Ehlers indicators: signal processing applied to price
John Ehlers, an aerospace engineer, brought digital signal processing (DSP) to trading: instead of treating price as a line, he treats it as a SIGNAL made of cycles plus noise, and applies engineering filters to separate one from the other. It is among the least-known corners of technical analysis, but with serious mathematical grounding. Here are his flagship tools, what they're for and when they fail.
By the TradingCalculator.Pro team · Updated on · About us
Start your 7-day free trial →What you will learn
The idea: price as a signal
Classic averages face a dilemma: fast ones are noisy, smooth ones lag. Ehlers attacks that with DSP — the same maths as radar and audio: he models price as a dominant cycle plus noise and designs filters that strip the noise with the least possible lag. The goal isn't to guess, but to see the clean signal sooner.
Fisher Transform
Most indicators assume prices are distributed like a Gaussian bell, and they aren't. The Fisher Transform "stretches" the data toward that bell, with a practical effect: turns become much sharper and more pointed, giving earlier, clearer reversal signals than a normal RSI or stochastic.
MAMA / MESA (adaptive average)
MAMA (MESA Adaptive Moving Average) is an average that changes its speed with the market's cycle: it speeds up when price moves and slows down when it goes sideways. MESA (Maximum Entropy Spectral Analysis) is the engine that measures that cycle. The result: it tracks turns closely and avoids much of a normal average's lag.
Dominant cycle and Sinewave
Ehlers measures the "dominant cycle" — roughly every how many bars the up-and-down repeats — and builds the Sinewave from it, showing the PHASE of the cycle (where the wave is) instead of price. Its big use is warning when the market is in cycle mode (range) or trend mode, where the indicator steps aside and stops giving reversal signals.
Filters: SuperSmoother and Roofing
These are Ehlers' "engineering filters". The SuperSmoother smooths price with far less lag than a classic average; the Roofing Filter strips both the very short-term noise and the very long-term drift at once, leaving only the useful band of frequencies. With them, many old indicators (RSI, stochastic) improve when fed already-filtered price.
Limits (honesty)
These are advanced tools, not magic shortcuts. Market cycles aren't fixed: they change length and sometimes vanish, and in strong trends cycle indicators fail (which is why Ehlers himself adds trend detection). Getting value from them requires understanding what each filter does; used blindly, they're just another pretty indicator.