Weighted Trading System

Weighted Trading System is a modular quantitative trading framework designed to combine multiple technical indicators into a unified confidence score.

The main philosophy is:

No single indicator is perfect. Indicators should contribute weighted evidence toward a trading decision.

The system aims to provide:

• Flexible indicator scoring
• Confidence-based entries
• Multiple exit methodologies
• Automatic weight optimization
• Machine Learning integration
• Long historical backtesting support
• Migration compatibility from TradingView Pine Script

---

Features

Indicator Scoring

Every indicator contributes positive or negative confidence.

Example:

EMA200 Above Price +20

Triple Supertrend Green +35

ADX > 25 +20

Stoch RSI Bullish +30

MACD Cross Up +25

Bearish Divergence -15

Weak Momentum -10

Final confidence:

Bullish Score = 145

Bearish Score = 35

Net Confidence = 110

Trade decision:

Net Confidence >= Entry Threshold

→ Open Position

---

Supported Indicators

Trend

• EMA
• SMA
• WMA
• VWMA
• Supertrend
• ADX
• DMI
• Ichimoku
• Parabolic SAR

---

Momentum

• RSI
• Stochastic
• Stochastic RSI
• MACD
• CCI
• ROC
• Momentum

---

Volatility

• ATR
• Bollinger Bands
• Keltner Channel
• Donchian Channel

---

Volume

• OBV
• CMF
• MFI
• VWAP

---

Documentation

Dokumentasi lengkap tersedia di folder docs/:

Indicators

Dokumen	Deskripsi
EMA	Exponential Moving Average — fast & slow EMA
Stochastic RSI	StochRSI — RSI-based stochastic oscillator
ADX	Average Directional Index — trend strength
ATR	Average True Range — volatility measurement
Supertrend	Trend-following indicator with ATR bands

Strategy — Weighted System

Dokumen	Deskripsi
Overview	Arsitektur sistem, diagram alur, komponen utama
Scoring Engine	Detail scoring: Supertrend, StochRSI, EMA, ADX, CI, Persistence
Entry Methods	Single entry, Pyramid entry, Grid entry
Exit Methods	7 metode exit: ST1/2/3, Scoring, Bars Since, dll

---

Architecture

Weighted System

│
├── config/
│
├── data/
│
├── docs/
│   ├── indicators/
│   │   ├── adx.md
│   │   ├── atr.md
│   │   ├── ema.md
│   │   ├── stochastic_rsi.md
│   │   └── supertrend.md
│   └── strategy/
│       ├── overview.md
│       ├── scoring-engine.md
│       ├── entry-methods.md
│       └── exit-methods.md
│
├── src/
│   │
│   ├── datasource/
│   │
│   ├── indicators/
│   │
│   ├── features/
│   │
│   ├── core/
│   │
│   ├── backtest/
│   │
│   ├── optimizer/
│   │
│   └── ml/
│
└── tests/

---

Scoring Engine

Indicators do not directly generate entries.

Instead they contribute confidence.

Example:

score = 0

score += 20 if close > ema200 else 0

score += 15 if supertrend_green else 0

score += 10 if adx > 25 else 0

score += 25 if macd_cross else 0

score -= 15 if bearish_divergence else 0

Result:

net_score = bullish_score - bearish_score

Example:

Bullish = 135

Bearish = 25


Net = 110

---

Entry Engine

Several entry models are supported.

Single Entry

Single position only.

Position = 0

Open Buy

---

Pyramid

Increase exposure when confidence rises.

Example

Threshold = 100


Trade 1

score >=100


Trade 2

score >=110


Trade 3

score >=120


Trade 4

score >=130

---

Grid

Price based additions.

Example

Trade 1

Trade 2 at -1 ATR


Trade 3 at -2 ATR


Trade 4 at -3 ATR

---

Exit Engine

Supported methods

Supertrend Exit

ST1

ST2

ST3

---

Score Exit

score < 50

---

Score + EMA

score < 50

or


close < ema200

---

Persistence Exit

Example

score < 50


for 3 bars

---

ATR Exit

ATR Stop

ATR Trailing

---

Break Even

Supported

---

Partial Take Profit

Supported

---

Backtesting

Metrics

Supported

Net Profit

CAGR

Profit Factor

Sharpe Ratio

Sortino Ratio

Winrate

Expectancy

Max Drawdown

Recovery Factor

Ulcer Index

Calmar Ratio

---

Weight Optimization

Static

Manual weights.

---

Statistical

Feature Importance

Correlation Analysis

Mutual Information

Permutation Importance

---

Machine Learning

Random Forest

LightGBM

CatBoost

XGBoost

---

Reinforcement Learning

PPO

DQN

---

Adaptive Weights

The system can retrain periodically.

Examples

Train:

2020-2024

Validate:

2025

Trade:

2026

Or

Rolling Window

Train

Last 2 years


Trade

Current month


Retrain

Every Month

---

Development Roadmap

Phase 1

Indicator migration

Phase 2

Scoring engine

Phase 3

Entry system

Phase 4

Exit engine

Phase 5

Backtester

Phase 6

Weight optimization

Phase 7

Machine learning

Phase 8

Adaptive retraining

Phase 9

Live execution

---

Goal

The objective of this project is not to discover a perfect indicator.

The objective is to create a confidence aggregation framework capable of combining many weak signals into a robust decision-making process.

This project is inspired by concepts from:

• Ensemble Learning
• Factor Investing
• Quantitative Momentum
• Smart Money Concepts
• Adaptive Trading Systems
• Machine Learning Feature Engineering