Stocks & Commodities V. 24:7 (68-71): Product Review: MESA8 by Dennis D. Peterson
John Ehlers began applying the approach of maximum entropy spectral analysis (MESA) to trading in 1978, continuously improving it through the years. Now, MESA may be the best method of measuring market cycles.
MESA8 is the flagship of the company, a set of new indicators available as a TradeStation or NeuroShell Trader plug-in. What makes MESA8 exciting is that
the cycles approach uses long-established techniques, techniques that were used to characterize and predict sunspot activity as far back as 1927, so the science behind this product has had many years of development. On top of that, developer Ehlers, the author of MESA8, was the one who first derived MESA (short for maximumentropy spectral analysis), well
known in technical analysis.
FILTER COEFFICIENTS AND DOMINANT CYCLE
The process behind MESA8 is fairly straightforward: create an adaptive filter that uses an autoregression technique to find autocorrelation coefficients characterizing two signal inputs into the filter: Gaussian noise and the difference between the output of the filter and a time price series. This creates a set of autocorrelation coefficients that “matches” the price series in a white noise environment. After finding the matching coefficients, input a spectrum of frequencies to this adaptive filter. The frequencies that pass through the filter are the ones that reflect the cycle(s) in the price series — the one with the greatest amplitude is the dominant cycle.