The past 10 days serves as an excellent illustration with regards to the proper selection of significant variables going forward.
As the popular press has stressed, quantitative funds who have modeled the markets based on the historical relationships (be they correlative or factually dependent) between sets of prices have suffered some unsettlingly correlated losses.
These numerical relationships are, to put it mildly, "under stress". There are certain members of the press who see the results and quote some amporphous quality called "trading wisdom" and harken back to open-outcry systems, which by implication must be better during these conditions. But what do they mean by this?
Simple: The ability to utilize a deep-thinking form of educated common sense. It is the ability to "connect the dots" using metaphorical and narrative arguments as opposed to data-driven statistical or combinatory reasoning.
And, giving a nod to Soros, any system with thinking participants will produce self-sustaining bubbles as competition will chase profits and drive down the opportunity costs of prudent restraint..."No covenants on your Sub-prime backed equity tranche mortguage securities?? Who are we to argue?? Nobody else is either!! Give us the fee and see you later."
Quantitative analysis is no different. The investment industry has been flooded with brilliant mathematical minds all employing similar (or worse, functionally indentical) trading algorithms. Many have spectacular historical results which bred competition and duplication, using the same arbitrage arguments, the same financing vehicles to fund the trades and the same amount of leverage to make the crumbs on the floor look like loaves of bread.
And then someone yells "FIRE!!!" in the factory with one emergency exit, and the devil takes the hindmost.