb Two works in my mailbox: One Formulation of size and fragility http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2741198&download=yes   One Book on Antifragility http://link.springer.com/book/10.1007/978-3-319-30070-2   Permalink 6:13 AM – 23 Mar 2016

@nntaleb Here is the video on the metadistribution of p-values https://youtu.be/8qrfSh07rT0   and the text is here: http://fooledbyrandomness.com/pvalues.pdf   Permalink 7:55 PM – 22 Mar 2016

@davidmoore411 He’s one of few people on the internet who could insult you publicly and you’d still buy his next book. His stuff is that good. @nntaleb Permalink 6:08 PM – 22 Mar 2016

@nntaleb OK, OK, full proof of an exact metadistribution of p-values (covers all hypotheses). Shd I do a video comment? pic.twitter.com/YLuY3UTY1v Permalink 3:21 PM – 22 Mar 2016

@nntaleb Syrian Christians behave as Muslims in the East, Christians in the West, & are viewed as Muslims in the West and Christians in the East. Permalink 5:26 AM – 22 Mar 2016

@nntaleb Walid Phares is a Christian Lebanese @dandrezner Permalink 4:54 AM – 22 Mar 2016

@nntaleb What is slowing me down w #SkinintheGame is Isocrates’ Silver Rule for Nations: Treat other nations the way you want them to treat you. Permalink 3:26 PM – 21 Mar 2016

@nntaleb Q: Should we intervene more to “help” countries as we brilliantly did in Lybia, Iraq, Syria, Vietnam, Iran (1953)? https://twitter.com/simon_schama/status/712014036993499137   Permalink 2:37 PM – 21 Mar 2016

@nntaleb “Funding” & “Investments” are the last thing you need. All you need: authenticity & risk. https://www.facebook.com/nntaleb/posts/10153701042473375   pic.twitter.com/lMyA1wTQ56 Permalink 2:34 PM – 21 Mar 2016

@cdipaola62 @nntaleb @NoceraNYT Definition of Irony- A writer named “Nocera” finds discussion of Primum non Nocere, unintelligible Permalink 1:47 PM – 21 Mar 2016

@nntaleb Actually, from his name John of Damascus was not Syrian/Aramaic, but likely to be an Arab (*Yahya* Ibn Sarjun, not Yuhanan) @iandavidmorris Permalink 1:38 PM – 21 Mar 2016

@nntaleb 9 days is nothing. There is the 2005 pre-crisis discussion of my skepticism of models that will eventually be re-opened. @NoceraNYT Permalink 10:29 AM – 21 Mar 2016

@nntaleb It is called “first do no harm”, avoid “policies with side effects” (primum non nocere) translated: “no nocere, Nocera” @robert_cookson Permalink 10:09 AM – 21 Mar 2016

@nntaleb If you have trouble understanding this, you really have problems. @robert_cookson http://www.zerohedge.com/news/2016-03-12/nassim-taleb-sums-americas-election-17-black-swan-words   Permalink 9:57 AM – 21 Mar 2016

@nntaleb A journo who says “unintelligible” reveals more about his intellect than the intelligibility of the subject matter https://twitter.com/noceranyt/status/708650770380546049   Permalink 9:36 AM – 21 Mar 2016

@nntaleb Thanks to Scott Robeson, this is the expected pseudo-significant results <5% for true p-value @indianaclimate pic.twitter.com/sZ1863QyzA Permalink 8:05 AM – 20 Mar 2016

@nntaleb Friends, the metadistribution of p-values: rewrote, looking for mistakes bef publ on ArXiv. http://fooledbyrandomness.com/pvalues.pdf   pic.twitter.com/hn25LnI37C Permalink 6:08 AM – 20 Mar 2016

@nntaleb For all his defects, more pple would like to become Trump (yacht/American dream) rather than some reasonable-sounding hack criticizing him. Permalink 3:05 PM – 19 Mar 2016

@nntaleb The more efficient you try to be at a given task, the more you will hate it. Permalink 3:51 PM – 17 Mar 2016

@lhfang New Saudi PR effort SAPRAC led by guy who recently helped regime smear Nimr al-Nimr, gov critic they executed. pic.twitter.com/9IbCNKDxFz Permalink 2:08 PM – 17 Mar 2016

@nntaleb A 3% significance is really 16% @pehr_bjornbom @antitrustly @ReneMaric pic.twitter.com/ZcQdCggnmf Permalink 12:31 PM – 17 Mar 2016

@nntaleb Real-world rigor (Feynman-like): bring mathematics to deeply understood problems, don’t take deeply understood mathematics to problems. Permalink 12:00 PM – 17 Mar 2016

@nntaleb To get a real p-val of 5% you need 1.1%; 5% is ~ 26% https://dl.dropboxusercontent.com/u/50282823/p-val.pdf   @antitrustly @ReneMaric @BrianNosek pic.twitter.com/EHXHV4xVvZ Permalink 11:33 AM – 17 Mar 2016

@yaneerbaryam Military strategy in a complex world — how antifragility undermines traditional military strategies http://www.necsi.edu/research/military/strategy.html   @nntaleb Permalink 10:08 AM – 17 Mar 2016

@nntaleb Why didn’t anyone pull out the “p-value” of the “p-value” before? @ReneMaric @BrianNosek Permalink 5:46 AM – 17 Mar 2016

@nntaleb Only solution: it you “really” want a p-value of .05, require .0125! @BrianNosek @ReneMaric Permalink 5:35 AM – 17 Mar 2016

@nntaleb I mean p-values are random variables that do not necessarily reflect statistical significance. I can’t find in lit. @ReneMaric @BrianNosek Permalink 5:26 AM – 17 Mar 2016

@nntaleb Why aren’t psychologists (@BrianNosek) aware that p-values have HUGE sampling error? #replicationcrisis https://dl.dropboxusercontent.com/u/50282823/p-val.pdf   Permalink 5:19 AM – 17 Mar 2016

@nntaleb typos pic.twitter.com/z9BQiLlsQp Permalink 6:41 PM – 16 Mar 2016

@nntaleb On plane proof p-values are extremely random/volatile & explicit variance as function of n. https://dl.dropboxusercontent.com/u/50282823/p-val.pdf   pic.twitter.com/IYIN215S3a Permalink 6:34 PM – 16 Mar 2016

@nntaleb Evidence based? That is ignorant BS. Anti-GMO is scientific scrutiny. https://twitter.com/NinjaEconomics/status/710263345153093632   Permalink 6:08 PM – 16 Mar 2016

@nntaleb Experiment: Farro w/ squid ink. pic.twitter.com/Qzzkxud18w Permalink 5:18 PM – 16 Mar 2016

@nntaleb People are so concerned about my age. pic.twitter.com/JHXlNBinkr Permalink 3:57 PM – 16 Mar 2016

@nntaleb Victory! Senate Rejects the DARK Act http://ecowatch.com/2016/03/16/senate-rejects-dark-act-gmo-labeling/   via ecowatch Permalink 2:19 PM – 16 Mar 2016

@nntaleb This is the kind of BS people clueless about tail risk utter. Farmers’ modifications belong to a narrow basin. https://twitter.com/neiltyson/status/710093894847959040   Permalink 1:45 PM – 16 Mar 2016

@nntaleb Heuristic: when in a hotel for the very first time, reject the first room they give you. Permalink 5:49 AM – 16 Mar 2016

@nntaleb Monsanto forces farmers to bear all the liability for its seeds, making itself immune to adverse consequences. pic.twitter.com/khRh3Sff1J Permalink 1:48 AM – 15 Mar 2016

P-Values, Trump, Maximum Entropy Methods, Kahneman, Lindy Effect | Twitter

@nntaleb In addition to the metadistribution of p-values, a technique for large scale simulations via Radon-Nikodym deriv https://dl.dropboxusercontent.com/u/50282823/Radon%20Nikodym%20Monte%20Carlo.nb.pdf   Permalink 12:23 PM – 14 Mar 2016

@nntaleb Sicilian squid ink at Taormina in London. pic.twitter.com/OIzC6vojNq Permalink 6:29 AM – 14 Mar 2016

@nntaleb Some people, when they change language, if they don’t have an accent in it, also change personality. Permalink 4:57 AM – 14 Mar 2016

@nntaleb I don’t take face value @limitlesskait @realDonaldTrump @nytimes Permalink 12:08 PM – 13 Mar 2016

@nntaleb Donaldo, tell us something we don’t know. BTW,Trump wass loved by his butlers while Hillary was hated by bodyguards. https://twitter.com/realDonaldTrump/status/709089928324489216   Permalink 12:03 PM – 13 Mar 2016

@nntaleb Proud of being gentle with the waiter & tough with powerful ass***les rather than reverse. https://www.facebook.com/johnfaithful.hamer/posts/10153499619627683?pnref=story   pic.twitter.com/sdCggdk3D7 Permalink 11:28 AM – 13 Mar 2016

@nntaleb The metadistribution of the p-value and the #replicationcrisis https://dl.dropboxusercontent.com/u/50282823/p-val.pdf   Permalink 8:22 AM – 13 Mar 2016

@nntaleb To summarize from above, P-values are highly stochastic and the metadistribution is not factored in. Compiling notes. #replicationcrisis Permalink 7:58 AM – 13 Mar 2016

@ModeledBehavior Scary final act twist would be Trump successfully runs to the middle & before it’s too late Democrats realize he’s burning down both parties Permalink 7:05 AM – 13 Mar 2016

@nntaleb (cont) More rigorous proof via Maximum Entropy Methods give a truncated exponential that mimics a beta. pic.twitter.com/caKFyoYtq1 Permalink 7:02 AM – 13 Mar 2016

@nntaleb Answer to QUIZ du JOUR: The Distribution of a P-Value. Psychologists seem to ignore p is stochastic! pic.twitter.com/wWnYkB4SEk Permalink 3:05 AM – 13 Mar 2016

@nntaleb Reader:”I have read all your books. How do you recommend an average young man navigate the modern world? “. Me:”Never ask a vague question”. Permalink 1:25 AM – 13 Mar 2016

@nntaleb QUIZ DU JOUR. What is the probabilisty distribution of a P-Value? (may be offline for 16 hours please put answer in tweet) Permalink 3:09 PM – 12 Mar 2016

@nntaleb 1) You remove segment of samples and redo P-Values to ensure stability:ignore P, it is its fragility (convexity) @learnfromerror @alxpr1c3 Permalink 3:02 PM – 12 Mar 2016

@nntaleb The only solid use I found for P-Values is a test of robustness across jacknife (here 1-p) @learnfromerror @alxpr1c3 pic.twitter.com/1k3Yko2wyl Permalink 2:54 PM – 12 Mar 2016

@nntaleb My catalogue of how what is robust in psychology today needs to abide by #LindyEffect https://twitter.com/VergilDen/status/708776124441042946   Permalink 2:08 PM – 12 Mar 2016

@zerohedge Nassim Taleb Sums Up America’s Election In 17 “Black Swan” Words http://www.zerohedge.com/news/2016-03-12/nassim-taleb-sums-americas-election-17-black-swan-words   Permalink 12:36 PM – 12 Mar 2016

@nntaleb Kahneman is exceptionally scientific as work matches the #LindyEffect . I don’t see him as a psychologist but a decision-theorist. @RiberJan Permalink 12:32 PM – 12 Mar 2016

@nntaleb It turned out Mr Waugh was spreading rumors of “climate denial” in 2009 & he will not escape. @neilchriss @paulwaugh @PCollinsTimes Permalink 12:00 PM – 12 Mar 2016

@aral Abject moral failure illustrated in 140 characters. https://twitter.com/JohnKerry/status/708368631508262913   Permalink 6:38 AM – 12 Mar 2016