One of the things to keep in mind about our performance results is that the numbers could be made to look even better via optimization.  Optimization is the process of running variables from a given low to a given high, and assigning an increment so you can see all the possibilities in between.

We routinely display nine markets, and the methodology is the same across them all.  We don't do that for “best results,” but rather to maintain some reasonable degree of robustness.  If we were to custom fit numbers to markets �" say, shorten the 40-day moving average rule to 20 days in the currencies or change the 2-day-5 day-rule to 20 and 50 days, we could observe a significantly higher outcome, which may or not mean a lot.  With each individual market, specific parameters will show eye-popping returns.  It should not surprise us that this will almost certainly be arbitrary and over-optimization can therefore become a way of fooling yourself with random results.  You're actually painting a perfect portrait of the past that will have no value in future profitability. 

Obviously, we want to avoid that.  We're here to make money, not to churn out exciting looking numbers.  Granted it's a balancing act, and also a source of debate among mechanical system experts, as I found during the course of writing Market Beaters.  Some maintain that treating all markets the same during all time frames is the surest way to avoid data-mining.  They insist that the numbers they developed years ago have held up over time and they have no intention of re-inventing the wheel.

Others acknowledge that markets do have some particular characteristics.  The indexes, for example, were widely believed to be significantly dissimilar to the commodity world at large.  Many technicians re-optimize periodically, once a year or so, in recognition of the fact that markets do drift subtly over time.

Again, this can present a slippery slope, but in our “live-as-it-happens” quest to construct a day trading guide, we'll have to do something in the way of comparing and contrasting.  Years ago, I learned not to merely find the best numbers in each market and go with them (to the tune of a few tens of thousand of dollars lost).  There might be some useful middle ground though.  If, in a broad sense, we accept the idea of specific market characteristics, we might be on safe ground optimizing according to sectors.  We've been looking at three of them �" currencies, bonds and indexes.  Tomorrow, we'll see how individual sector scrutiny changes our original results.

P.S. I'd love to receive feedback from you. Please leave a comment or discuss the article by clicking on "Make a comment on this article" below.

Art Collins is the author of Market Beaters, a collection of interviews with renowned mechanical traders. He is currently working on a second volume. E-mail Art at artcollins@ameritech.net.