━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
WAVE RIDER v6.5 — Update
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Hey everyone,
First of all, thanks for the DMs since I shared v6.1. I received mostly this request: "Can you add this baseline?" Fair point. A few of the original 8 were research-grade algorithms that I personally found fascinating but that aren't exactly usable on live trading.
The baseline roster has been expanded from 8 to 15 algorithms, replacing the more exotic ones with options that are better understood and more useful even in real prop-firm conditions (I run this on NAS100 FTMO).
─────────────────────────────────────────
WHAT WAS REPLACED — and why
─────────────────────────────────────────
Out: Epanechnikov Kernel, PKF (Parametric Kalman Filter), EPF (Edge Preserving Filter)
These were the ones people flagged as "too experimental" — the Kalman variants in particular had parameter sensitivity that made them tricky to configure without deep signal processing knowledge. Replaced with more interpretable, widely-used alternatives.
─────────────────────────────────────────
THE FULL v6.5 BASELINE LIST
─────────────────────────────────────────
- KAMA — Kaufman Adaptive Moving Average
The classic adaptive MA. Speeds up during trending conditions, slows down during consolidation based on the Efficiency Ratio. Solid all-rounder, great for avoiding whipsaws in ranging markets. Still the default.
- Ahrens Moving Average
Triple-smoothing structure with a built-in trending gate — it only follows price when momentum is sustained. Very clean on higher-velocity moves.
- Regularized EMA
An EMA with a regularization term (controlled by the new CustomSigma input) that penalizes rapid direction changes. Produces a smoother line than a standard EMA while staying more responsive than a heavy SMA. Great for reducing noise without sacrificing too much lag.
- Rational Quadratic Kernel
A kernel-based smoother with heavy-tail weighting — it gives more influence to recent bars while still incorporating older price memory. Unlike the Epanechnikov version (compact support, hard cutoff), the RQ kernel has no hard boundary, making it more robust to volatility spikes. Now has a dedicated KernelRelativeWeight input.
- TRAMA — Trend Regularity Adaptive MA (LuxAlgo)
Adapts its length based on the regularity of the trend, not just raw volatility. Fast in clean trends, slow in choppy conditions.
- VHMA — Vertical Horizontal MA (alexgrover)
Uses the VHF (Vertical Horizontal Filter) ratio as an adaptive coefficient. Excellent at distinguishing trend from range. The VHF² weighting makes it more aggressive when the market is clearly going somewhere.
- Zero Lag Hull MA
Double Hull construction that cancels the lag inherent in a standard Hull MA. The result is a line that closely hugs price in trending conditions without the oscillation artifacts of a simple fast MA. Very reactive, even too much...
- Corrected Super Smoother
Ehlers' Super Smoother with a correction pass that reduces the phase lag. Two-pole Butterworth-style filtering that eliminates high-frequency noise while staying directionally clean. Underrated baseline for trend confirmation.
- Ridge Regression
Regression-based MA with an L2 (ridge) penalty. Excellent at producing stable, non-oscillating trend lines even in sideways markets. The CustomSigma input controls the regularization strength.
- Huber Regression
Like Ridge but uses a Huber loss function, which makes it robust to price spikes and outliers. Less sensitive to wicks than any standard MA. Good choice when trading instruments prone to sharp news-driven moves.
- G-Channel Midline
Takes the midpoint of an adaptive upper/lower channel built from a running max/min structure. Produces a very stable, centered line with minimal noise. No slope lag — pure structural midpoint.
- ZLEMA — Zero-Lag EMA
The well-known John Ehlers / Raff zero-lag construction. Uses price momentum to offset the inherent EMA delay. Fast, clean, widely understood.
- Hull MA — Standard 3-step WMA (true implementation)
Implemented correctly as the full 3-step WMA process: WMA(n/2), WMA(n), then WMA(√n) of the 2×half − full series. Many indicators get this wrong. This one doesn't.
- McGinley Dynamic
Self-adjusting MA that corrects for speed differences in up and down markets. Unlike EMA or SMA, it adapts to the actual velocity of price movement. Historically a strong trend-following baseline — rarely gets whipsawed.
- MedianFlow — HL2 → WMA → EMA
One of my personal creation: a two-pass smoother built on the midpoint (HL2) instead of close price. Naturally more centered on the candle body, less distorted by wick spikes. The AppliedPrice input is bypassed for this baseline — it always uses HL2 by design. Very smooth baseline, try to lower the period if too much smooth.
─────────────────────────────────────────
OTHER CHANGES IN v6.5
─────────────────────────────────────────
- Pullback Slope Gate — filters pullback entries based on slope direction or threshold vs. the entry threshold multiplier
- LT Filter Mode now defaults to Smooth (was None) — helps reduce false LT band breakouts
- KernelRelativeWeight now exposed as a dedicated input (was hardcoded internally)
Feedback welcome as always.
AlgoAlex81
WHAT WAS REPLACED — and why
THE FULL v6.5 BASELINE LIST 
Screenshot_798.png
