@run-out
Yes, that works. Thank you very much for your input, very much appreciated.

@marsario There have been a lot of conversations on this topic. Here is one I remember. https://community.backtrader.com/topic/2240/how-to-speed-up-almost-100-times-when-add-data-and-preload-data?_=1626988579575

Try a thorough search through the community, I'm sure this topic has been bantied about a lot.

@andi I just realized that the second column is already the actual portfolio value (i.e cash + positions).

@pshiw There's two parts to your question.

Yes you can build an execution strategy using Replay. This allows you to have two different time frames for your data.

If you have implemented options in backtrader, great. But my understanding is that backtrader does not work well or at all with options. You might want to google that before you commit to using backtrader.

@andi Yes you can build reports at the portfolio level. There is an area called analyzers with built in modules or you can create your own custom reports to suit your needs.

@tony-0 Do you mind if I ask why you don't want to pre-process this?

@guyue We can't help you without a clear question

Still digging... looks like this deal is actually the one which may be causing the issue...

It's somehow placing the sell before the first buy:

*** NEW DEAL *** BUYING: size 0.007998528270798173 price 1250.23 DEBUGGING: Size 0 Take Profit Price 0.0 2021-01-07T13:26:00, SELL EXECUTED, Size: -0.04059 Price: 1261.41, Cost: -51.20, Comm 0.00 2021-01-07T13:26:00, BUY EXECUTED, Size: 0.00800 Price: 1250.61, Cost: -10.09, Comm 0.00 DEBUGGING: Size 0.007998528270798173 Take Profit Price 1266.2426249999999 2021-01-07T16:39:00, SELL EXECUTED, Size: -0.00800 Price: 1265.86, Cost: -10.12, Comm 0.00

Given how I have my def next() setup, I am unsure how this is possible...

@run-out I thought I had this working with backtrader.analyzers.PyFolio() however I then realized it is not meant to handle 1m candles (or anything < 1d), so although the report was populating, the data seemed incorrect.

Is it possible to load data into quanstats reports using 1m candles? I am trying to do it with backtrader.analyzers.TimeReturn() but running into some issues... and before I kept digging wanted to see if this is the right approach.

def backtest_pair(pair, tp, bo, so, mstc, os, sos, ss): # Create a cerebro entity cerebro = bt.Cerebro() # Add the Data Feed to Cerebro data = load_data(pair) cerebro.adddata(data) # Add Strategy cerebro.addstrategy(DCA3C, tp=tp, bo=bo, so=so, mstc=mstc, os=os, sos=sos, ss=ss) # Add Analyzer cerebro.addanalyzer(bt.analyzers.TimeReturn, _name='Returns', timeframe=bt.TimeFrame.Minutes, compression=1) # Set starting cash starting_cash = calc_max_funds(bo, so, mstc, os) cerebro.broker.setcash(starting_cash) # Set the commission - 0.1% ... divide by 100 to remove the % cerebro.broker.setcommission(commission=0.001) print('Starting Portfolio Value: %.2f' % cerebro.broker.getvalue()) # Run results = cerebro.run() # cerebro.plot() strat = results[0] portfolio_stats = strat.analyzers.getbyname('Returns') returns = pd.DataFrame.from_dict(portfolio_stats.get_analysis(), orient="index") returns.index = pd.to_datetime(returns.index) quantstats.extend_pandas() # returns.index = returns.index.tz_convert(None) # Save bot_settings = 'TP: <b>{}</b> BO: <b>{}</b> SO: <b>{}</b> MSTC: <b>{}</b> OS: <b>{}</b> SOS: <b>{}</b> SS: <b>{}</b>'.format( strat.params.tp, strat.params.bo, strat.params.so, strat.params.mstc, strat.params.os, strat.params.sos, strat.params.ss) quantstats.reports.html(returns, output=f'{pair.replace("/", "_")}.html', title=f'Pair: <b>{pair}</b> {bot_settings}<br>', trading_year_days=365) print('Final Portfolio Value: %.2f' % cerebro.broker.getvalue())

Getting this error:

Traceback (most recent call last): File "/Users/ethan/PycharmProjects/backtraderbot/main.py", line 315, in <module> backtest_pair('ETH/USDT', tp=1.75, bo=10, so=20, mstc=30, os=1.05, sos=2, ss=1) File "/Users/ethan/PycharmProjects/backtraderbot/main.py", line 250, in backtest_pair trading_year_days=365) File "/Users/ethan/PycharmProjects/backtraderbot/venv/lib/python3.7/site-packages/quantstats/reports.py", line 101, in html by='max drawdown', ascending=True)[:10] File "/Users/ethan/PycharmProjects/backtraderbot/venv/lib/python3.7/site-packages/pandas/util/_decorators.py", line 311, in wrapper return func(*args, **kwargs) File "/Users/ethan/PycharmProjects/backtraderbot/venv/lib/python3.7/site-packages/pandas/core/frame.py", line 6254, in sort_values k = self._get_label_or_level_values(by, axis=axis) File "/Users/ethan/PycharmProjects/backtraderbot/venv/lib/python3.7/site-packages/pandas/core/generic.py", line 1776, in _get_label_or_level_values raise KeyError(key) KeyError: 'max drawdown'

Thanks!

@run-out Thanks!

@vladisld much appreciated mate :) That worked out

UPDATE:
Solved it by defining trading hours in calendar

class MyCalendar(bt.TradingCalendar): params = { 'open': time(9, 15), 'close': time(15, 30), }

Thanks for your help! This is a popular and well-tested trading strategy. And it's one of the best crypto trading methods for beginners. bitcoin news - here you can read more about btc trading strategies.

@hghhgghdf-dfdf

Can you please how your backtesting P/L looks like. I'm assuming we buy at 21.84 and sell at 15.90 is it? are there any other confirmation studies we can add to it ? as mostly Zig Zag repaints(overwrites the signals)

@run-out Thanks! I think your idea works, but that is what I want.

before I start to try backtrader, I use talib+pandas, they are easy to transfer signals from :

signalA=[0,0,1,0,0,0,0,-1,0,0] # triggered at 3rd day,not yet signalB=[0,0,0,1,0,0,-1,0,0,0] # triggered at 4rd day,not yet signalC=[0,0,0,0,0,1,0,-1,0,0] # triggered at 6rd day, buy here

to

signalA=[0,0,1,1,1,1,1,-1,-1,-1] # triggered at 3rd day,not yet signalB=[0,0,0,1,1,1,-1,-1,-1,-1] # triggered at 4rd day,not yet signalC=[0,0,0,0,0,1,1,-1,-1,-1] # triggered at 6rd day, buy here

then , when signalA==1 and signalB==1 and signalC==1, it trigger the long signal.
Since I am a newbie for backtrader, I don't know to datafeeds work that way.