Well, I've finished reading The Black Swan. As I mentioned, the first half of the book is basically looking at all of the ways people make bullshit arguments. The second half of the book then quickly and succinctly demonstrates that Gaussian statistics are, in many cases, bullshit.It's interesting to me, that when the book was reviewed on the Freakonomics Blog, the questions that the readers were asking had to do with what the author's investment strategy was. Put in simple terms, as someone there attempted, he thinks that unlikely event futures are under-priced.That, however, seems to me to completely miss the point of his book. At least, that could be true for reasons that have nothing to do with his point.The point of his book is that we have learned about randomness and probability from the bell curve and normal distributions, and that many of the most important things in our lives are not normally distributed.It's easier to explain with examples. If you're dealing with body weight, taking a random sample of 1000 people then adding the person with the highest individual body weight on the planet, you will not affect the average very much. It will go up, but not monstrously. On the other hand, take the same random sample of 1000 people, then add Bill Gates. The average wealth in the room will skyrocket.So the basic argument he's making is that there are some variables, like weight, that are normally distributed, and that there are many other variables, like wealth, that are not. Further, if you use normal distributions to understand probabilities in the realm of this second category of variables, you are setting yourself up to get screwed. Because there will be lots of "corroborating evidence" to tell you that your model is working, right up until it doesn't.The book essentially argues against the mis-application of a mathematical representation of reality in places where it doesn't apply.He suggests that if where normal distributions don't apply you model unpredictability according to power laws (where probabilities do not reduce exponentially as the observation moves away from the average), you will make better decisions regarding extremely unlikely events, both positive and negative.Let me talk for a minute about the author's style. The book was written in a conscious effort to avoid categorization (or as he calls it, platonification). Parts of it are fiction, parts of it are biography, parts of it are statistics text book. Throughout, it is riddled with the names of scientists and philosophers alive and dead that the author has studied in trying to understand his topic. The bibliography to the text is simply ridiculous. For a person such as myself who's used to reading academic works with lots of name-dropping, and who has developed the ability to ignore the names where the ideas are reasonably reproduced, this may not be problematic. To anyone else, it would undoubtedly be totally frustrating.The author is also "irreverent" as one of his editors complained. Again, coming from the academic reading that I've been doing for the last few years, in academic circles this text would be considered violent, not merely irreverent, toward a number of prominent individuals with whom the author disagrees. As I said to my wife last night, he writes like a person who has spent 30 years trying to teach idiots the simplest of truths. He is frustrated, and it comes through in his writing.Third, his writing has this strange capacity to simply stop. Chapter headings and sub-headings seem of almost random placement and importance. There are only brief introductions to the four sections of the book, none to the chapters, nor to the chapters recap themselves effectively. On a larger scale, there is simply not as much in the second half of the book, outside of a technical explanation of the problems with the bell curve, as there is in the first.This last isn't a major complaint, and it's consistent with one of the many points he makes. It's not so important to have some alternative that works well as it is to know that the method you're using now is bullshit. And proving that something is false takes a lot less time and effort than proving it's true.Now, on to the interesting question for us political types: What does his book say about politics? The answer is nothing, explicitly. The question that immediately comes to mind, however, is whether or not the things that we're measuring when we poll people (be it in an election or in the statistical sense) are things that belong to the category of variables that include weight, or the category of variables that include wealth. The question is whether or not the way people make voting decisions necessarily fits into the models that have been created for understanding it.My intuition is that they do not. I have not thought about it much further than this, but here is the basic idea: Voting is driven by opinion and perception, which are not binary, and which are not under any physical restriction that would make them vary according to a bell curve. The idea that there is a political centre, therefore, like the thick part of the bell curve, may be totally unfounded. The political centre may be simply a manifestation of the fact that we've got this artificially one-dimensional political spectrum in our narrative, and a limited number of political parties attempting to grab segments of popular opinion, making it easier to rank them on a single spectrum.What could that mean? It could mean any number of things. Take, for example, the possibility that the method for the Conservatives to take a majority in the house might not be for them to appear more liberal, but to find some ultra-conservative glut of non-voters and motivate them to come out with ultra-conservative policies. That would be a Black Swan.The perhaps more interesting aspect is whether or not the ideas in The Black Swan can be used as a criticism of political polling, and the decision-making that takes place on the basis of Gaussian statistics applied to human opinion. I'm not there yet. I just don't know. I kind of hope some political bloggers with knowledge and experience in such things (Cough.. Dan ... Cough) will take up the issue and talk about it more. I'm going to think about it more.It's perhaps fitting that after having finished a book that absolutely rails against many social sciences, I'm off today to convocate with an honours degree in one of those social sciences (political science) with a minor in another (economics is easily the biggest offender that The Black Swan attacks). I will therefore accept my degree with an undoubtedly justified and healthy dose of humility.As for The Black Swan, if you're a person who has been taught statistics and has applied them to a social science or finance, you need to read this book. Otherwise, the question is whether or not the above discussion bored you to death. If it did, don't bother getting the book. But make sure that the people making decisions with your money have read it.