What do an electronic saw, a Paris metro ticket, and modern anesthetic equipment have in common? They are all designed for failure. The electric saw comes with something called a dead man switch which turns the saw off instantly the moment a grip lets go. The Paris metro ticket can be inserted into a turnstile face up or face down and the machine will read it either way. Modern anesthetic equipment makes it nearly impossible to mix up the delivery of various drugs, by making the nozzle of each connection utterly unique.

In their wonderful new book, Nudge: Improving Decisions About Health, Wealth, and Happiness, Richard Thaler and Cass Sunstein talk about the fact that the best, most intelligently designed products in our lives assume that human beings will make mistakes and accommodate for that dynamic. In our day to day life what can be a better example of that than the spell checker? A software tool so common and yet so extraordinary that it came make illiterate brutes like me to become well polished, published authors. Anyone who has ever had the misfortune to read my research reports before a spell checker corrected my two finger typing knows what a pathetic experience that could be.

On the other hand, take something like a credit card reader. From the parking meters on New York streets, to the movie theaters to the local supermarkets each reader has a unique way of accepting the card. In fact out of four possible ways to swipe a card through a reader exactly three of them are wrong. Furthermore, each card reader has a different “right” way of accepting the card. How many of us have spend countless frustrating minutes, trying to figure out how to get the various readers to work? The credit card reader is a horrible product that should take some design cues from the Paris metro. Its designed for cyborgs, not human beings.

When it comes to trading, so many of us make same critical mistake of optimizing for success rather than for failure. Is there anything more human, more error prone, more irrational that the financial markets? Little wonder then that fancy Ph.d.-engineered algorithms on relative valuation and reversion to the mean strategies blow up with predictable regularity. In fact, as the recent high frequency trading scandal shows, the only computer driven models that succeed in the market are the ones that cheat by front running other orders.

Every novice trader approaches the market with a mindset of an engineer, but every expert trader leaves the market with the wizened attitude of a bar tender who has seen the human condition close up. When traders approach me asking for the EXACT recipes of my strategies telling me that otherwise my ideas are all totally useless because they cannot be replicated I cannot help but laugh and ruefully shake my head. As I show in this week’s video, precision is the enemy of success in trading.

One of the things that I realized as I was reading Nudge is that our BK trading model has organically involved over time optimized for failure which is ironically enough why it has been able to succeed. We trade with trend only. We try not to chase price. We always look for flow and fundamentals to align our way and most importantly in our own version of “dead man switch” we always take partial profits the moment we go in the money. In short, instead of assuming that everything will go right we believe that a million things can go wrong and trade accordingly.

Of course, optimizing for failure does not in any way guarantee success. Saws still chop off limbs, turnstiles fail and unfortunately some patients still die from the wrong anesthetic. As to trading you can be assured that mistakes will always occur in our human driven financial markets. However, by recognizing that fact and giving ourselves the widest possible margin for error we stand a much better chance of being profitable in the end.

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