D018 · TREND-FOLLOWING SYSTEM

KalmanDrift Navigator

Rocket-tracking math meets market persistence — trade the drift, skip the noise.

Concept

Most trend tools lag because they average the past. The Kalman Filter instead maintains a live statistical estimate of the market's "true" price level and its drift velocity, updating both with every bar. The Hurst Exponent (rescaled-range method) then answers one question: is this market actually persistent right now, or mean-reverting chop? Only when both agree do we act — entering on the classic pattern of a pullback below the Kalman path followed by a decisive reclaim.

The Three Pillars

🛰️ Kalman State Engine

A constant-velocity Kalman filter (Q=0.001, R=1.0) outputs a smooth price path plus a drift speed. Direction requires |velocity| > 0.05 × ATR(14).

📈 Hurst Regime Gate

R/S Hurst estimate over 100 bars must exceed 0.55 — statistical evidence of trend persistence. Below that, all signals are ignored.

🎯 Residual Reclaim Trigger

The residual (close − Kalman level) must cross from negative to positive for longs (mirror for shorts): a pullback absorbed, drift resuming.

Entry Rules

LONG — Hurst > 0.55, Kalman velocity > +0.05×ATR, and residual crosses up through zero on bar close.
SHORT — Hurst > 0.55, Kalman velocity < −0.05×ATR, and residual crosses down through zero on bar close.
Signals evaluated once per closed H1 bar; one position at a time; spread filter 30 points.

Exit & Risk Management

Stop loss: 1.5 × ATR(14) beyond the Kalman level (structure-referenced, not entry-referenced).
Take profit: 2.5 × ATR(14) from entry — reward:risk ≈ 1.67:1.
Position sizing: fixed 1% account risk per trade, lot size derived from stop distance.

Parameters

Parameter	Default	Role
KalmanQ / KalmanR	0.001 / 1.0	Filter responsiveness (process vs. measurement noise)
HurstWindow	100	Lookback for R/S Hurst estimate
HurstThreshold	0.55	Minimum persistence to allow trades
SlopeATRfrac	0.05	Minimum drift speed as ATR fraction
ATRPeriod	14	Volatility unit for slope, SL, TP
SL_ATR / TP_ATR	1.5 / 2.5	Stop and target multiples
RiskPercent	1.0	Account risk per trade (%)
Timeframe	H1	Recommended chart period

Why It's Different

No moving-average crossovers, no oscillator thresholds, no channel touches. The edge comes from combining an optimal state estimator (Kalman) with a regime statistic (Hurst) — trading only when the market is measurably trending and the trend has just absorbed a counter-move. Works on FX majors, gold, and indices on H1; parameters are scale-invariant.

⚠️ Risk Disclosure: Trading involves substantial risk of loss and is not suitable for every investor. Past performance and backtests do not guarantee future results. This material is for educational purposes only and is not investment advice. Always forward-test on a demo account before committing real capital.