Roll over image to zoom in
Statistical Consequences of Fat Tails: Real World Preasymptotics, Epistemology, and Applications (Revised Edition) (Technical Incerto)
£18.60
The monograph investigates the misapplication of conventional statistical techniques to fat tailed distributions and looks for remedies, when possible.
Switching from thin tailed to fat tailed distributions requires more than “changing the color of the dress”. Traditional asymptotics deal mainly with either n=1 or n=∞, and the real world is in between, under of the “laws of the medium numbers” –which vary widely across specific distributions. Both the law of large numbers and the generalized central limit mechanisms operate in highly idiosyncratic ways outside the standard Gaussian or Levy-Stable basins of convergence.
A few examples:
+ The sample mean is rarely in line with the population mean, with effect on “naive empiricism”, but can be sometimes be estimated via parametric methods.
+ The “empirical distribution” is rarely empirical.
+ Parameter uncertainty has compounding effects on statistical metrics.
+ Dimension reduction (principal components) fails.
+ Inequality estimators (GINI or quantile contributions) are not additive and produce wrong results.
+ Many “biases” found in psychology become entirely rational under more sophisticated probability distributions
+ Most of the failures of financial economics, econometrics, and behavioral economics can be attributed to using the wrong distributions.
This book, the first volume of the Technical Incerto, weaves a narrative around published journal articles.
Switching from thin tailed to fat tailed distributions requires more than “changing the color of the dress”. Traditional asymptotics deal mainly with either n=1 or n=∞, and the real world is in between, under of the “laws of the medium numbers” –which vary widely across specific distributions. Both the law of large numbers and the generalized central limit mechanisms operate in highly idiosyncratic ways outside the standard Gaussian or Levy-Stable basins of convergence.
A few examples:
+ The sample mean is rarely in line with the population mean, with effect on “naive empiricism”, but can be sometimes be estimated via parametric methods.
+ The “empirical distribution” is rarely empirical.
+ Parameter uncertainty has compounding effects on statistical metrics.
+ Dimension reduction (principal components) fails.
+ Inequality estimators (GINI or quantile contributions) are not additive and produce wrong results.
+ Many “biases” found in psychology become entirely rational under more sophisticated probability distributions
+ Most of the failures of financial economics, econometrics, and behavioral economics can be attributed to using the wrong distributions.
This book, the first volume of the Technical Incerto, weaves a narrative around published journal articles.
Read more
Additional information
| Publisher | STEM Academic Press (25 July 2023) |
|---|---|
| Language | English |
| Paperback | 473 pages |
| ISBN-13 | 979-8218248031 |
| Dimensions | 15.24 x 2.72 x 22.86 cm |
by Serban Dragne
Fat tails are one of the the most misunderstood statistical phenomena and Taleb explains this in an easy to understand format (you do need a fair understanding of mathematics and probabilistic theory though). It has made me question some of the risk models we use in cybersecurity as well as in financial hedging strategies. The takeaway for me from this book isn’t a formulaic simplification of rules, it gives you a better understanding of how mistakes are made, how risks are misunderstood both in terms of volatility and predictability.
by Serban Dragne
Taleb has this knack of mixing hardcore math alongside jibes at things he finds annoying about how stats is (mis)applied in the real world from topics that seem to range from finance to wars. I have only had it a few days but already I’ve highlighted the PDF version to death with notes I need to come back and reread.
Content aside, you can get the PDF freely available on Monsieur’s website, but I strongly recommend you buy a physical copy because it’s legitimately the nicest book I have ever owned. It’s gorgeous! I tried to do it justice in the photos but I don’t think it expresses how NICE this book feels. Even though it’s not a big book it’s got this heft to it (you could crack some skulls with this thing). The pages are beautifully printed with this a lovely rich ink. It even has a great smell! Yes. A great smelling book. My wife thought i was mad i was sniffing a book. ‘Technical’ books aren’t meant to be shown this much love and attention
by Thien Huong Vu
Taleb shows in this tome, through his profligate use of algebraic equations that he’s on a separate plain to any other trader in the City (of London and elsewhere).
My own 21 year career in finance breaks down as follows:
1. 4 years as a clerk in retail banking
2. 4 years as a management accountant in an investment (merchant) bank in London
3. 2 years as a stock (equity) trader on the European markets.
4. 8 years as a financial derivatives sales person.
5. 2 more years (the final two) as a freelance consultant to the financial sector in the UK.
by M
This is a book that tries to achieve two things. First to explain why fat tails are and why they matter to ordinary people who haven’t got a PhD in Advanced Mathematics. The second part is trying to illustrate the mathematics involved that underpins this theory and why it is different from the type of statistical analysis most commonly done by all kinds of experts, whether they be laymen like Hospital consultants trying to understand the efficacy of a new medical treatment, epidemiologists attempting to understand or predict the future or Taleb’s favorite past time buying insurance policies in equity markets (options trading).
If you are a young budding statistician, economist, researcher or a day trader, medium or long term trader and you want to understand something about the work you do, which you probably will not find in your curriculum, then this book will probably provide 200-300% on top of your undergraduate studies and put you miles ahead from your peers.
If you want to understand current affairs, Covid-19, from a statistical point of view then the book is full of examples which anyone could understand. It may be hard for non-mathematicians to find value, but Taleb have attempted to categorise chapters into who they are intended for.
Note however, that like some of his prized partners, Taleb is sometimes a very lazy person who draws conclusions without penetrating a subject thoroughly and so when he is commenting about stuff in a casual manner, do pay attention to not be drawn in by his fallibility to become that which he detests most – the expert who comments on stuff he has no real understanding of.
by Rob Peach
So, I was looking forward to this book. Sadly I was disappointed. I was disappointed by the price and the content. In summary, it is made up of the all the previous material in the Incerto collection. Which is to be expected. No surprise there. Only this time its underpinned with maths. This is the technical part referred to in the title. Luckily I’ve a background in Maths. Sadly, it still didn’t make the book any more interesting. I’m afraid I wont be recommending this book.
by Dionysius the Areopagite
The book is necessarily more heavy-going than Taleb’s other work, but far more useful for it. It was invaluable to use during my Master’s thesis. It is both more useful and much more affordable than other textbooks in the EVT space.
Stefan Gasic’s comic strips break up the text and provide light relief, although they become less common as the book goes on.
My only criticism of the work is that Taleb sometimes leaves his graph axes unlabeled, although you can always work out what the graph means eventually through context.