mns-backtest

MNS Contrarian Investment Strategy Backtest Skill

Steps

1. Build

cargo build --release

2. Run backtest

# 默认行为：先输出多资产回测（美股+红利低波+黄金）对比，再输出单资产（纳指）参数对比
mns backtest

# 使用自定义配置文件（仅单资产回测）
mns backtest run --config path/to/config.toml

Compare multiple configuration files (single-asset backtest only)

mns backtest run --compare config1.toml,config2.toml

View adjustable parameter descriptions (including current default values)

mns backtest params

### 3. 参数调优

查看当前配置：
```bash
mns config

Modify a single parameter:

mns config buy_ratio.extreme_fear 70
mns config thresholds.fear 40

Use preset configuration files (located in .agents/skills/mns-backtest/data/）：

Configuration file	Features
config_defensive.toml	Defensive configuration (low drawdown, neutral no buying,Current default）
config_balanced.toml	Balanced configuration
config_circuit_breaker.toml	Circuit breaker mechanism
config_swing.toml	Swing trading
config_extreme_contrarian.toml	Extreme contrarian
config_value.toml	Value-oriented
config_historical_aggressive.toml	Historically aggressive configuration (⚠️ overfitting risk, for research reference only)

mns backtest run --config .agents/skills/mns-backtest/data/config_defensive.toml

Data description

The backtest uses embedded data (include_str! compiled into the binary):

• fgi_2016_2020.csv — daily CNN fear and greed index (high confidence)
• fgi_2020_2025.csv — monthly approximate fear and greed index (low confidence)
• monthly_real_final.csv — multi-asset monthly prices (Nasdaq / dividend low-volatility / RMB gold price)

Note: Updating data requires modifying the CSV file and re-running cargo build --release。

Core logic files

• src/backtest.rs — backtest engine,run_backtest()、run_multi_asset_backtest()、run_param_comparison()
• src/strategy.rs — strategy core; the backtest reuses the same buy/sell/risk judgment functions (calculate_buy_suggestions etc.); the execution flow is a backtest-specific simulation
• src/config.rs — parameter structure definition,AppConfig::default_config() is the defensive configuration default value

Interpreting results

After running, a comparison table is output (annualized return, total return, maximum drawdown, number of buys and sells); focus on:

1. vs buy and hold: Is the strategy's annualized return close to the benchmark
2. Maximum drawdown: The lower the drawdown, the stronger the strategy's defensiveness
3. Number of buys: Setting buy_ratio to 0 in the neutral range will significantly reduce the frequency

The value of the strategy lies indisciplinerather than excess returns—it helps overcome emotional decision-making and enforces buy and sell signals during extreme markets (panic/greed).