The problem with most technical indicators is lag. By the time a moving average or RSI confirms a turning point, you are already well past the optimal entry. For a mean reversion or cycle-based trader, this is not a minor inconvenience – it can be the difference between a profitable strategy and a guaranteed loser.
John Ehlers has spent decades applying digital signal processing to financial markets. An engineer by background, he brought the mathematics of DSP – the same field used in audio processing, radar, and communications – into trading indicator design. His goal is always the same: eliminate lag, identify cycles, and predict turning points before they happen rather than after.
In this episode, John explains why indicator lag is such a serious problem for short-term traders, how exponential moving averages distort signals in a specific and exploitable way, and how he constructed a new indicator by computing an EMA in reverse – from right to left – to cancel out those distortions and create an oscillator that responds to market turning points faster than anything built on traditional approaches.
Watch the full episode below, then read on for the complete breakdown.
Why Cycles Matter in Trading
John’s foundational observation is that most stocks and index futures contain a persistent cycle of approximately 20 trading days – one calendar month. Within that cycle, price will spend roughly 10 days moving in one direction and 10 days moving in the other. Understanding this structure is what makes cycle-based and mean reversion trading viable.
“Most technical analysis is reactive,” John explains. “If you look at a moving average, it’s telling you what has happened in the past. When you’re trading, you’re trading off the right-hand side into the future. You need a model that is not reactive but predictive – and that’s where cycles come in.”
The goal is to identify where a cycle is in its current phase and anticipate the turning point before it occurs, rather than waiting for confirmation that it has already turned.
The Lag Problem: A Concrete Example
John illustrates the problem with RSI. A typical RSI calculation introduces three bars of lag. Standard practice adds another three bars waiting for the indicator to confirm its turning point, then entry on the next bar after that. Total lag: seven bars into a ten-bar move.
You are past the 70 percent point of the move before you enter. “That’s a guaranteed loser,” John says. “The process is instead of waiting for confirmation, you really need to predict when that turning point is going to happen. That cancels out your computational lag.”
The same problem applies to moving averages, with longer periods compounding the lag. A 200-day moving average is trend-following territory – useful for trend context, but useless for timing short-term turns.
How the Reverse EMA Indicator Works
John’s innovation starts with a known property of exponential moving averages. An EMA applied from left to right introduces distortion: longer waves are delayed more than shorter waves. This uneven delay is typically seen as a problem to be minimised. John realised it could be exploited.
The insight: if you apply an EMA from left to right, and then apply a moving average of that EMA from right to left, all distortions and lags cancel out. The result would be a perfect, zero-lag indicator at all frequencies. The problem: you cannot compute a right-to-left EMA in real time.
John’s solution: convert the infinite impulse response of the EMA into a finite series by truncating it where the coefficients become negligibly small (within a few percent of zero). This makes the right-to-left calculation realizable in real time. The reverse EMA can then be subtracted from the original left-to-right EMA to create an oscillator.
“By making it a finite impulse response, you can place it at the right-hand side of your chart, reverse the characteristics of the EMA to go from right to left, put in enough lag to make it realizable – and you have a calculated reverse EMA,” he explains.
What the Oscillator Reveals
The resulting oscillator is parametric – its sensitivity can be adjusted by changing the EMA fraction, just as you would change the period of a standard EMA. A faster setting responds quickly to short-term cycle turns. A slower setting identifies longer-term momentum or trend conditions.
John’s preferred application is mean reversion and cycle trading:
- When the oscillator reaches a valley and its rate of change crosses from negative to positive, that is a buy signal.
- When the oscillator reaches a peak and its rate of change crosses from positive to negative, that is a short signal.
- The entry should be made when the rate of change crosses zero – by that point, you should already be in the position.
The oscillator guarantees a zero-mean value because it is the difference of two calculations of essentially the same EMA. This makes it straightforward to assess maximum departure from mean and time entries and exits accordingly.
Why This Indicator Is More Useful Than Traditional Alternatives
The advantages John identifies:
- Eliminates lag that makes traditional oscillators enter trades too late.
- Amplifies the distortion in the original EMA – the distortion that represents market inefficiency – rather than smoothing it away.
- Responds quickly to short-wave (high frequency) turning points while preserving long-wave context.
- Works across timeframes and can be slowed down for trend-following applications if needed.
“It’s fast indication of the turning point as a result of using this indicator,” John summarises. “You get a fast indication of the turning point as a result of the reverse EMA structure.”
The Cycles Approach in Practice
John’s broader framework – cycles, DSP, and predictive indicator design – starts from the observation that casual observers of market data can see cycles present in the data, but most traders do not use them because the mathematics is unfamiliar and cycles are ephemeral.
The key is not that cycles are perfectly regular – they are not. It is that if a cycle is present in recent data and you understand its characteristics, you can extrapolate it far enough into the future to take a position. “Your hope is that the cycle will continue long enough for you to take advantage of it,” John explains. That is all cycle trading requires: not perfect prediction, but enough persistence to trade profitably.
Related episodes
- John Ehlers on indicators, DSP, MESA and cycles
- David Aronson on indicators that identify market regime
Get the show notes & transcript
Want more on trading indicators and quantitative research? Subscribe to the Better System Trader podcast for weekly interviews with the world’s top systematic traders and quantitative researchers.
