How to run a multi indicator strategy in bt?

I want to run a multi indicator strategy. I want to long a stock when stock price > sma and also RSI > 70 else short the stock if stock price < sma and RSI <30 . I have tried this way

import bt
import numpy as np

# download data
data = bt.get('aapl', start='2018-01-01')

# calculate moving average DataFrame using pandas' rolling_mean
import pandas as pd
# a rolling mean is a moving average, right?
data['sma'] = data.rolling(50).mean()


import talib as ta

d=data['aapl'].values.reshape(len(data))
data['rsi']=(ta.RSI(d))

signal=(data['aapl']>data['sma']) & (data['rsi']>70)

s = bt.Strategy('multi', [bt.algos.SelectWhere(signal),
                               bt.algos.WeighEqually(),
                               bt.algos.Rebalance()])

# now we create the Backtest
t = bt.Backtest(s, data)

res = bt.run(t)

This doesn’t work?

[davidhuang123]

It seems it’s easier to dissect your strategy into 2 strategies. You can write 2 script for each and combine them to see the results .

On Wed, Jul 10, 2019 at 13:17 fightthepower [e-mail] wrote:

@daviduang123 Sorry for the late reply and also thank you for the code. But can I ask one doubt. In the sample SMA strategy http://pmorissette.github.io/bt/examples.html#sma-strategy. The signal signal = data > sma gives true when price is greater than sma

2019-06-18 True 2019-06-19 True 2019-06-20 True 2019-06-21 True 2019-06-24 True 2019-06-25 True 2019-06-26 True 2019-06-27 True 2019-06-28 True 2019-07-01 True 2019-07-02 True 2019-07-03 True 2019-07-05 True 2019-07-08 True 2019-07-09 True

But in my case data[‘signal’]=(data[‘aapl’]>data[‘sma’]) & (data[‘rsi’]>70) signal will only be true when rsi is greater than 70 and when it decreases below 70 it becomes False.

What I want is I should buy the aapl stock when rsi is 70 and sell it when it reach rsi 30 a long position. Then when rsi is below 30 go for short that is sell at rsi<30 and buy back this at 70

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/pmorissette/bt/issues/183?email_source=notifications&email_token=AC255UYCNV2QUB6UEGRK7FLP6VWF5A5CNFSM4H6VEUH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZSKEDI#issuecomment-509911565, or mute the thread https://github.com/notifications/unsubscribe-auth/AC255U2IXIOIKPYXMBRHPUDP6VWF5ANCNFSM4H6VEUHQ .

[davidhuang123]

Try below, I guess you need to feed the right format so that the library understand.

import bt import numpy as np import pandas as pd import talib as ta import copy #import copy function

download data

data = bt.get(‘aapl’, start=‘2018-01-01’)

calculate moving average DataFrame using pandas’ rolling_mean

data[‘sma’] = data.rolling(50).mean()

d=data[‘aapl’].values.reshape(len(data)) data[‘rsi’]=(ta.RSI(d))

data = data.fillna(0) data[‘signal’]=(data[‘aapl’]>data[‘sma’]) & (data[‘rsi’]>70)

tw_data = copy.deepcopy(data) #make copy of data without association #Assign fixed weight to column tw_data[‘aapl’]=data[‘signal’]

#drop column Unused data=data.drop([‘sma’,‘rsi’,‘signal’], axis=1) tw_data=tw_data.drop([‘sma’,‘rsi’,‘signal’],axis=1)

s = bt.Strategy(‘multi’, [bt.algos.SelectWhere(tw_data), bt.algos.WeighEqually(), bt.algos.Rebalance()])

now we create the Backtest

t = bt.Backtest(s, data)

res = bt.run(t)

%matplotlib inline res.plot()