V.12:4 (135-145): Optimizing Momentum by Anthony W. Warren, Ph.D.

Optimizing Momentum by Anthony W. Warren, Ph.D.

Longtime readers may remember the body of written work from Stocks & Commodities contributor Anthony W. Warren during the early days of this magazine, in particular Fourier transform and triple exponential smoothing (TRIX). More recently, he has written about data filtering for trend channel analysis. This time, Warren explains how to identify the cyclical component to price data and tuning WAMI, a momentum indicator that he has designed, to indicate buy and sell points.

One of the oldest and most popular technical anal ysis methods is using a momentum indicator for timing trades. Typically, momentum is measured either as a di fference between a current price and a past price value or as a ratio of a current price and a past price. The former method measures the absolute rate of change of a market in dollars per period, whereas the latte r, the ratio, measures relative rate of change in percent per period. If data filters such as moving averages and e xponential moving averages in the momentum indicator are used, then a number of momentum indicators are alread y in use: moving average conve rgence/divergence, Herrick pa yoff index, price rate of change ( ROC), price oscillato r, triple exponential smoothing oscillator ( TRIX) and so forth.

These indicators are popular because momentum indicators will often identif y important market bottoms and tops. However, momentum indicators must be carefull y tuned to the market being anal yzed to avoid unprofitable trading. As a result of this caveat, momentum indicators are often used with other technical methods to improve trade reliability.

Here, we will introduce a cycle-based method for momentum trading that signals important market moves but avoids unprofitable trades by optimal parameter tuning to the current market. This momentum method is a three-step process:

1. Identification of the most important cycle frequencies using Fourier (FFT) based spectrum analysis

2. Optimization of the Warren momentum indicator (WAMI) filter parameters to the identified market cycles

3. Application of the WAMI indicator and a trigger function to find important buy/sell trading points.

This process provides a practical method for selecting momentum parameters optimized to current market conditions. However, the method requires learning some advanced cycle analysis concepts and the use of cycle-based tools to implement steps 1 and 2. Here, I will explain the cycle analysis concepts and the use of spreadsheet/technical analysis software to implement the first two steps. In another article, we will demonstrate the full process by applying it to a number of diverse markets. The trading examples use weekly data, but the method is easily modified to use hourly, daily, monthly or other sample intervals.