Noise Elimination Technology Oscillator

Almost every smoothing method - moving averages, EMAs, even most digital filters - is a magnitude-weighted average of recent values. That means one outsized bar (a spike on thin liquidity, a news print) can swing the whole reading, even though a single bar is not, by itself, meaningful evidence of a new trend. John Ehlers' Noise Elimination Technology (NET) takes a different approach entirely: instead of averaging magnitudes, it measures Kendall's rank correlation between a window of values and time - it only asks whether each later value in the window ranks above or below each earlier one, never by how much. A handful of huge-range bars carries exactly the same weight in that count as a handful of ordinary ones, which is the specific noise problem magnitude-weighted smoothing cannot avoid. This version applies NET to price momentum (the change over InpMomentumPeriod bars) rather than to raw price, then computes Kendall's tau across a rolling InpNetPeriod window of that momentum series. Because time is monotonically increasing across any window by construction, the general two-series Kendall formula collapses to a single pairwise comparison: a pair is concordant when the later momentum value ranks above the earlier one, discordant when it ranks below, and a tie contributes neither way. The result sits in -1..+1: values near +1 mean momentum has been consistently rising through the window (an efficient uptrend in rank terms), near -1 means consistently falling, and near zero means no consistent rank ordering at all - genuine noise. Ehlers published NET in a TASC "Noise Elimination Technology" article; this is a from-scratch reimplementation of the published Kendall-correlation description, applied to a momentum series rather than raw price as Ehlers' original examples did - not a copy of his DSP source code, which was never released for this exact variant. Notes and limitations - Bounded to -1..+1 by construction (it is a correlation coefficient, not a magnitude), so it reads on a fundamentally different scale from a typical momentum oscillator. - The pairwise Kendall computation is O(InpNetPeriod squared) per bar. At the sensible default of 20 this is trivial; pushing InpNetPeriod into the hundreds will visibly slow recalculation on deep chart history. - Does not repaint - each bar's value only ever uses momentum values at or before it. - This is an original reimplementation from the published method description, not a verified match to any specific commercial NET implementation's exact internal numeric routine.

Indicator · MT5 · v1.0 · Free

Tags: Moving Average · Momentum · Oscillator · Trend · News Trading · MT5 Indicator · MQL5 Source Code

MetaTrader 4 & 5 indicators · PolyMaria