@run-out Cool. Took your advice! Here's what I ended up with.

First, I wrote a simple CAGR analyzer that anyone can use:

class CAGRAnalyzer(bt.analyzers.Analyzer): """ Analyzer returning CAGR of the portfolio """ def nextstart(self): self.rets = AutoOrderedDict() self.rets.start_value = self.strategy.broker.getvalue() self.rets.start_date = self.strategy.datetime.datetime().date() def stop(self): self.rets.end_value = self.strategy.broker.getvalue() self.rets.end_date = self.strategy.datetime.datetime().date() self.rets.num_years = (self.rets.end_date - self.rets.start_date).days / 365.25 self.rets.cagr = self.calculate_cagr(self.rets.start_value, self.rets.end_value, self.rets.num_years) def calculate_cagr(self, start_value, end_value, num_years): """ The CAGR formula is: EV / BV ^ (1/n) – 1. EV and BV are the ending and beginning values, while n is the number of time periods (usually months or years) for which you are calculating an average. The ^ character means “to the power of”; we take the ratio of EV / BV and raise it to the power of 1/n. Subtracting one (one = 100%) """ return ((end_value / start_value) ** (1 / num_years)) - 1

Next, I derived a real CAGR analyzer from it which pulls in a csv file of CPI data and calculates the annualized inflation rate and finally the real CAGR:

class RealCAGRAnalyzer(CAGRAnalyzer): params = (('cpi_data_file_path', None),) """ Analyzer returning the real CAGR (adjusted for inflation using values from a csv passed in with a "date" column and a "us_cpi" column """ def __init__(self): self.cpi_data_file_path = self.params.cpi_data_file_path self.cpi_data = pd.read_csv(self.cpi_data_file_path, parse_dates=['date'], index_col=0) def nextstart(self): super(RealCAGRAnalyzer, self).nextstart() self.rets.start_cpi = self.cpi_data.loc[pd.to_datetime(self.rets.start_date)]['us_cpi'] def stop(self): super(RealCAGRAnalyzer, self).stop() self.rets.end_cpi = self.cpi_data.loc[pd.to_datetime(self.rets.end_date)]['us_cpi'] self.rets.cagr = self.calculate_cagr(self.rets.start_value, self.rets.end_value, self.rets.num_years) self.rets.annualized_inflation_rate = self.calculate_cagr(self.rets.start_cpi, self.rets.end_cpi, self.rets.num_years) self.rets.real_cagr = self.annualized_real_return(self.rets.cagr, self.rets.annualized_inflation_rate) def annualized_real_return(self, nominal_cagr, annualized_inflation_rate): """ Annualized Real Return = ((1 + CAGR / 1 + annualized inflation rate) – 1) X 100 """ return (1 + nominal_cagr) / (1 + annualized_inflation_rate) - 1

Though im not an expert in finance, everything looks reasonable:

start_date: 1913-01-31 end_date: 2020-11-30 start_value: $100,000.00 end_value: $38,158,746.88 num_years: 107.83 start_cpi: $9.8 end_cpi: $260.23 cagr: 5.67% annualized_inflation_rate: 3.09% real_cagr: 2.50%

Thanks for the help!

@run-out It worked! I had tried running the next() within the indicator class but I wasn't including the subscript [0] in median[0].
THANKS A LOT!

@run-out if you start out with a portfolio with no cash and then increase it to 50% cash, the expected return and standard deviation are also halved, thereby leaving the Sharpe ratio unchanged.

@run-out Hey I found the error!!!

here I found the solution: https://community.backtrader.com/topic/2582/run-multiple-data-in-for-loop

I gave my trade_list class an other name. In my example I gave it the name "Trade_The_List" and than I replaced it in the analyzer like this:

cerebro.addanalyzer(Trade_The_List, _name='trade_list')

now it works :)

thank you for your help!

@run-out yes thats what i was referring to.However i am noticing certain issues with the file as well which may be leading to the issue

when i am sampling 1 min tf to 5 min tf,i notice empty rows are coming for saturday and sunday.
How can i remove those as i have already tried
df.dropna(inplace=True) without any luck.

Here df=pandas dataframe

@petehe The line you posted issues buy trailing stop order only.

@run-out Since in the backtrader docs(backtrader.com/docu/data-multitimeframe/data-multitimeframe/) it says it needs runonce to False.

@mhk Here's an example here

Just search for backtrader pandas extend lines and you can find lots of example in the community. Let us know how it goes.

@shravan-venkataraman

I am looking to backtest option strategies (i.e. long/put) using futures/index data, exactly the same question you mentioned at the beginning of this thread.

Are you able to use backtrader to backtest options based on Futures? If yes, have you find any similar article? Thanks!

@tigerandy May I ask why you are concerened about the way this is starting up? Have you looked at the built in indicator at all?

@coder1111 I just tested the following:

self.ols = bt.ind.OLS_TransformationN(period=10) self.ols = bt.ind.OLS_TransformationN(self.datas[0], self.datas[1], period=10) self.ols = bt.ind.OLS_TransformationN(self.datas[0].close, self.datas[1].close, period=10)

And they are all working fine. Perhaps as I suggested in my earlier post you could share your data, and while you are at it, share the rest of your code altogether, and we can help you find your problem. As it stands we don't have enough information.

@vladisld Thankyou @vladisld it worked.

@marsario said in math.log with linebuffer._LineDelay object:

self.lines.mean[0]= bt.Sum(self.change_sum)/len(self.weights)

I think you are close. This line:

self.lines.mean[0]= bt.Sum(self.change_sum)/len(self.weights)

The self.change_sum is already a sum from the previous line. It's cummulative. So I think this line should read:

self.lines.mean[0]= self.change_sum[0]/len(self.weights)

Also, in your first line.

self.close_log[0] = math.log(self.datas[0].close[0])

Import math and try that. self.datas[0].close[0] is a scalar and math.log should work.