@run-out You are correct. Needed to go back to the drawing board as a result. Thank you.

Does this have anything to do with the "short interest" indicator? I'm trying to figure out how to create it in backtrader but I don't know if what this post is about is the same thing.

Your help & support is greatly appreciated. Thanks a bunch! @run-out

@vladisld said in How to backtest a group of stocks？:

Cerebro logic for each stock separately ( in a loop)

It works! I misunderstood what you meant. I'm just an idiot without your help. Thanks for your help.

@carsten said in how to upload several stocks data in backtrader?:

cerebro.adddata(data)

Use:

cerebro.adddata(data, name="insert your security name here")

The guy in your link didn't include fees & slippage.

Further more it should be noted that every single EMA combination tested (and most SMAs) outperformed the buy and hold annualized return of 6.32%^ during the test period (before allowing for transaction costs and slippage)Further more it should be noted that every single EMA combination tested (and most SMAs) outperformed the buy and hold annualized return of 6.32%^ during the test period (before allowing for transaction costs and slippage)

Thank you very much for all your help, much appreciated!

I've played a bit with boundoff, daily and resampled weekly data trying to get shifted weekly bar, but wasn't able to make it. Maybe it should be same timeframe specified for both original and target bars in order to force it to work.

@taewoo-kim whats the final script to convert dukascopy data?

Haha - it was hiding in plain sight! or I was just being blind! Thanks a lot guys!

Many thanks and thank you for your patience

See Blog - Plotting Date Ranges

OCO orders are all active at the same time.

Bracket orders (the ones you submit with parent) have a different lifecycle:

parent is 1st alive. children are inactive If the parent is cancelled the children are simply discarded (they weren't active) If the parent is executed, the children are activated Both children behave now like OCO orders

I can confirm that backtrader is by far the best option compared to zipline, and quantconnect. The zipline code is a mess, they implemented it, then decided to create a new library to implement basic statistics which they got wrong then created a new library to implement the risk metrics. It is inefficient and Quantopian holds back enough code to just make zipline unusable.

Quantconnect has a better architecture but it also has a lot of old legacy code.

Not mine.

Only if you want to use:

def pretty(d, indent=0): for key, value in d.iteritems(): if isinstance(value, dict): print '\t' * indent + (("%10s: {\n") % str(key).upper()) pretty(value, indent+1) print '\t' * indent + ' ' * 12 + ('} # %s #\n' % str(key).upper()) elif isinstance(value, list): for val in value: print '\t' * indent + (("%30s: [\n") % str(key).upper()) pretty(val, indent+1) print '\t' * indent + ' ' * 12 + ('] # %s #\n' % str(key).upper()) else: print '\t' * indent + (("%10s: %s") % (str(key).upper(),str(value))) tradeAnalyzer = strat.analyzers.getbyname('TradeAnalyzer') pretty(tradeAnalyzer.get_analysis())

The idea is always to have a generic model, like it is the case with the timezones, in which you may pass pytz instances or something that, more or less, complies to the interface of pytz.

A similar approach will be taken.

@richard-o-regan Good afternoon. I just joined BackTrader. Long-time futures trader; technical. Years ago with aid of programmers created own linux-based backtesting and then real time trading. Now looking for back testing platform. Found QuantConnect on IB site, but others recommended Backtrader. I will be working with programmer, but based on number of years experience backtesting tens of thousands of iterations, what we did was to rather than merely take output, and rearrange, place all output into a large spd sheet. That included say all 500 iterations for a test, along with all the indiv vars, and columns such as you suggest, % pft tdes, etc. However, with spd sheet, could easily sort on columns, as well as sub-sorts. Regards,

The problem is difficult to comprehend, because things seem to work like they:

Snippet1

self.b_bands = bt.ind.BollingerBands() self.rsi = bt.ind.RelativeStrengthIndex(period=self.params.rsi, upperband=70, lowerband=30)

Two (2) indicators are declared in the __init__ method of the Strategy and two (2) indicators show up in the char

BollingerBands which defaults to plotting along the data (because the values from this indicator are meant to be in line with the value) RSI which defaults to be plotted on a subplot because it has fixed values ranging from 0 to 100

Snippet 2

# self.b_bands = bt.ind.BollingerBands() self.rsi = bt.ind.RelativeStrengthIndex(period=self.params.rsi, upperband=70, lowerband=30) self.bb = BBandsIndicator(mode=self.params.mode)

And again two (2) indicators there. The story of RSI is already known.

BBandsIndicator is a direct descendent of Indicator and like most indicators the default is to plot on a subplot

Both charts are ok

Release 1.9.56.122 contains an exception for volume fields which are not convertible to float.