Portfolio backtesting & Benchmark

marketwizard

Hello,

I'm currently backtesting my strategies on multiple stocks using mutiple datafeeds.

I would like to compare my portfolio strategy returns to a benchmark (like the S&P500).

In the backtrader documentation I found this article:
https://www.backtrader.com/blog/posts/2016-07-22-benchmarking/benchmarking/

However, with the solutions proposed in the article, I suppose that if I add the S&P500 to cerebro, that cerebro will also backtest my strategy on the S&P500 as well.

What I'm trying to do is a bit different. For example:

Test the strategy on MSFT, BAC, JNJ, FB and APPL
Compare the results to the S&P500

And Ideally plot the results of the portfolio and the Benchmark

I hope that someone can help !

Cheers
Marketwizard

run-out

@marketwizard Put your benchmark line in first, then the rest of your stocks. Then when you are in next, simply trade self.datas[1] and greater.

marketwizard

@run-out How can I trade self.datas[1] and greater ?

Here is the code of my Init and Next funtions:

    def __init__(self):

        self.inds = dict()

        for i, d in enumerate(self.datas):

            self.inds[d] = dict()

            # Define the indicators
            self.inds[d]['ema50'] = bt.indicators.EMA(d.close, period=50)
            self.inds[d]['ema200'] = bt.indicators.EMA(d.close, period=200)
            self.inds[d]['atr'] = bt.indicators.ATR(d, period=14)

            # Define the crossover signals
            self.inds[d]['bull_cross'] = bt.indicators.CrossOver(self.inds[d]['ema50'], self.inds[d]['ema200'])
            self.inds[d]['bull_cross'].plotinfo.subplot = False

            self.inds[d]['bear_cross'] = bt.indicators.CrossOver(self.inds[d]['ema200'], self.inds[d]['ema50'])
            self.inds[d]['bear_cross'].plotinfo.subplot = False

            if i > 0:  # Check we are not on the first loop of data feed:
                if self.p.oneplot == True:
                    d.plotinfo.plotmaster = self.datas[0]

    def start(self):
        self.order = None  # sentinel to avoid operations on pending order
        self.curpos = None

    def prenext(self):
        self.next()

    def next(self):

        for i, d in enumerate(self.datas):

            #dt, dn = self.datetime.date(), d._name
            pos = self.getposition(d).size

            # Get the Amount of cash in the Portfolio
            cash = self.broker.get_cash()

            if self.order:
                return  # pending order execution

            if not pos:  # check if we already have an open position on the stock
               
                if self.inds[d]['bull_cross'][0] > 0.0:
                    # Calculation of the Stock Qty to buy depending on our risk strategy
                    # calculate the absolute stop loss distance with atr
                    pdist = self.inds[d]['bull_cross'][0] * self.p.atrdist

                    # calculate the stop price
                    self.pstop = d.close[0] - pdist

                    # calculate the amount of shares to buy / sell depending on the stop loss and the risk strategy
                    qty = math.floor((cash * self.p.trade_risk) / pdist)

                    # calculate the maximum exposure depending on the risk strategy
                    portfolio_exposure_calc = qty * d.close[0]

                    portfolio_exposure_strategy = cash * self.p.portfolio_expo
                    qty = math.floor(portfolio_exposure_strategy / d.close[0])
                    self.order = self.buy(data=d, size=qty)

                                   
            else:  # in the market

                if self.inds[d]['bear_cross'][0]:
                    self.close(data=d)  # stop met - get out

run-out

@marketwizard said in Portfolio backtesting & Benchmark:

for i, d in enumerate(self.datas):

datas is just a list, so like this:

for i, d in enumerate(self.datas[1:]):

marketwizard

@run-out thank you, it works for trading the datas excluding the first one.

However, when I try to use the benchmark observer I receive an error.
Here is the code I use for adding the datas to cerebro:

    for i in range(len(datalist)):
        data = bt.feeds.GenericCSVData(

            dataname=datalist[i][0],
            # Do not pass values before this date
            fromdate=datetime(fromDate.year, fromDate.month, fromDate.day),
            # Do not pass values before this date
            todate=datetime(datetime.today().year, datetime.today().month, datetime.today().day),

            nullvalue=0.0,

            dtformat=('%Y-%m-%d'),

            datetime=0,
            open=1,
            high=2,
            low=3,
            close=4,
            volume=5,
            openinterest=-1
        )

        cerebro.adddata(data, name=datalist[i][1])

    # Set the Cash for the Strategy
    cerebro.broker.setcash(icap)
    # Set the comissions
    cerebro.broker.setcommission(commission=0.005)

    #Add the benchmark Observer
    cerebro.addobserver(bt.observers.Benchmark, data=data[0])

and the error message:

Traceback (most recent call last):
  File "C:/Users/marketwizard/PycharmProjects/marketwizard_Backtests_PF/02_MW_EMA_RSI_V2_PF_Benchmark/main.py", line 339, in <module>
    cerebro.addobserver(bt.observers.Benchmark, data=data[1])
  File "C:\Users\marketwizard\PycharmProjects\Algotrading_libraries1\lib\site-packages\backtrader\lineseries.py", line 467, in __getitem__
    return self.lines[0][key]
  File "C:\Users\marketwizard\PycharmProjects\Algotrading_libraries1\lib\site-packages\backtrader\linebuffer.py", line 163, in __getitem__
    return self.array[self.idx + ago]
IndexError: array index out of range

Process finished with exit code 1

vladisld

@marketwizard said in Portfolio backtesting & Benchmark:

cerebro.addobserver(bt.observers.Benchmark, data=data[0])

I believe it should be:

cerebro.addobserver(bt.observers.Benchmark, data=data)

marketwizard

@vladisld now it works, thank you !

However I couldn't understand why... I thought that data was a list ?