4 ms·

A critique of The Black Swan

72 points by biohacker42 18y ago

biohacker42 18y ago
This is a very good response to Nassim Nicholas Taleb. In fact, it may have already been on HN, that's probably where I first read it and have been trying to find it again, ever since. Taleb is now turning up almost everywhere, so I thought people might find this interesting.
- hardik 18y ago
  You might be referring to this: http://news.ycombinator.com/item?id=353264 http://news.ycombinator.com/item?id=353264
thomaspaine 18y ago
Taleb raises a lot of valid concerns about the state of mathematical finance. You wouldn't believe the number of financial models that assume normal or log-normal returns, despite a vast amount of evidence to the contrary. Value at Risk (VaR) is a terribly deficient risk metric, and I've never met anyone who actually believes that Black-Scholes is an accurate model for predicting options prices, yet both of these are widely used in the financial industry. My problem with Taleb and all the attention he's been getting lately because of the financial crisis is that I think it's a logical jump to go from those criticisms to "these models are bad, because of black swans". Black swans are defined as unpredictable random events with large impact. 9/11 was a black swan, but I don't think that the financial crisis we're seeing today is a result of a black swan. Insurers ignored systemic risk and over leveraged based on bad asset pricing, but were those assets priced poorly because of black swans? I think that's a tough argument to make, because mortgage foreclosures don't seem nearly as unpredictable as say, terrorists flying airplanes into buildings. There also seems to be a bit of observation bias in concluding that events are black swans. Prediction markets can't be right 100% of the time. To quote the article: "What does it mean to say such a prediction is right or wrong? In 2008, the day before John McCain was scheduled to announce his VP choice, the Intrade prediction market gave Sarah Palin a 4% chance. Was this right or wrong? Unlikely events will sometimes happen just by chance."
- yters 18y ago
  What other models can they use, though? Every other distribution assumes even more about the problem domain, and would thus correlate even worse with black swans. Sounds kind of like a no free lunch theorem for finance. Anyways, do you know if anyone is trying to come up with something better?
  - thomaspaine 18y ago
    There are other types of distributions which fit black swan type data better, such as power-law distributions, Gumbel distribution, etc. I don't know if research in extreme value theory is still a hot topic, but it tries to address these problems. Here is a paper on risk management and extreme value theory: http://tinyurl.com/aa7t6m http://tinyurl.com/aa7t6m Here is a nice paper on risk measures, and points out some of the problems with VaR: http://tinyurl.com/ar3336 http://tinyurl.com/ar3336 My apologies if you're unable to grok the contents of these papers...they're both a little technical but I can't think of anything off the top of my head that's more accessible.
- MaysonL 18y ago
  My problem with Taleb and all the attention he's been getting lately because of the financial crisis is that I think it's a logical jump to go from those criticisms to "these models are bad, because of black swans". These models are bad because the companies that have been bet on their accuracy go bust. They go bust because they assume that the probability of "black swans" such as housing prices throughout the country all going down at the same time is vanishingly small when it isn't.
- mstoehr 18y ago
  The orthodox school of thought in financial economics was the efficient market hypothesis, which more or less states that the best valuation on current assets is the market price. I suppose there is something to be said for pointing out that the "best valuation" may not be very good. Taleb certainly wasn't the first to bring this up, but we probably aren't worse off for being reminded of that fact.
- mattobrien 18y ago
radu_floricica 18y ago
This isn't really a... well, it is a critique in the european sense, but it is not really a rebuttal. Most of it agrees with NNT, and when it doesn't it's usually on a matter of form.
- Agathos 18y ago
  Wait. Are you saying "critique" and "rebuttal" are synonyms? I never thought so, and I've never even been to Europe.
kurtosis 18y ago
Statistical natural language processing is a very prominent example of a field which has developed effective methods for coping with "black swans" or the occurance of improbable events that have never been seen before. If you take any large corpus of text and start counting the frequencies that different words occur you will rapidly notice that a huge fraction of the words have never been seen before or occur only once. Biblical scholars called these words hapax legomena instead of black swans. Methods for estimating the probabilities of these events go back to Alan Turning and his codebreaking work at Blechley Park. One need not assign zero probability to unseen events a la maximum likelihood. Taleb's rants against VAR and mediocristan have always seemed like borderline straw man bashing to me. Sure there are many fools who believe in the gaussian or lognormal returns model, but the best knowledge in the field doesn't make these assumptions. Why doesn't he give authors like bouchard and potters who have built on mandelbrot's work their due? or does he?
- navanit 18y ago
  I'm assuming you're referring to Jean-Philippe Bouchaud and Marc Potters. Taleb does give credit to them in terms of them understanding the presence of fat tails and focussing on the failure of the gaussian. However, both Bouchaud and Potters, coming from physics backgrounds, believe that the tail-exponent can be calibrated accurately from a finite sample-set. When in fact estimating the tail-exponent (the 'alpha' of a power law process) is fraught with small-sample effects and one cannot reliably make decisions even when you do come up with an estimate of the tail-exponent. This is Taleb's main beef with the breed of physicists with power-law models. The best paper to understand this is: Weron 2001: "Levy-stable distributions revisited: Tail index > 2 does not exclude the levy-stable regime" http://citeseer.ist.psu.edu/448515.html http://citeseer.ist.psu.edu/448515.html More on my website if you're interested: navanitarakeri.com A Black Swan is not just a rare event - it is defined as a rare event with _high impact_. Statistical NLP may handle rare occurrences well (are you talking about smoothing?), but, by it's very nature, a rare event in a block of text is not going to wipe out a library now, is it? So the appearance of a rare word in a block of text is hardly in the same class of problems as epidemics, wars, market crashes etc.
  - 18y ago
lgriffith 18y ago
Calling an event "random" does not mean its causeless. It simply means its not predicable. Its not predicable because we don't know and/or haven't measured the relevant causes. Hence, Black Swan events are caused. Perhaps it would be an interesting and profitable exercise to look for and measure causes of Black Swan events.
- cdog46 18y ago
  So long as people continue to confuse explanation with description-we'll have confusion. Math is description; science is description with comparison and the contrasting of said description(s). Then what is explanation? I have no clue but to say it is part of consciousness, and,in the case of homo sapiens, partial consciousness. To say that us(humans) with our four dimensional rules of calculation & perception can "know" the how of everything is no more proposterious than to believe the earth is flat. Put another way-scientists who too often reject a new theory put forth that challenges current theories or assumptions-it reminds me of religious zealots defending any challenge to their dogma

72 points by biohacker42 18y ago

This is a very good response to Nassim Nicholas Taleb. In fact, it may have already been on HN, that's probably where I first read it and have been trying to find it again, ever since. Taleb is now turning up almost everywhere, so I thought people might find this interesting.

hardik 18y ago
You might be referring to this: http://news.ycombinator.com/item?id=353264 http://news.ycombinator.com/item?id=353264

thomaspaine 18y ago

Taleb raises a lot of valid concerns about the state of mathematical finance. You wouldn't believe the number of financial models that assume normal or log-normal returns, despite a vast amount of evidence to the contrary. Value at Risk (VaR) is a terribly deficient risk metric, and I've never met anyone who actually believes that Black-Scholes is an accurate model for predicting options prices, yet both of these are widely used in the financial industry. My problem with Taleb and all the attention he's been getting lately because of the financial crisis is that I think it's a logical jump to go from those criticisms to "these models are bad, because of black swans". Black swans are defined as unpredictable random events with large impact. 9/11 was a black swan, but I don't think that the financial crisis we're seeing today is a result of a black swan. Insurers ignored systemic risk and over leveraged based on bad asset pricing, but were those assets priced poorly because of black swans? I think that's a tough argument to make, because mortgage foreclosures don't seem nearly as unpredictable as say, terrorists flying airplanes into buildings. There also seems to be a bit of observation bias in concluding that events are black swans. Prediction markets can't be right 100% of the time. To quote the article: "What does it mean to say such a prediction is right or wrong? In 2008, the day before John McCain was scheduled to announce his VP choice, the Intrade prediction market gave Sarah Palin a 4% chance. Was this right or wrong? Unlikely events will sometimes happen just by chance."

yters 18y ago
What other models can they use, though? Every other distribution assumes even more about the problem domain, and would thus correlate even worse with black swans. Sounds kind of like a no free lunch theorem for finance. Anyways, do you know if anyone is trying to come up with something better?
- thomaspaine 18y ago
  There are other types of distributions which fit black swan type data better, such as power-law distributions, Gumbel distribution, etc. I don't know if research in extreme value theory is still a hot topic, but it tries to address these problems. Here is a paper on risk management and extreme value theory: http://tinyurl.com/aa7t6m http://tinyurl.com/aa7t6m Here is a nice paper on risk measures, and points out some of the problems with VaR: http://tinyurl.com/ar3336 http://tinyurl.com/ar3336 My apologies if you're unable to grok the contents of these papers...they're both a little technical but I can't think of anything off the top of my head that's more accessible.
MaysonL 18y ago
My problem with Taleb and all the attention he's been getting lately because of the financial crisis is that I think it's a logical jump to go from those criticisms to "these models are bad, because of black swans". These models are bad because the companies that have been bet on their accuracy go bust. They go bust because they assume that the probability of "black swans" such as housing prices throughout the country all going down at the same time is vanishingly small when it isn't.
mstoehr 18y ago
The orthodox school of thought in financial economics was the efficient market hypothesis, which more or less states that the best valuation on current assets is the market price. I suppose there is something to be said for pointing out that the "best valuation" may not be very good. Taleb certainly wasn't the first to bring this up, but we probably aren't worse off for being reminded of that fact.
mattobrien 18y ago

radu_floricica 18y ago

This isn't really a... well, it is a critique in the european sense, but it is not really a rebuttal. Most of it agrees with NNT, and when it doesn't it's usually on a matter of form.

Agathos 18y ago
Wait. Are you saying "critique" and "rebuttal" are synonyms? I never thought so, and I've never even been to Europe.

kurtosis 18y ago

Statistical natural language processing is a very prominent example of a field which has developed effective methods for coping with "black swans" or the occurance of improbable events that have never been seen before. If you take any large corpus of text and start counting the frequencies that different words occur you will rapidly notice that a huge fraction of the words have never been seen before or occur only once. Biblical scholars called these words hapax legomena instead of black swans. Methods for estimating the probabilities of these events go back to Alan Turning and his codebreaking work at Blechley Park. One need not assign zero probability to unseen events a la maximum likelihood. Taleb's rants against VAR and mediocristan have always seemed like borderline straw man bashing to me. Sure there are many fools who believe in the gaussian or lognormal returns model, but the best knowledge in the field doesn't make these assumptions. Why doesn't he give authors like bouchard and potters who have built on mandelbrot's work their due? or does he?

navanit 18y ago
I'm assuming you're referring to Jean-Philippe Bouchaud and Marc Potters. Taleb does give credit to them in terms of them understanding the presence of fat tails and focussing on the failure of the gaussian. However, both Bouchaud and Potters, coming from physics backgrounds, believe that the tail-exponent can be calibrated accurately from a finite sample-set. When in fact estimating the tail-exponent (the 'alpha' of a power law process) is fraught with small-sample effects and one cannot reliably make decisions even when you do come up with an estimate of the tail-exponent. This is Taleb's main beef with the breed of physicists with power-law models. The best paper to understand this is: Weron 2001: "Levy-stable distributions revisited: Tail index > 2 does not exclude the levy-stable regime" http://citeseer.ist.psu.edu/448515.html http://citeseer.ist.psu.edu/448515.html More on my website if you're interested: navanitarakeri.com A Black Swan is not just a rare event - it is defined as a rare event with _high impact_. Statistical NLP may handle rare occurrences well (are you talking about smoothing?), but, by it's very nature, a rare event in a block of text is not going to wipe out a library now, is it? So the appearance of a rare word in a block of text is hardly in the same class of problems as epidemics, wars, market crashes etc.
- 18y ago

lgriffith 18y ago

Calling an event "random" does not mean its causeless. It simply means its not predicable. Its not predicable because we don't know and/or haven't measured the relevant causes. Hence, Black Swan events are caused. Perhaps it would be an interesting and profitable exercise to look for and measure causes of Black Swan events.

cdog46 18y ago
So long as people continue to confuse explanation with description-we'll have confusion. Math is description; science is description with comparison and the contrasting of said description(s). Then what is explanation? I have no clue but to say it is part of consciousness, and,in the case of homo sapiens, partial consciousness. To say that us(humans) with our four dimensional rules of calculation & perception can "know" the how of everything is no more proposterious than to believe the earth is flat. Put another way-scientists who too often reject a new theory put forth that challenges current theories or assumptions-it reminds me of religious zealots defending any challenge to their dogma