Fractional Candlestick Long Only Experimental V10

Fractional Candlestick Long-Only Strategy – Technical Description

This document provides a professional English description of the "Fractional Candlestick Long Only Experimental V6" strategy using pure CF/AB fractional kernels and wavelet-based filtering.

1. Fractional Candlesticks (CF / AB)
The strategy computes two fractional representations of price using Caputo–Fabrizio (CF) and Atangana–Baleanu (AB) kernels. These provide long-memory filtering without EMA approximations. Both CF and AB versions are applied to O/H/L/C, producing fractional candlesticks and fractional Heikin-Ashi variants.

2. Trend Stack Logic
Trend confirmation is based on a 4-component stack:
- CF close > AB close
- HA_CF close > HA_AB close
- HA_CF bullish
- HA_AB bullish
The user selects how many components must align (4, 3, or any 2).

3. Wavelet Filtering
A wavelet transform (Haar, Daubechies-4, Mexican Hat) is applied to a chosen source (e.g., HA_CF close). The wavelet response is used as:
- entry filter (4 modes)
- exit filter (4 modes)
Wavelet modes: off, confirm, wavelet-only, block adverse signals.

4. Trailing System
Trailing stop uses fractional AB low × buffer, providing long-memory dynamic trailing behavior. A fractional trend channel (CF/AB lows vs HA highs) is also plotted.

5. Exit Framework
Exit options include: stack flip, CF<AB cross, HA fractional flip, kernel trailing, with optional wavelet refinement.

6. Practical Advantages
- Stronger noise suppression
- Fractional memory structure extraction
- Multi-scale impulse detection
- Reduced false entries/exits
- Long-only alignment with structural trend

This system integrates fractional calculus, wavelet transforms, and trend stacking into a cohesive adaptive long-only model.