Of models and Black Swans

Image via WikipediaInteresting discussion on Friendfeed about Black Swans and models (both scientific and financial). Teed off by a post by Ashutosh Jogelekar, this discussion is an interesting one. Let’s start with something Ashutosh writes

But let’s face one big truth about models; most of them in fact are not designed to represent “reality”. In fact models don’t care as much about causation as they do about accurate correlation and prediction. While this may sound like shooting ourselves in the foot, it often saves us a lot of time, not to mention philosophizing. We use models not because they are “real” but because they work

Fred Wilson also has an interesting piece about models and financial markets where he quotes Umair Haque

We can’t fix today’s problems unless we change yesterday’s rules. But economists — and the models they rely on — are bounded by yesterday’s rules.

As I see it, in science a model is the best theoretical representation of our knowledge. The model could be statistical, physical, mathematical, or perhaps purely hypothetically derived to explain a series of observations. There was a time when I didn’t pay attention to statistical models, but as Ashutosh points out, they work, so I’ve come around on that. So given that our aim as scientists is to design experiments that break the underpinnings of the model, which Black Swan events are supposed to do. Since these events improve our scientific knowledge, they are a good thing, and we can go back and refine our models. We have the checks and balances in place (pre-clinical studies, clinical trials, etc), to make sure that flaws in our knowledge do not end up creating havoc, something the financial markets don’t have in place. Since scientific models are a moving target by the very nature of the scientific method, there is no such thing as yesterdays rules. It is our best representation of the facts as we know and understand them. Ideally (hopefully), no individual biases, and no psychological variables either.

And when we can’t find any Black Swans or break our models they become law. That’s how science works. Models are good, and, sorry Chris Anderson, they are not going any where.