A Moving Average type of indicator that attempts to identify trends and trend changes.
Overview
Developed by Dennis Meyers and described in his article entitled “The Japanese Yen, Recursed” published in the December 1998 issue of Technical Analysis of Stocks and Commodities magazine, the Recursive Moving Trend Average is in mathematical terms a “Recursive Moving Polynomial Fit”.
Meyers describes this technique as requiring only a small number of past values of the estimated price and the current (today’s) price to predict the next (tomorrow’s) price whereas the Least Squares Polynomial Fit (which is very computationally intensive) requires large numbers of past price points.
- “The system is basically a moving average crossover system used to identify changes in trends. It compares a Recursive Moving Trendline (which attempts to predict tomorrow’s price using an exponential moving average slope calculation) to an Exponential Moving Average of the price plot. A crossover of the two lines triggers a buy or sell signal. It might make a useful addition for those who use a linear regression-based system.
Overall, it seems to identify trends and trend changes quite well, though with the weak money management capabilities of MetaStock, it is difficult to fine tune the system to avoid whipsaws during retracements.” (Trader.Online.pl)
The Recursive Moving Trend Average on the AUD/NZD Daily chart showing April-Nov 2010
Source: VT Trader
Interpretation
Meyers suggests creating a simple “trend oscillator” which he refers to as the “tocs” which is the difference between the Recursive Moving Trend Average and an Exponential Moving Average of the same n-periods.
He mentions looking for changes in the price series that are above the normal noise fluctuations indicating that a potential uptrend or downtrend has started. This is accomplished by examining the plot of the tosc.
A potential buy signal is generated when the tosc’s value crosses above the “dup” level and a potential sell signal is generated when the tosc’s value crosses below the “-ddn” level.
During Meyer’s testing, the optimum “dup” and “-ddn” levels were calculated using optimization, walk-forward testing, out-of-sample testing, and averaging of all testing results.
For additional information, including the mathematical formulas for the Recursive Moving Trend Average and TOSC, please review Dennis Meyer’s whitepaper entitled “The Japanese Yen, Recursed” published in 1998, which is available from his website.
