Good judgement comes from good experience and good experience comes from a lot of bad experience.

My daughter, who is still in high school, just started taking a sociology class at Hunter College. Up to now, she had exactly zero interest in the subject of social studies, steadfastly refusing all my attempts to teach her about Greeks, Romans, the Enlightenment, or any other part of Western civilization. In fact, her only source of news has been the Daily Show.

But now that she is taking this course, her interest has suddenly perked up and she is bombarding me with fresh found knowledge about Locke, Hobbes, Marx, and Weber as we discuss individual rights, the idea of surplus labor, class divisions and the ever present human desire for status. My daughter's new found enthusiasm for the subject made me realize just how much I missed studying sociology purely for it academic value, but it also made me appreciate sociology's very practical contribution to the art of trading.

When we think of trading models, most of us inevitably fall into the quantitative trap. We think that trading models are nothing more than Metatrader Expert Advisors -- some cobbled computer code that uses well worn technical indicators as its inputs to generate a buy or sell signal with a smattering of money management rules thrown in. Typically, these Expert Assistants trade 24/7 across multiple currency pairs, providing you with the illusion that you can "bank profits" without even looking at the quote screen again.

So how many EA millionaires are there? I don’t know, but in all my time in the markets I have never been able to find one. One of my earliest jobs in the currency markets involved monitoring the overnight execution of the algorithmic desk for our customers. I saw just about every idea you can imagine. From relative value to momentum breakout strategies to some of the most convoluted martingale schemes ever created. They all lost money in the end.

Why? Because when you are constructing a trading model, it is sociology, not statistics that matter. When we are trying to model the market, we are trying to understand human behavior that governs the price action, not the price action itself which merely reflects that behavior. The single greatest flaw of all the quant-driven trading models is that they often miss the proper context of their actions.

Boris Schlossberg serves as director of currency research at GFT, and runs bktraderfx.com.