Nassim Taleb on the black swans of war

[citizen k]

Nassim Taleb on the black swans of war
Yahoo Finance By Justine Underhill


As the frequency of wars among great powers dwindles, has the world entered a new paradigm of peace and prosperity? Not quite, says Nassim Taleb, author of Fooled by Randomness and The Black Swan.

In their new paper, statistician Pasquale Cirillo and Taleb examine wartime events of the last 2000 years, and they find no statistical evidence that the frequency or magnitude of wars are declining. The "long peace" theory, which claims that violence has declined, is a falsehood says Taleb. He notes that people are lulled into a false sense of security because of the nature of so-called fat tail – or extreme – events.   

"You cannot make claims that violence has dropped," he says. "Simply because, under fat tails, events can take hundreds of years.  Large events can take a long time...and between events is totally unpredictable...they occur randomly with no time structure, and we saw no evidence of decline."

They also find that catastrophically fatal wars have fatter tails than those of financial crises, which suggests a propensity for extremes. Taleb notes that the severe cases of war are more deleterious than those of finance.

"It's much worse in the sense that a smaller number of events determine a larger share of the casualties," he says.

Get the Latest Market Data and News with the Yahoo Finance App

Taleb adds that the role of randomness is misunderstood. He explains that using a past average to forecast a future average leads to a considerable underestimation of the number of casualties.

Regarding the role of statistical tools, Taleb says, “Physicists get the point. But social scientists do not get it. Their statistical tools do not work for what we call fat tails, as defined as something prone to black swan events," he says. "They can write whatever narrative they want. But we call that journalistic. That's not statistical."

Taleb clarifies the point, as he says, "If someone tells you violence has dropped and shows you fifty years of data, just laugh at the person. That's what we're saying."

video at link: http://finance.yahoo.com/news/nassim-taleb-on-the-black-swans-of-war-194841536.html

[jimmy olsen]

Complete bullshit.

[grumbler]

Seems like doubletalk.  They are correct to note that we don't know the "proper" time period for evaluating levels of violence, but, to paraphrase Taleb, "If a statistician tells you he has a better timeframe for determining random war events than you do, just laugh at the person."

[The Minsky Moment]

Regarding the role of statistical tools, Taleb says, "Physicists get the point. But social scientists do not get it. Their statistical tools do not work for what we call fat tails, as defined as something prone to black swan events," he says.

The assumption here is that statisticians can only work with normal distributions which is false.

[HisMajestyBOB]

His book on black swans is interesting in that it provokes discussion and thought on "how do we plan for the unknown?", but he does get pretty obsessive. It's fun watching his rants on Youtube against economists.

[MadImmortalMan]

Regarding the role of statistical tools, Taleb says, "Physicists get the point. But social scientists do not get it. Their statistical tools do not work for what we call fat tails, as defined as something prone to black swan events," he says.

The assumption here is that statisticians can only work with normal distributions which is false.

I think it's more that statisticians tend to discard outliers.

[Admiral Yi]

I think it's more that History and Poli Sci folks stop at intro to stats.

Not sure how fat tails changes the frequency.  That's just a descriptor.

[Eddie Teach]

We're in the calm before the storm, folks. 

[DGuller]

Regarding the role of statistical tools, Taleb says, "Physicists get the point. But social scientists do not get it. Their statistical tools do not work for what we call fat tails, as defined as something prone to black swan events," he says.

The assumption here is that statisticians can only work with normal distributions which is false.

Yes and no.  Obviously there are skewed distributions, but you may not have enough data to fit them accurately.  The more skewed the distribution, the more data you need.

[jimmy olsen]

The murder rate has plunged in the last thousand years across most of the world. There hasn't been a major war in the Americas in a century and a half. There hasn't been a major war in Europe in seventy years. East Asia has been at peace for forty.  It's ludicrous to argue that the world is not more peaceful than it used to be when everywhere in the world in 17th century looked like present day Syria.

[DGuller]

Not sure how fat tails changes the frequency.  That's just a descriptor.

Fat tail and frequency are directly related.  The lower the underlying frequency of the event, the more skewed the distribution of the number of events is (i.e. the more fat-tailed it is).

[DGuller]

The murder rate has plunged in the last thousand years across most of the world. There hasn't been a major war in the Americas in a century and a half. There hasn't been a major war in Europe in seventy years. East Asia has been at peace for forty.  It's ludicrous to argue that the world is not more peaceful than it used to be when everywhere in the world in 17th century looked like present day Syria.

It's ludicrous to argue by just repeating the very point being refuted.  Taleb's point is that wars may be distributed in such a way that we simply have no way of knowing whether expected losses from wars became less likely over the last 7 decades.  Is he right?  I don't know, that depends on whether he got his numbers right.  Is it conceivable that he's right?  Most definitely.

[Admiral Yi]

Fat tail and frequency are directly related.  The lower the underlying frequency of the event, the more skewed the distribution of the number of events is (i.e. the more fat-tailed it is).

I was talking about this part.

"In their new paper, statistician Pasquale Cirillo and Taleb examine wartime events of the last 2000 years, and they find no statistical evidence that the frequency ... of wars [is] declining. "

Surely that's just a counting and dividing operation.

[DGuller]

Fat tail and frequency are directly related.  The lower the underlying frequency of the event, the more skewed the distribution of the number of events is (i.e. the more fat-tailed it is).

I was talking about this part.

"In their new paper, statistician Pasquale Cirillo and Taleb examine wartime events of the last 2000 years, and they find no statistical evidence that the frequency ... of wars [is] declining. "

Surely that's just a counting and dividing operation.

"Surely"?  So what is it that you think statisticians do that arithmeticians can't do?  No, there is a hell of a lot more involved than just counting and dividing.  Counting and dividing just gives you the estimate.  The important part of statistical analysis, the one embedded in "statistical evidence" part, is calculating the confidence interval around the counted and divided number.  And the confidence interval is most surely a function of how fat the tail is.

[Razgovory]

It's occasionally interesting when we have someone who actually knows what the fuck they are talking about.