跳转至

Example 20: Support & Resistance Portfolio Strategy

Strategy Overview

A multi-stock portfolio strategy that identifies support and resistance levels using historical price patterns, then buys near support and sells near resistance with confirmation from RSI, MACD, ATR, and Donchian channels.

Buy Signals

  • Price near support level (within 2%) + RSI oversold (< 30) or MACD golden cross
  • Price at or below Donchian lower band + RSI not overbought

Sell Signals

  • ATR trailing stop (highest price since buy - 2.5 * ATR)
  • Price near resistance + RSI overbought (> 70) or MACD death cross
  • Price at or above Donchian upper band

Position Management

  • Max 25% of portfolio per stock (4 stocks max)
  • Equal-weight allocation among eligible candidates
  • Per-trade commission: 0.03%, sell tax: 0.1%

Stock Pool

Code Name (EN) Name (CN) Sector
601390 China Railway 中国中铁 Infrastructure
600916 China Gold 中国黄金 Gold
002594 BYD 比亚迪 EV/New Energy
601088 China Shenhua 中国神华 Coal
601857 PetroChina 中国石油 Oil
600536 China Soft 中国软件 Technology
601398 ICBC 工商银行 Banking
518880 Gold ETF 黄金ETF Gold ETF

Backtest Results

Metric Value
Period 2020-01-01 to 2026-03-30 (~6 years)
Starting Cash 1,000,000
Final Value 2,371,889.70
Total Return +137.19%
Buy Orders 114
Sell Orders 112
Total Trades 226

Files in This Directory

File Description
sr_strategy.py Strategy code
run_backtest.py Backtest runner script
sr_backtest.png Portfolio chart (PNG)
sr_backtest.html Interactive HTML report
sr_backtest.md Markdown summary report
sr_backtest.json Full backtest data (JSON)

How to Run

# From the project root directory:
python examples/20_sr_strategy/run_backtest.py

Backtest note: order* calls are queued first and filled at the next trading day open in local backtests.

New reports (timestamped PNG/HTML/MD/JSON) are written under the repository root reports/ directory.

Requires local data files in data/ directory. If you don't have them: 

# Download data first (example 19)
python examples/19_local_data_backtest.py --download-all

Technical Indicators Used

  • Support/Resistance: Price level clustering from 80-day lookback
  • RSI (14): Relative Strength Index for overbought/oversold detection
  • MACD (12,26,9): Trend direction and momentum
  • ATR (14): Average True Range for volatility-based stop loss
  • Donchian Channel (20): Price breakout/breakdown detection

Key Design Principles

  1. Mean reversion with trend confirmation: Buy dips at support, but confirm with MACD/RSI
  2. Dynamic stop loss: ATR-based trailing stop adapts to volatility
  3. Diversification: 8 stocks across different sectors reduces single-stock risk
  4. Position limits: Max 25% per stock prevents over-concentration