Copyright: 2007

Publisher: Random House

ISBN: 978-1-4000-6351-2

Nassim Taleb writes an interesting book about unexpected events.  His theories about probability provide a different perspective on some old topics. Taleb's writing style is engaging if not eccentric.  He injects dry humor along with plenty of parenthetical comments.  He is a self-described philosopher - trader, however unlike most philosophers, he doesn't seem to be too concerned with the reader thinking he is smart.  Most modern philosophers seem more concerned with impressing the reader than informing the reader.  Taleb does share a similarity with most philosophers however in that he takes himself quite seriously...  something I find strange in people who make their living proving that most of the people who lived before us took themselves too seriously. 

Unknown Unknowns

The central theme of Taleb's Black Swan theory seems to be the fact that there are a lot of things we simply don't know and can't know.  He spends quite a bit of time pointing out that no matter how many white swans you see... you can't prove that black swans don't exist (in fact... they do exist.)  Taleb uses what he calls the "round trip fallacy" to explain this.  Just because there is no evidence of  black swans does not mean there is evidence of no black swans. 

 This is reinforced by the turkey story.  The turkey gets fed in a nice warm barn for 1,000 days straight and is perfectly rational in assuming that day 1,001 is going to be more of the same.  Of course if day 1,001 happens to fall a day or two prior to Thanksgiving a terrible change of fortune for the turkey.  No evidence of the terrible fate of the turkey certainly was not evidence that there would be no terrible fate. 

One of the keys to understanding Taleb's philosophy is understanding that from the point of view of the turkey, it was impossible to know what was going to happen.  However, the farmer knew full well what was going to happen.  So the unknown unknowns in life are not necessarily unknowable... it just depends on your point of view.

Silent Evidence

Another key is to realize that it is generally impossible to predict the future.  Looking at the past events are not necessarily an indication of what the future might hold. We like to try to reverse engineer history and try to find the causes of the outcomes that we observe.  The problem with this exercise is that we lack the full evidence of the historical narrative.  The outcomes of history only show those explorers that survived, those kings that managed to rise to power (and whose historical records survived) and those ships that did not sink.  Since we don't see the silent evidence of history, we may attribute causes to the outcomes that could easily have led to different outcomes.  As an example, consider an author that finds that most millionaires are risk takers.  The author might say that is a cause of being a millionaire.  However if they were to analyze bankruptcies, they would likely find a similar common cause.  We simply have a hard time projecting the past outcomes into future predictions.

Planning Failures

Our inability to predict the future due to the "unexpected" leads to planning failures.  "The unexpected has a one-sided effect with projects" according to Taleb, meaning that the unexpected almost always leads to delays and higher costs for a project rather than the opposite.  One notable exception was the Empire State building which was completed under budget and early, but these are the rare exception to the rule.

One of the things that Taleb notes is that we are better at predicting the "regular events" or things that have happened before.  It is the large, irregular events that we aren't good at predicting.  That suggests to me that one of the keys to project management is to "predict" on a smaller scale.  That is to say, limit what you need to predict in order to make a decision. 

Agile software methods treat this symptom by using short iterations and releasing software early and often.  Many non-Agile project managers get frustrated with this approach because they feel the need to create an ROI calculation based on the entire product rather than incremental pieces.  This is a valid concern, however as Taleb points out, even the best ROI calculations are going to fail to predict accurately due to the uncertainty of the future.  ROI tends to be a calculation of "risk" based on the known unknowns... the size of the market and the probability of capturing certain percentages of it for instance.  What it fails to consider is the unknown unknowns, for instance, what if the economy suddenly surges forward and your niche market suddenly becomes flooded with competitors with access to large amounts of VC? 

 Taking the small loss

The closest thing to practical advice that Taleb gives in the book deals with how to use our inability to predict the future to help.  The majority of traders in the markets base their risk assessments on bell curve probability theory... which Taleb believes is wrong.  He believes they should be using power law distributions rather than bell curve distributions.  This means that small shifts in the market can have dramatic effects.  So if you understand that most people are betting with the Bell curve but you know the market to be conforming to a power law distribution, you can afford to take the small losses that no one else is willing to take.  By consistently exposing yourself to small losses, with potentially large upsides, you can take advantage of the occassional small shift with large gains. 

On the flip side, you want to make sure you don't expose yourself to the potentially fatal large losses.  These are the risks that most traders are willing to take since, in theory, the probability is so small that they aren't concerned with them. 

Scalability in Career choices

One of the observations I found interesting in the book dealt with the scalability of careers / products. Taleb talks about careers that are scalable and thus desirable (or undesirable depending on your view point.)  An example of a non-scalable career would be a baker.  A baker arrives at the bakery every day and bakes his goods.  Once those goods are sold and eaten, he must then prepare another batch.  Every time he sells an item, he must make that item. 

An author on the other hand has a scalable profession.  He is able to write the book once and send it off to the publisher who then can print hundreds of thousands of copies.  The author does not need to be present to sell his book nor does she need to write a new chapter every time someone reads one. 

On first glance it might seem that choosing a scalable profession is the right choice... after all... who doesn't like "easy money"?  But scalable professions tend to suffer from the "winner takes all" side effect.  In this scenario, there are hundreds (even thousands) of authors all writing books, but only one or a very few ever sell enough books to be full time authors. 

On the flip side a dentist (a non-scalable profession) is likely to make roughly the same amount of money as every other dentist out there.  There will be some distribution of wealth but it won't be characterized by the same disparity that you will find in scalable professions.  So the "safe" route is to choose a non-scalable profession... but if you feel lucky, go for the scalable one! 

Conclusion

I find Taleb's writing to be fascinating if not terribly practical.  As a dabbler in economics, it is refreshing to see his views (even his disparaging ones) on the subject.  It is also enlightening to learn new ways of thinking... which is something that philosophers are supposed to do.  Taleb manages to challenge you to think without sounding like a snob... just a wee bit arrogant and self-important at times.  I enjoyed this book and my next one will be one of his earlier books...  Fooled by Randomness.