Backtesting TW stock with pandas, backtesting, talib, and FinMind

stock_backtesting.py

from backtesting import Backtest, Strategy
from backtesting.lib import crossover

from FinMind.data import DataLoader
import pandas as pd

import talib
from talib import abstract

## 取得資料
dl = DataLoader()
df = dl.taiwan_stock_daily(stock_id='2330', start_date='2008-01-01', end_date='2023-1-1')
## 整理資料格式
df = df.rename(columns={"date": "Date"})
df.set_index("Date" , inplace=True)
df = df.set_index(pd.DatetimeIndex(pd.to_datetime(df.index)))
## backtesting.py 格式
df1 = df.rename(columns={"open": "Open", "max": "High", "min": "Low", "close": "Close", "Trading_Volume": "Volume"})
## ta-lib 格式
df2 = df.rename(columns={"max": "high", "min": "low", "Trading_Volume": "Volume"})
## 取得 KD 值
df_kd = abstract.STOCH(df2,fastk_period=9, slowk_period=3,slowd_period=3)
## 合併資料
df1 = pd.merge(df1, df_kd, on="Date")

"""
KD 策略(K<20買；K>80賣)
"""
class KdCross(Strategy):
    def init(self):
        super().init()
        
    def next(self):
        if crossover(20, self.data.slowk): ## K<20 買進
            self.buy()
        elif crossover(self.data.slowk, 80): ## K>80 平倉
            self.position.close()
            
bt = Backtest(df1, KdCross, cash=10000, commission=0.001798) ## 交易成本 0.1798%
output = bt.run()
print(output)
bt.plot()

"""
跌破季線買進；K>80賣出
Strategy: 跌破60日均線買進, K>80賣出
"""

# class PriceBelowMAandKDJ(Strategy):
#     def init(self):
#         super().init()
#         self.ma_60 = self.I(talib.MA, self.data.Close, timeperiod=60)  # 60-day moving average

#     def next(self):
#         if self.data.Close < self.ma_60:  # Price falls below 60-day MA
#             self.buy()
#         elif self.data.slowk > 80:  # KDJ > 80
#             self.position.close()

# bt = Backtest(df1, PriceBelowMAandKDJ, cash=10000, commission=0.001798)  # 交易成本 0.1798%
# output = bt.run()
# print(output)
# bt.plot()


"""
跌破月線的兩個標準差買進；收盤價大於5日線的兩個標準差平倉
"""
## Calculate 20-day mean and standard deviation
# df1['5_Day_MA'] = df1['Close'].rolling(window=5).mean()
# df1['5_Day_STD'] = df1['Close'].rolling(window=5).std()
# df1['20_Day_MA'] = df1['Close'].rolling(window=20).mean()
# df1['20_Day_STD'] = df1['Close'].rolling(window=20).std()
# class MeanReversionStrategy(Strategy):
#     def init(self):
#         super().init()
    
#     def next(self):
#         if self.data.Close < self.data['20_Day_MA']:
#             self.buy()
#         elif self.data.Close > self.data['5_Day_MA']: 
#             self.position.close()

# bt = Backtest(df1, MeanReversionStrategy, cash=10000, commission=0.001798)  # 交易成本 0.1798%
# output = bt.run()
# print(output)
# bt.plot()