T2W Trading Community

The Proof is in the Trading

In my recent TASC article, “The Hunt for Superior Signals – Two Moving Function Hybrids” (September 2005), I recommended that the Moving Slope Rate of Change (“MSROC”) be a part of every trader and researcher’s toolbox.   A quick look at this indicator demonstrates its value:

As you can see, the moving slope rate of change is remarkably smooth.  Instead of using just a change in price to calculate slope (as with rate of change or “ROC”), we use the slope of a least-squares line to calculate MSROC.  [Instructions for MSROC calculation are at the end of the article.]  Least-squares lines are notoriously smooth.  This smoothness is a major asset, as it minimizes the need for further qualification or subjective input, giving fewer “false signals”.  For example, from 1993 to early 2005, you would have observed the 25-day ROC of the close of the S&P 500 signaling 1493 changes in direction, whereas the 25-day MSROC signaled only 492 changes in direction.

Graphs look smart and anecdotes make good copy; but the real proof is the trading performance.  There was not enough room in the TASC article to present further proof of MSROC’s value.  Allow me to illustrate a previously-documented trading system and its improvement by switching the signal mechanism from a standard momentum calculation to one using the Moving Slope Rate of Change. 

One of the favorite trading systems appearing in newsletters is one in which a 12-month ROC of the NASDAQ Composite is plotted against the 12-month moving average of that ROC.   The mechanics of the system are quite simple; a crossover of the 12-month ROC of the NASDAQ Composite close against its 12-month moving average will trigger either a buy (if crossing up) or an exit (if crossing down).  Since 1987, this system which was either long or out of the market, had 14 buy signals, 11 of which were winners.  That profitability and the fact that the system rules are uncomplicated is the tease.  However the system would not be suitable for a professional who has to be mindful of daily drawdowns, as it is based solely on end-of-month data. 

That is, the trader is to wake up on the last trading day of the month, make a decision and then fall asleep until the next month-end, regardless of what happens in the interim.  Trading the month-end system as proposed is not impossible, but creates huge professional liability.  For example, how does a professional trader in a litigious society explain that regardless of what is happening in the world, he can only make decisions and effect trading on the last day of the month?  Those same constraints, minus the fear of litigation, affect individual traders as well.

Should one make the documented system realistic by using daily data, one would plot the 252-day ROC against the 252-day moving average of that ROC (since 252 trading days usually constitute a year).  You are still comparing an annual rate of change to its annual moving average, but you now see the differences on a daily basis.  When viewed in this realistic light, the number of trades increases to 45 and the winning ratio drops from 78 percent to 42 percent. As they say, sometimes realism hurts.