新能源锂电行业与CTA增强策略研究

New Energy Lithium Industry and CTA-Enhanced Strategy Research

中文简介

本项目基于 A 股新能源锂电产业链股票池，构建动量、估值、成长、盈利质量、低波动和相对强弱等多维因子，使用 Rank IC / ICIR 评估因子有效性，并进一步构建 ICIR 加权多因子选股策略、行业 CTA 趋势择时策略，以及“多因子选股 + CTA 仓位控制”的融合增强策略。

English Summary

This project studies a China A-share new-energy lithium battery industry universe. It builds a multi-factor dataset covering momentum, valuation, growth, profitability, low-volatility, and relative-strength signals, evaluates factors with Rank IC / ICIR, and then tests three strategy layers: an ICIR-weighted multi-factor stock-selection strategy, a CTA-style moving-average trend overlay, and a fused multi-factor plus CTA position-control strategy.

Key Findings

The public README only includes summary statistics from the research report. Raw data and generated result files are intentionally excluded.

Item	Result
Stock universe	20 A-share lithium battery industry leaders
Final sample period	2023-04-04 to 2025-12-31
Final daily observations	10,782
Best factor by ICIR	20D low-volatility, ICIR 0.4607
Momentum finding	20D momentum ICIR -0.2019, suggesting short-term reversal
Best CTA parameter set	MA10 / MA30
Multi-factor + CTA switch total return	81.02%
Multi-factor + CTA switch annual return	25.13%
Multi-factor + CTA switch Sharpe	0.95
Multi-factor + CTA switch max drawdown	-24.75%
Equal-weight lithium benchmark total return	-3.39%

The core research takeaway is that lithium stocks in this sample were not well suited to naive momentum chasing. Low volatility, gross margin, net profit growth, and valuation repair signals were more useful for stock selection, while the CTA trend overlay helped reduce industry beta exposure during downtrends.

For a longer report-style summary, see docs/research_report.md.

Project Name

• Chinese: 新能源锂电行业与CTA增强策略研究
• English: New Energy Lithium Industry and CTA-Enhanced Strategy Research
• Suggested repository slug: new-energy-lithium-cta-enhancement

Research Pipeline

1. Build the master factor dataset from Tushare market, valuation, and financial data.
2. Run Rank IC / ICIR analysis for candidate factors.
3. Backtest an ICIR-weighted multi-factor stock-selection strategy.
4. Backtest a CTA trend strategy on the equal-weight lithium industry index.
5. Fuse multi-factor stock selection with CTA trend-based exposure control.

Repository Structure

new_energy_lithium_cta_enhancement/
  README.md
  .env.example
  .gitignore
  requirements.txt
  pyproject.toml
  docs/
    github_release_checklist.md
    research_report.md
    research_design.md
  src/
    lithium_cta/
      __init__.py
      backtest.py
      config.py
      data_fetch.py
      factors.py
      ic_analysis.py
      metrics.py
  scripts/
    01_build_master_dataset.py
    02_factor_ic_analysis.py
    03_icir_multifactor_backtest.py
    04_cta_trend_strategy.py
    05_fusion_multifactor_cta.py
  data/
    .gitkeep
  outputs/
    .gitkeep
  tests/
    test_factors.py
    test_metrics.py

Data Policy

Raw data and generated research outputs are not included in this repository.

• data/ is for local raw and processed data.
• outputs/ is for local backtest results, Excel reports, and charts.
• Both directories are ignored by Git except their .gitkeep placeholders.
• Tushare credentials must be provided through environment variables, not hard-coded in source code.

Quick Start

cd new_energy_lithium_cta_enhancement
python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt

Set your Tushare token:

$env:TUSHARE_TOKEN="your_tushare_token_here"

Build the master dataset:

python scripts/01_build_master_dataset.py

Run factor IC analysis:

python scripts/02_factor_ic_analysis.py

Run the ICIR-weighted multi-factor backtest:

python scripts/03_icir_multifactor_backtest.py

Run the CTA trend strategy:

python scripts/04_cta_trend_strategy.py

Run the fused multi-factor + CTA strategy:

python scripts/05_fusion_multifactor_cta.py

GitHub Upload Checklist

Before publishing this project, review:

• Initialize Git from inside new_energy_lithium_cta_enhancement, not from the parent research folder.
• No API token is present in source code, Markdown files, commit history, or logs.
• No raw CSV/XLSX data files are staged.
• No generated backtest output files are staged unless intentionally shared.
• README.md, .gitignore, .env.example, and requirements.txt are included.
• Public release is acceptable after secrets, raw data, and generated outputs are excluded.

Research Limitations

• The default universe is a hand-picked lithium battery industry stock pool, so it may contain survivorship and selection bias.
• Factor validity depends on data quality, reporting-date alignment, transaction-cost assumptions, and sample size.
• Backtests are simplified research simulations and do not model all real-world execution constraints.

Disclaimer

This project is for quantitative research, learning, and engineering practice only. It does not constitute investment advice. Any strategy or signal should be validated with stricter data controls, out-of-sample tests, transaction-cost modeling, and risk review before real-world use.