Jump to content

Talk:Taleb distribution

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Fix reference

[edit]

The first reference doesn't seem to link to anything

Unsigned comment of 05:00, 2009 February 12 212.249.2.158
Thanks – fixed!
—Nils von Barth (nbarth) (talk) 22:30, 15 March 2009 (UTC)[reply]

Not a statistical term

[edit]

This article should be deleted. This is a nonsensical term, which is used nowhere else, neither in statistics, nor in probability, economics or finance. This seems like a lame attempt at publicity by Taleb. The only person who has used this term is apparently 'John Kay' who appears to have a confused view of Taleb's contributions and statistical notions. 82.120.131.17 (talk) 12:01, 13 December 2014 (UTC)[reply]

Taleb distribution is not a term used by statisticians, and is not a term used by market practicioners. This article should be deleted. A google search shows that this term is not commonly used term, other than in mirrors of this piece, or by a man named John Kay who is quoted in the piece. http://www.google.co.uk/search?q=%22taleb+distribution%22&sourceid=ie7&rls=com.microsoft:en-us:IE-SearchBox&ie=&oe=&redir_esc=&ei=Pb-cTPn3AZuN4gbsnt2iDQ

86.179.192.71 (talk) 15:11, 24 September 2010 (UTC)[reply]

The term does not necessarily mean "distribution" in the statistical sense; it is there to signify MORAL HAZARD. More persons than John Kay have used the term. The article should not be deleted. Someone has visibly started a new account just for that. 207.134.68.130 (talk) 15:44, 24 September 2010 (UTC)[reply]

This page should be deleted as the term is not notable and commonly used, particularly it is hardly used by economists, statisticians, or probability theorists. It seems to be primarily used by two authors of the Financial Times. As such, this page could appear as part of an effort to invent and popularize the term. To avoid this, I suggest that this page be deleted for now. If the term catches on without being on Wikipedia, then it can be relisted. Seabonn (talk) 21:51, 28 May 2014 (UTC)[reply]

Given that the two practitioners at FT, both Wolf and Kay, are world renowned experts in Economics and Kay is a professor at LSE, I think the belittlement above is quite over the top. Moreover, it is disingenuous as it suggests Taleb gave this distribution his name, and it was Wolf and Kay who did. Moreover, their characterization is perfectly proper, as it is a qualitative description of a set of leptokurtic distributions, here, ones pertaining to the mathematics of finance, which necessarily includes statistics of finance. The claim that "Taleb distribution is not a term used by statisticians" is spurious, since statisticians use a lot of inconsistent terminology, depending upon their field of practice, e.g., Cauchy distribution versus Lorentz distribution. Google Search is hardly a signature of definition in a technical field.

I would have more respect for the objection if claims like "[Kay] appears to have a confused view of Taleb's contributions and statistical notions" were manifested with technical evidence of what precisely is confused, preferably with mathematical specificity.

There is no reason why a statistical distribution need only be characterized by a closed form expression. Indeed, most statistical distributions in practice cannot be so characterized.

I say the page and characterization is fine.

bayesianlogic.1@gmail.com

This user is a member of WikiProject Statistics.


Alternative title

[edit]

"Peso problem" redirects to this page. They express similar concepts, and "Peso Problem" is a much more widely used term that has been around since the 70's. I'm happy to build out the page more (I have a bit of experience in this area of finance), but want to see if anyone thinks it makes sense before I spend the time to revamp. Thanks Jlhender (talk) 22:27, 5 February 2011 (UTC)[reply]

Agreed that this article should be expanded with Peso Problem background. Regarding the merge proposal by Sailsbystars, I prefer to keep this article separate, with the same reasoning that "Peso Problem" has been in use since before Fooled by Randomness / Black Swan. Quarl (talk) 07:26, 14 March 2011 (UTC)[reply]
Hello, the Peso problem is included in the Taleeb distribution, but not the reverse since the subject of the Taleeb distribution is hiding risk in the tails so there is a "probabilistic agency problem" in it. — Preceding unsigned comment added by 64.134.67.255 (talk) 14:46, 8 June 2011 (UTC)[reply]

Mathematical model

[edit]

Is there at least a mathematicaly described example? — Preceding unsigned comment added by 79.117.44.240 (talk) 15:54, 10 December 2013 (UTC)[reply]

One example that can be described mathematically is the strategy of forgoing property insurance. The chance of a catastrophic event that destroys your house is low, and every insurance payment you skip is money gained, as long as your house doesn't burn down, fall over, and sink into the swamp. But if some catastrophe does occur, you're wiped out.
No one knows the probability of any specific structure suffering a catastrophe (except maybe an arsonist). The property owner doesn't know, the insurance company doesn't know, and the fire department doesn't know. But the insurance industry has a very good idea of the probability of a loss, and prices its services accordingly. The fire department has an idea of the probability, and uses it to decide how many fire stations, trucks, and firefighters are necessary to reduce the risk of property losses and loss of lives to a level that is tolerable to taxpayers.
But the collective probability of disaster is less important to an individual property owner than the unknown risk to his or her own property. Because of the catastrophic risk of forgoing insurance (and the demands of mortgage lendets), the property owner buys insurance, even knowing that, on average, the insurance company wins.
This example is probably described in the literature somewhere, but it was easier to describe it than find a reliable reference. As such, this is original research, and thus doesn't qualify for the main article. But a sourced version of this example might be a useful illustration. — Steve98052 (talk) 20:00, 24 January 2018 (UTC)[reply]

Regarding the Claim of Incentuous References

[edit]

Clearly this is invalid. For example, a Google Scholar search reveals, among others: McConnell, Patrick (21 March 2012). "The governance of strategic risks in systemically important banks". Journal of Risk Management in Financial Institutions. 5 (2): 128–142.

bayesianlogic.1@gmail.com

This user is a member of WikiProject Statistics.


Graph inconsistency

[edit]

The pictures of the distribution show positive mean averages, but the "payoff" or investment returns profile says that "the expected value is very much less than zero." These seem inconsistent. — Preceding unsigned comment added by 72.139.203.117 (talk) 12:48, 12 May 2022 (UTC)[reply]