双均线策略

双均线策略是量化投资中最经典、最基础的策略，也是入门量化投资的第一个实战项目。

---

策略原理

当短期均线上穿长期均线时买入（金叉），当短期均线下穿长期均线时卖出（死叉）。

为什么有效？

• 金叉意味着短期趋势走强，可能开启一波上涨
• 死叉意味着短期趋势走弱，可能开启一波下跌
• 本质是趋势跟踪——不去预测趋势，而是跟随趋势

---

策略实现

import pandas as pd
import numpy as np

def dual_ma_strategy(df, short_window=5, long_window=20):
    """
    双均线策略
    
    参数:
        df: 包含 close 列的 DataFrame
        short_window: 短期均线窗口
        long_window: 长期均线窗口
    
    返回:
        增加了 signal 和 position 列的 DataFrame
    """
    # 1. 计算均线
    df = df.copy()
    df['ma_short'] = df['close'].rolling(window=short_window).mean()
    df['ma_long'] = df['close'].rolling(window=long_window).mean()
    
    # 2. 生成交易信号
    df['signal'] = 0
    df.loc[df['ma_short'] > df['ma_long'], 'signal'] = 1   # 持仓
    df.loc[df['ma_short'] <= df['ma_long'], 'signal'] = 0  # 空仓
    
    # 3. 计算持仓变化（交易点）
    df['position'] = df['signal'].diff()
    # position = 1: 买入信号, position = -1: 卖出信号
    
    return df

---

信号可视化

import matplotlib.pyplot as plt

def plot_dual_ma(df, short_window=5, long_window=20):
    """可视化双均线策略信号"""
    df = dual_ma_strategy(df, short_window, long_window)
    
    fig, (ax1, ax2) = plt.subplots(2, 1, figsize=(14, 10), 
                                     gridspec_kw={'height_ratios': [3, 1]})
    
    # 上图：价格和均线
    ax1.plot(df.index, df['close'], label='收盘价', alpha=0.6)
    ax1.plot(df.index, df['ma_short'], label=f'MA{short_window}', linewidth=1)
    ax1.plot(df.index, df['ma_long'], label=f'MA{long_window}', linewidth=1)
    
    # 标记买卖点
    buy_signals = df[df['position'] == 1]
    sell_signals = df[df['position'] == -1]
    ax1.scatter(buy_signals.index, buy_signals['close'], 
                color='red', marker='^', s=100, label='买入', zorder=5)
    ax1.scatter(sell_signals.index, sell_signals['close'], 
                color='green', marker='v', s=100, label='卖出', zorder=5)
    
    ax1.set_title(f'双均线策略 (MA{short_window} / MA{long_window})')
    ax1.legend(loc='upper left')
    ax1.grid(True, alpha=0.3)
    
    # 下图：持仓信号
    ax2.fill_between(df.index, 0, df['signal'], alpha=0.3, color='blue')
    ax2.set_ylabel('持仓')
    ax2.set_ylim(-0.1, 1.1)
    ax2.grid(True, alpha=0.3)
    
    plt.tight_layout()
    plt.show()

---

策略回测

def backtest_dual_ma(df, short_window=5, long_window=20, 
                      initial_capital=100000, commission=0.0003):
    """
    双均线策略回测
    
    参数:
        df: 包含 close 列的 DataFrame
        short_window: 短期均线窗口
        long_window: 长期均线窗口
        initial_capital: 初始资金
        commission: 手续费率（默认万三）
    """
    df = dual_ma_strategy(df, short_window, long_window)
    
    # 计算每日收益率
    df['market_return'] = df['close'].pct_change()
    
    # 策略收益 = 持仓信号 * 市场收益（前一天的信号后一天生效）
    df['strategy_return'] = df['signal'].shift(1) * df['market_return']
    
    # 扣除交易手续费
    trades = df['position'].abs()  # 交易发生时
    df['strategy_return'] = df['strategy_return'] - trades * commission
    
    # 累计收益
    df['cum_market'] = (1 + df['market_return']).cumprod()
    df['cum_strategy'] = (1 + df['strategy_return']).cumprod()
    
    # 绩效指标
    total_return = df['cum_strategy'].iloc[-1] - 1
    annual_return = (1 + total_return) ** (252 / len(df)) - 1
    sharpe = np.sqrt(252) * df['strategy_return'].mean() / df['strategy_return'].std()
    max_drawdown = (df['cum_strategy'] / df['cum_strategy'].cummax() - 1).min()
    win_rate = (df['strategy_return'] > 0).sum() / (df['strategy_return'] != 0).sum()
    
    print(f"=== 双均线策略回测 (MA{short_window}/MA{long_window}) ===")
    print(f"累计收益率: {total_return*100:.2f}%")
    print(f"年化收益率: {annual_return*100:.2f}%")
    print(f"夏普比率:   {sharpe:.2f}")
    print(f"最大回撤:   {max_drawdown*100:.2f}%")
    print(f"胜率:       {win_rate*100:.2f}%")
    print(f"交易次数:   {int(df['position'].abs().sum())}")
    
    return df

---

参数优化

不同股票和周期适合不同的均线参数，可以通过网格搜索来优化：

def optimize_ma_params(df, short_range, long_range):
    """寻找最优均线参数组合"""
    results = []
    
    for short in short_range:
        for long in long_range:
            if short >= long:
                continue
            df_bt = backtest_dual_ma(df.copy(), short, long, verbose=False)
            results.append({
                'short': short,
                'long': long,
                'sharpe': sharpe,
                'total_return': total_return
            })
    
    results_df = pd.DataFrame(results)
    best = results_df.loc[results_df['sharpe'].idxmax()]
    
    print(f"最优参数: MA{int(best['short'])} / MA{int(best['long'])}")
    print(f"最佳夏普比率: {best['sharpe']:.2f}")
    
    return results_df

---

策略注意事项

1. 震荡市不适用：横盘整理时会产生大量假信号
2. 交易成本不可忽视：频繁交易的手续费会蚕食利润
3. 滑点：实际成交价与信号价之间的差异
4. 避免过度优化：参数调得太精致反而在实盘中表现差

---

下一步：因子选股策略 →