Anyone facing ValueError for RSIIndicator for ta backtesting?

Question

Trying to get some idea on how to debug this. Just started to use backtesting.py + TA to conduct some backtesting. Seems to work for EMA indicator. RSI indicator did not work: Tried with self.rsi instead of self.rsi [-1] as well. But the same ValueError was raised.

Error seems to start from backtesting.py on line 139:

File ~\anaconda3\Lib\site-packages\backtesting\backtesting.py:139, in >Strategy.I(self, func, name, plot, overlay, color, scatter, *args, **kwargs)

ValueError: attempt to get argmax of an empty sequence.

Was expecting RSI Indicator to work like any other indicators.

import pandas as pd
import yfinance as yf
import ta
from backtesting import Backtest, Strategy, set_bokeh_output
from backtesting.lib import crossover

df = yf.download('AAPL', start='2022-12-01', end='2023-12-31')

class EMA_200(Strategy):

n1 = 200
n2 = 14
rsi_btm = 30
rsi_top = 70

def init(self):
    close = self.data.Close
    self.ema_200 = self.I(ta.trend.ema_indicator, pd.Series(close), self.n1)
    self.rsi = self.I(ta.momentum.RSIIndicator, pd.Series(close), self.n2)

def next(self):
    if self.data.index[-1] >= pd.to_datetime('2023-01-01'):
        if not self.position: # no existing position
            if (self.data.Close < self.ema_200) and (self.rsi[-1] <= self.rsi_btm):
                self.buy()
            elif self.data.Close > self.ema_200 and (self.rsi[-1] >= self.rsi_top):
                self.sell()
        elif self.position: # existing position
            if (self.position.is_long) and (self.data.Close > self.ema_200) and (self.rsi[-1] >= self.rsi_top):
                self.position.close()
            elif (self.position.is_short) and (self.data.Close < self.ema_200) and (self.rsi[-1] >= self.rsi_btm):
                self.position.close()
bt = Backtest(df, EMA_200, cash=1000000, commission=.002,
exclusive_orders=True)
stats = bt.run()
stats

Don C · Accepted Answer · 2024-02-10 20:50:59Z

1

I was encountering the same issue with the StochasticRSI. self.I takes in a function, this function should return an array of the same length of (price) data https://kernc.github.io/backtesting.py/doc/backtesting/backtesting.html#backtesting.backtesting.Strategy.I

However, StochRSIIndicator & RSIIndicator do not return indicator data arrays they return an indicator object. The array can be accessed using the stochrsi()/rsi() method.

The example below is how I got it sorted for StochRSIIndicator. Should be the same for RSIIndicator but using the rsi() method.

rsi = lambda : StochRSIIndicator(close=close, window=self.RSI_WINDOW).stochrsi()
self.rsi = self.I(rsi)

Please do not hesitate to respond to this with a cleaner solution should you find one.

answered Feb 10, 2024 at 20:50

Don C

111 bronze badge

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Anyone facing ValueError for RSIIndicator for ta backtesting?

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related