Naive empiricism, terrorism & Ebola.
Channel:N N Taleb's Probability Questions · Length:10:08 · Published:2017-02-14 · Views:51586
Summary
Nassim Taleb argues that evidence on terrorism risk presented in media is statistically naive, using extreme value theory without math to separate fat-tailed risks like Ebola and terrorism from thin-tailed ones.
Why HN recommends it
HN cites it when arguing that different risks follow different distributions and should be judged by growth rates and tails rather than current counts.
Watch this to understand that we're talking about two different distributions — HN, 2017
Data science & statistics · Probability & game theory
Subtitles (excerpt)
friends um thank you for your interest uh this discussion is going to be about public misunderstanding of risk and uh how uh when it comes to terrorism or Ebola the uh the the press and actually many academics people who should know better uh conduct some kind of naive and prism and mislead us completely you should never compare things that are fat tailed to things that don't have the same tail risk attributes at the collective level thin tails and fat tails should never be compared so let's start uh let me quickly explain the difference between mediocris Stan and extremist Stan medran is a domain in which tail events do not have very large consequences and exem is one in which everything comes from the tals so let's take uh uh the following um example as a thought experiment you have a tail event you just uh sampled randomly from the population and um and and uh uh picked two people with a total height of 4.1 M what's the most likely combination 1.1 and uh three no 10 cm and four no the most likely combination is 4.05 2.05 and 2.05 simply the probability of exceeding three sigmas in a gan distribution assuming three sigas okay with no loss of generality um is something like 0134 the probability of exceeding that twice is a square but the probability of
On Hacker News
No it isn't, in this case. I think Taleb has the best summary of this, for all his flaws: [VIDEO] In short, risk due to coconuts vs sharks is a valid comparison because we have good priors about growth rates, and they are similar. Coconuts vs covid is a terrible, terrible comparison because again we have good priors on growth but they are mind-bogglingly different functions of time. — nmca, 2021-02
I will give an alternative explanation why people fear terrorism: Falling from a ladder also kills more people than terrorism, right, but this is not what people are afraid of. Watch this to understand that we're talking about two different distributions: [VIDEO] Terrorism SUDDENLY increases the odds of you being killed by it. And that is what - rightfully - scares people. Example: Before 9/11 th… — s3nnyy, 2017-10
I think Taleb has a point on this topic - the concern is about a forecasted growth rate, not about the current rate. [VIDEO] — nmca, 2021-09
Funny, Nassim Taleb, a worldwide (if controversial) risk probability expert, would say that you are the one which is being very bad at reasoning with probabilities, because you are putting Ebola and terrorism in the same class of risk as falling off a ladder: [VIDEO] — 21, 2018-04
Related videos
- Do we have to accept Quantum weirdness? De Broglie Bohm Pilot Wave Theory explained · Looking Glass Universe · 11:59
- Avoiding Negative Side Effects: Concrete Problems in AI Safety part 1 · Robert Miles AI Safety · 9:33
- Reinventing Organizations · Reinventing Organizations · 1:42:52
- A Better Default Colormap for Matplotlib | SciPy 2015 | Nathaniel Smith and Stéfan van der Walt · Enthought · 19:08
- Why Being Delusional is a Superpower · Veritasium · 12:04
- Why Your Vote Doesn't Count · CGP Grey · 6:31
- Is Most Published Research Wrong? · Veritasium · 12:22
- Once you understand game theory, you understand the world. · Veritasium · 27:18
- The Voting System That Should Replace Every Election · CGP Grey · 4:27
- How not to be ignorant about the world | Hans and Ola Rosling · TED · 19:09
- I don't like notebooks.- Joel Grus (Allen Institute for Artificial Intelligence) · O'Reilly · 56:12
- VisiData Lightning Demo at PyCascades 2018 · Saul Pwanson · 4:50
- Emacs + org-mode + python in reproducible research; SciPy 2013 Presentation · Enthought · 21:16
- I Like Notebooks · Jeremy Howard · 52:09
- golden balls. the weirdest split or steal ever! · spinout3 · 6:25
- Interactive notebooks ~ Pluto.jl | Fons van der Plas | JuliaCon 2020 · The Julia Programming Language · 23:37