Taleb’s Black Swan
Nassim Nicholas Taleb The Black Swan: The impact of the highly improbable Penguin, Revised 2010 (Original Random House 2007)
I first read this in 2007, hoping that it might help avoid a financial crisis. I found it full of good stories but, as so often with popular books, I found it difficult to extract a clear line of reasoning and gave up before the end. However some colleagues and friends noted that it was still being referred to as insightful, and even bought me the second edition. So I have read it.
On first reading the assumption seemed to be that Black Swans could meaningfully be assigned probabilities, and that their value was low. The second reading has a substantial postscript that confirms my reading. As a mathematician, the assumption seems to be that everything is precisely measurable, which seems a ludicrous view. But reading beyond where I got to in the first edition I find that while Taleb says many silly things about uncertainty, in the end his argument doesn’t rest on them. As with so many popular writings Taleb seems compelled to write about things that his audience will think important, but only to take care about those things that he actually regards as pertinent.
The second edition introduces a quadrant diagram, of the kind that was in widespread use prior to 2007, making distinctions between circumstances that should affect how one views decision-making. In effect he distinguishes between tame and more or less wicked uncertainties and between tame and wicked impact, noting a problem when both uncertainty and impact are wicked. This is the realm of Black Swans. Taleb’s book is mainly about characterising these wickedness. Many people seem to ‘get’ his conceptualisation, so it currently seems preferable to more technical attempts (such as Keynes’).
As a mathematician, I couldn’t help notice that – like many thoughtful practitioners in this area – he has a downer on mathematicians. What I hadn’t seen before is something of an explanation. Apparently, certain types of people (‘nerds’) are more likely to become mathematicians, and also more likely to over systematize. So the problem is that they are systematizers, not that they are mathematicians. Also, there is no reason to think that the mathematics as such is at all to blame: indeed, I would argue that it has key parts of a solution.
Taleb provides an example involving coins, that I comment on. I think that many mathematicians work in environments where they are expected and even coerced into behaving like nerds. So the problems are not because they are mathematicians, but may be despite their being good mathematicians.
Taleb’s main suggestion for handling his doubly wicked problems is to arrange things so as to remove the wickedness of impact, so making it only a singly wicked problem. This seems to me to fit some good prior practice, and is good advice – so far as it goes.
In a footnote in the prologue Taleb talks about recursive uncertainty, but does not develop the idea further. Someone should. The financial and economic crises from 2000-2010 seem highly recursive, and it is not clear that anyone or any group of people would have been able to simplify the impact on the global economy. Moreover, it seems to me, that many people are not faced with a problem that is ‘out there’, but are a part of their problem. For example, a large-scale investor will impact on the markets they are investing in. Government policies influence the economies that they are monitoring, parents’ behaviour affects their children, and so on. So, to me, ‘recursive Black Swans’ merit a book all of their own. While Taleb provides extensive links to prior work, I would also appreciate more links to more of the works on my blog.