I liked this post about the limitations of models in understanding the market.
Funnily enough, I was thinking along the same lines in the pool this lunchtime. There is no reason to expect that it is possible to create a model that will allow you to predict any market indefinitely. Markets focus on, or are driven by, different things at different times. Even a model that incorporated the market’s shifting attention would not be able to accommodate totally new drivers (Greek bond spreads, for example) that were not part of the original model, and even if it could it would likely not be good enough at understanding how the market’s focus is changing.
A key part of understanding the markets is knowing what is driving price changes at any given time. Because the drivers can change, this exercise is not fundamentally about having a good model; it is about having a sense of the kinds of things that drive markets and the kinds of things that do not, and drawing analogies between past episodes and current market action. This kind of understanding can only come from experience (and wide reading — reading is accelerated experience).
Of course, it has to be real, deliberative experience, a constant effort to understand what is going on and why. Just hanging about and reading the news won’t cut it.