In "Inside the Black Box", Rishi Narang lifts the veil away from the secretive world of quantitative trading, explaining in clear, easy to understand English the key concepts that drive so much of hedge fund trading today. Instead of becoming bogged down in the inscrutable math of the quants. Mr. Narang expertly navigates the reader through the key theoretical concepts behind systematic trading and in the process provides multiple morsels of wisdom for the retail trader to consider. I will come back to many of the ideas he explores in the book in the next few months, but this week I want to focus only on one.

Mr. Narang shows how even the simplest systematic strategy can carry unintended risk. In quant world "relative alpha" - where a trader decides to go short "expensive” instruments and go long "cheap" ones - is a very common strategy. However, if the trader were to follow it blindly, according to some valuation algorithm he would be in danger of making unintended sector bets. For example, if the computer screen revealed that software companies were "cheap" and oil refiners were expensive, then an unfiltered relative alpha strategy whose intended function is to bet on a variety companies in order to diversify risk would in fact be making one implied bet of high tech vs. oil. Mr. Narang shows how the biggest blow up in hedge fund history, LTCM, was guilty of precisely such a flaw. Although traders at LTCM put on thousands of trades ranging from Argentine bonds to US Treasuries they were all essentially different variation of one trade called the convergence trade. When that trade went against them, all their positions lost money simultaneously.

I still remember several years ago being approached by someone at a Trade Show in Las Vegas who proudly told me that K and I sucked making a puny 100-200 points a month when he was making more than 500 points a week. "Really?" I asked a bit surprised and impressed by his claim. "What kind of trades are you doing?" He proudly whipped out his account statement and showed me that he was long USD/JPY, long GBP/JPY, long AUD/JPY and long EUR/JPY. This was in the heyday of the carry trade so naturally the guy was minting money. Needless to say I never heard from him again when the carry trade crashed and burned at the end of 2008.

We always stay in our trade even if data is wrong because often the overall trend saves us. Alas not today. Sometimes the only thing you can learn from your mistakes is that you have to accept them.

In currencies making this type of mistake is very common. If you are trading with a $10,000 account and have a rule that you never want to put more than $100,000 notional into any one position and you happen to be long one standard lot of EUR/USD, long one standard lot of GBP/USD and long one standard lot of AUD/USD - guess what? You are way over your limit. You are essentially making the same anti-dollar bet three times over. Although euro, pound and Aussie are not completely correlated they are correlated enough that you a very likely to incur much larger losses than you anticipated if the position moves against you. Instead of being diversified you are actually highly concentrated and should be aware of the hidden risk in your trading position.

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