扮演 Test Python Algorithmic Tra…

# 沪深300双均线策略回测脚本

一、环境准备

```python

import pandas as pd

import numpy as np

import akshare as ak # pip install akshare

```

• 数据源：akshare 免费获取指数日线，也可替换为 tushare/本地CSV
• 依赖：pandas、numpy、akshare

二、完整代码

```python

import akshare as ak

import pandas as pd

import numpy as np

# 1. 获取沪深300近三年日线数据

df = ak.index_zh_a_hist(symbol="000300", period="daily",

start_date="20220101", end_date="20241231")

df.rename(columns={"日期":"date","收盘":"close"}, inplace=True)

df["date"] = pd.to_datetime(df["date"])

df.set_index("date", inplace=True)

df.sort_index(inplace=True)

# 2. 计算均线与信号

df["ma5"] = df["close"].rolling(5).mean()

df["ma20"] = df["close"].rolling(20).mean()

df["signal"] = np.where(df["ma5"] > df["ma20"], 1, 0) # 金叉持仓，死叉空仓

df["position"] = df["signal"].shift(1) # 避免未来函数：次日执行

# 3. 计算策略与基准收益

df["ret"] = df["close"].pct_change()

df["strategy_ret"] = df["position"] * df["ret"]

df["cum_strategy"] = (1 + df["strategy_ret"]).cumprod()

df["cum_bench"] = (1 + df["ret"]).cumprod()

# 4. 年化收益率（按交易日250天计）

years = (df.index[-1] - df.index[0]).days / 365

ann_ret = df["cum_strategy"].iloc[-1] ** (1/years) - 1

ann_bench = df["cum_bench"].iloc[-1] ** (1/years) - 1

# 5. 最大回撤

def max_drawdown(cum):

peak = cum.cummax()

return ((cum - peak) / peak).min()

mdd = max_drawdown(df["cum_strategy"])

mdd_bench = max_drawdown(df["cum_bench"])

print(f"策略年化收益: {ann_ret:.2%}, 最大回撤: {mdd:.2%}")

print(f"基准年化收益: {ann_bench:.2%}, 最大回撤: {mdd_bench:.2%}")

```

三、关键设计要点

• 避免未来函数：信号用 shift(1)，当日收盘产生信号、次日才生效
• 交易假设简化：未计手续费（可按每次交易扣除万几至千几级别成本调整）、满仓进出、无滑点
• 年化天数：A股惯