This produced a win-loss ratio leaving little doubt that something was going on beyond mere chance. Fifteen major markets were tested over a period spanning 1989 through 1997. Out of 16,350 trials, 53.7 percent conformed to the "fade-the-close" theory. That may not seem like much compared to what the pit wizards might enjoy, but keep the following in mind.

1. The concept was raw, robust and simple â€" not jerry-rigged in any way.

2. Bookies enjoy roughly a 53 percent edge and Vegas settles for even less on many of their frontline bellwethers. The viability lies in sheer number of trials.

3. No one is advocating a mechanical system or any other kind of system here. The idea is that if you're trading anyway, you may be interested in knowing whether or not your position is conforming to the bias. Knowledge never hurts. (Well, maybe we should say "almost never.")

The specific above methodology will not be incorporated in the foreseeable future â€" it's gone bad lately. I didn't intend to tout it as an actual program. It's a mere illustration of the kinds of things I'll be exploring. At any rate, I'm neither surprised nor disheartened by the breakdown. My experience is that markets do drift, and part of the optimization process is periodically adjusting for that fact. The "nothing is eternal" point is arguable; some people swear they've invented the wheel for all time, pure and simple. What I've found, though, is that when one looks at thousands of trials over a daily field, the last two years or so is pretty much where the relevant focus lies.

I admit to the following personal conviction: Simple Is Best. My success as a mechanical system trader has â€" almost without exception â€"  hinged on very straightforward, unadorned, uncomplicated concepts. Mathematically, I'm in primary school compared to some of the math scholars that offer TradeStation online programming assistance, but I have to wonder if a significant number of them are better theoreticians than traders. I don't say that with rancor. It's borne out of personal experience. I had my second most successful day on Black Monday off an entry signal generated by subtracting the previous daily range from the low. That was my happy stop entry. My biggest windfall was another S&P shortâ€" this time, entering on momentum off an opening toward the direction of the previous five-day average. While I wish I were a better programmer, there isn't much correlation between my more esoteric ideas and my dumb ones. If anything, there's been a reverse relationship â€" not exactly a burning incentive to learn C++.

The idea of this feature is identify such basic building blocks. Each unit can be assigned a plus one-minus one value corresponding to long/short. The values can then be combined into something that will hopefully offer a significant combined whole. The ultimate "recommendations" â€" (quotes very deliberate) â€" will come from such compound ideas. It will be strictly based on past stats. And luckily for you, my abysmal trading judgment will have no part in the process

In the next few days, I'll show some other biases and hopefully immediately after, I'll start posting probable up/down daily trajectories. A continuous update of win/loss percentages will also be featured.

File all this under the "for-what-it's-worth" heading. Stats like these would at least give me a "hmm…check this out" pause. Hopefully, enough of you are similarly inclined. Let's see how it goes.