OpenSkillIndex← back to search
claude-code

ClawForge strategy-backtest

Runs SMA crossover backtests on historical OHLCV/candlestick data, calculating total return, Sharpe ratio, max drawdown, win rate, and trade log. Supports CSV files and JSON input with automatic AKShare/hhxg column normalization. Use when the user asks to backtest a trading strategy, evaluate strategy performance on historical price data, run quantitative analysis, or mentions OHLCV, candlestick data, or equity curves.

install
source · Clone the upstream repo
git clone https://github.com/leionion/ClawForge
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/leionion/ClawForge "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/04-Process/strategy_backtest" ~/.claude/skills/leionion-clawforge-strategy-backtest && rm -rf "$T"
manifest: skills/04-Process/strategy_backtest/SKILL.md
source content

strategy-backtest — Quantitative Strategy Backtesting

Runs strategy backtests on historical OHLCV data and returns performance metrics as JSON. Supports SMA crossover strategy with configurable fast/slow periods.

Usage

# Demo mode — uses built-in sample_ohlcv.csv
python3 strategy_backtest.py

# Backtest with custom CSV data
python3 strategy_backtest.py --data path/to/ohlcv.csv

# Backtest with JSON string input
python3 strategy_backtest.py --data '[{"open":10,"high":11,"low":9,"close":10.5,"volume":100}]'

# Custom SMA periods
python3 strategy_backtest.py --data prices.csv --fast 10 --slow 30

# Human-readable output
python3 strategy_backtest.py --data prices.csv --output print

Parameters

FlagDefaultDescription
--data
sample_ohlcv.csv
CSV path or JSON string (OHLCV)
--strategy
sma_crossover
Strategy type
--fast
5
Fast SMA period
--slow
20
Slow SMA period
--output
json
Output format (
json
or 
print
)

Supports column names in English (

open/high/low/close/volume
) or Chinese AKShare format (
开盘/收盘/最高/最低/成交量/日期
).

Example output

{
  "total_return": 0.0523,
  "sharpe_ratio": 1.2345,
  "max_drawdown": -0.0812,
  "win_rate": 0.6,
  "trade_count": 10,
  "trades": [
    {"date": "2024-01-15", "action": "buy", "price": 150.25},
    {"date": "2024-02-01", "action": "sell", "price": 158.50, "pnl": 0.0549}
  ]
}

Error handling

  • Missing pandas: prints 
    {"error": "pandas required: pip install pandas"}
  • Missing columns: reports which OHLCV columns are absent
  • Insufficient data: returns error if fewer rows than the slow SMA window
  • Unknown strategy: reports the unrecognized strategy name

Programmatic API

from strategy_backtest import run_backtest
metrics = run_backtest("prices.csv", strategy="sma_crossover", fast=5, slow=20)

Related skills

  • hhxg-top-hhxg-python: fetch A-share OHLCV data → feed into this skill
  • session-memory: store backtest metrics for later comparison