By: Dean Hoffman

To the new trading systems developer one of the most exciting things to play with is optimization. Optimization is using the power of the computer to look at every possible sequence of parameters and rules, and then using only those rules and or parameters that have worked the best. With enough computer crunching power, it is possible to find systems that perfectly “predicted” the past. We can run number crunching PC’s on automated routines and have them analyze billions of bits of data while we are sleeping! Many traders do this long enough and later “discover” the holy grail of trading systems. They jump into the markets with their new super predictive algorithms only to find they fall apart in real trading!

Trading Systems Optimization Failure

“What happened?” they ask. The answer is that what they created was likely a system that was a statistical coincidence (known as a “curve fit”). Curve fitting is where you force a system to conform to historical data. The problem is that the markets will behave much differently moving forward; therefore, a “perfect” trading system could be useless. For example, your computer finds the perfect dates historically to have bought and then sold the market. These dates are likely coincidental and have no future value, yet sometimes people will base a system on them. This is a clear example; however, most curve fits are some complex form of this basic concept.

Let’s look at another flawed example. Assume we wanted to optimize nickels that were most likely to land on heads. What we could do is flip millions of nickels and only select those that landed on heads. Then, we can take those remaining nickels and flip them again, once again only choosing those that land on heads. We could repeat this process over and over again each time only choosing those nickels that land on heads. At this point, we might conclude that we had narrowed down our nickels to only a small handful that were “optimized” to land on heads. We could then go out and wager large bets with those nickels putting all our money on heads. We would quickly make a fortune, right? WRONG!

We would quickly lose our money. Those nickels were not optimized for heads; they always had 50/50 odds. What might have confused some is that they thought they had found predictable nickels. All they found was a statistical coincidence!

Trading Systems Optimization Reality

Because there is so much data, and so much computing power available, these kinds of errors find their way into trading systems all the time. When developing a system it is imperative to avoid optimizing as much as possible. You need to find NON curve-fit, robust systems. There can be a place for some types of optimizing, but it must be handled correctly.

We design all our systems in a robust way that we feel avoids the optimization pitfalls of so many other systems.

Feel free to email or contact us with any questions or comments on this subject.