Take the virtuous advice that trading success hinges on "following your process." Please. WTF does that mean? Here's a process for you: I drink Turkish coffee every morning and then examine the coffee grounds at the bottom of the cup. If the grounds mostly settle at the top of the cup, it will be a bullish day in stocks and I buy at the open and sell the close. If the grounds settle at the bottom of the cup, it will be a bearish day in the market and I sell the open and cover at the close. If the grounds settle at the left side of the cup, it will be a quiet day and I'll sell volatility. If the grounds settle on the right side of the cup, volatility will expand and I will be a vol buyer. And if the grounds are evenly settled at the bottom of the cup, there's no edge for the day and I'll drink a second cup.
Now that *is* a process. But does my trading success hinge on fidelity to my process? Of course not! That's because, in the language of psychometrics, the process is reliable but not valid. It is repeatable but random. Being process driven is necessary, but not sufficient. Before one waxes poetic about following a process, it helps to define a process worth following.
Or let's take the platitude that one shouldn't try to predict markets but instead should listen to markets and follow their lead. I don't know what that means. Does it mean that you are supposed to naively extrapolate the last X bars on a chart and blindly assume that they will continue their pattern? Does it mean that you impose expectations of momentum and trend on every market regime? Does it mean that you utterly lack backtesting skills and are bravely turning that into a horse-whispering virtue? Whatever.
So here's a great experiment that anyone can conduct. Define a trading system that takes the last X bars and extrapolates from them to the next X bars, buying the X+1 bar open when the extrapolation is positively skewed; selling the X+1 bar open when the extrapolation is negatively skewed; and standing aside when the extrapolation displays no directional bias. That way you'll always have a replicable process *and* you'll be following what the market is telling you.
Just as a lark, I tried the experiment with historical data, buying SPX when the percentage of stocks above their five-day moving average was above 50% and selling SPX when the percentage of stocks above their five-day moving average was below 50%. The data were for SPX stocks specifically, going back to 2006, and the holding period was for the next five trading days. Buying strength gave an average five-day return of -.07%--a small loss. Selling weakness gave an average five-day return of -.44%, a considerable loss. The average five-day return over this period was a gain of +.15%. In other words, having a robust process and following the market's lead has ensured losses in both rising and falling market environments.
Interesting: the track record of the received wisdom is so poor that it's promising.
Further Reading:
Three Market Idiots
The Near-Perfect Market Indicator
.
Posts
