Moving Slope Rate of Change: Proven Value

The reason for the poorer performance is that the indicator chosen (rate of change), despite being smoothed with a humongous 252-day moving average, is still erratic.  But if you switch from using a 252-day ROC to a 252-day MSROC, you will acquire smoothness while eliminating two-thirds of the trades.  That is, by using MSROC you drop down to 15 trades (10 of which are winners) and get to operate the system on a daily basis, and improve your trading statistics practically across the board.  Please see the statistics in Table 1.

Testing the month-end ROC system and our two tradable variations (daily ROC and daily MSROC) on the Dow and S&P 500 market indices showed similar improvements.  However U.S. stock market indices have high correlations and the fact that a successful NASDAQ trading program also worked for the Dow and the S&P 500 does not prove the system is robust.  To gain comfort with our approach we therefore ran the identical systems on other datasets representing a broader sample of market activity.  The profitability improvement on all tested datasets constituting a diverse range of markets is presented in Table 2.

The worst performer here was the NIKKEI, but the proposed strategy was a long-only program, and during the test period the NIKKEI experienced a prolonged downtrend.  Note that this example was not anecdotally selected, but a published system that we have improved by a simple change of the yardstick.  This author does not recommend the published system, only that a trader enamored with that system consider changing the momentum measuring device to the MSROC.

Now I don’t have to tell this trading audience that market prices are fractal.  Which means that if a market trading tool works on annual or quarterly data, it most likely works on daily or intraday data also.  If you are using rate of change (ROC) or relative strength (RSI) as a momentum tool, you ought to look at using MSROC to improve your results. 

Most stand-alone software products (including Excel®) have the ability to fit a linear regression line through say the last 25 closing prices.  Calculate the vertical (price) difference between the beginning and ending values of that line, and divide it by the horizontal change (the number of days, i.e. 25).  That’s the slope of the line for that 25-day period.  Divide that slope by the beginning value of the line, and you will make your indicator price-independent.  Perform all of the above over successive 25-day periods, and you will make it moving.  The price independency makes the MSROC for one market comparable to that of another – which gives you a smooth and timely ranking tool.