Voss Predictive Filter

Most "predictive" indicators are a smoother relabelled - they still only know the past, and any apparent lead is really just less lag than a plain moving average. Ehlers' 2019 Stocks & Commodities article "Predictive And Successful Indicators" describes an actual negative-group-delay network, crediting the underlying design to David Voss. A narrow 2-pole bandpass filter is tuned to a chosen dominant cycle period and bandwidth, and its own recent output is combined with weights that add up to more than the current sample - a construction that produces genuine forward lead against a stable, dominant sine-wave-like cycle, not merely reduced lag. That lead is bought entirely by assuming the input keeps behaving like the cycle the filter is tuned to. The moment price does something the recent cyclic history does not contain - a break, a volatility spike, a genuine regime change - the projection is wrong in exactly the way any extrapolation is wrong when the pattern it extrapolates stops holding. This file plots both the underlying filtered line (Filt) and the projected line (Voss) together specifically so a trader can see how far ahead Voss is running relative to Filt, and judge for themselves whether the two are currently tracking a real, stable cycle or diverging. Notes and limitations - This is a predictive filter under an explicit cyclic-market assumption, not a forecast of price. It leads a stable, dominant cycle by design; it has no mechanism for anticipating a break, a news-driven jump, or any move outside the cyclic pattern it is tuned to, and will project confidently and wrongly across exactly those events. - The bandpass filter and the projection step are both recursive and reach roughly InpPredict x 3 bars into history, so the first stretch after a fresh attach (or after changing InpPredict) is a warm-up seeded at zero, not a true reading. - Because every value is computed once from closed-bar history with no lookahead into future bars, the plotted lines themselves never repaint - the risk here is not repainting, it is the projection being wrong when the cyclic assumption breaks, which is a different and more honest failure mode. - InpPeriod is a trader-set input, not something the filter measures - it does not adapt to the market's actual cycle length the way this catalogue's Cycle Adaptive RSI or MESA Stochastic do. Getting it wrong produces a plausible-looking but meaningless line, not an error.

Indicator · MT5 · v1.0 · Free

Tags: Moving Average · Alerts · MT5 Indicator · MQL5 Source Code

MetaTrader 4 & 5 indicators · PolyMaria