@dasch Do you know why the build is failing?

@backtrader Ah, that makes sense. Thanks a lot for your response.

@tianjixuetu said in I backtest a strategy about futures,but I find there something strange about the profit: there is something different between your case and my case. you let the margin equals 2000 forever,but in real world, the margin may be change awlays the price volitility. Real world is different, you may want to educate your self a little bit more on this. Sometimes margin doesn't fluctuate with the daily prices, even it still depends on volatility. Also an advice - post the script you use in the very first message, not in the very last. There are no mid readers here, so only you know what is ran. My simplified case shows that bt works as expected, therefore the reason of the discrepancy is in (a) your additions to the bt or (b) in your calculations posted here. Seems it is case (b).

@Paska-Houso this way maybe have somthing error,for example,the shorter data can get before it happened.

ok,this is a question I answer sombody,you can get the quesiton at the address,but I reply as a topic.I am very sorry.

This is one message which obviously slipped through my fingers, but this question was answered before: no. The default IB implementation is single account only.

@Ramesh-Gangappa said in AttributeError: module 'backtrader' has no attribute 'Cerebro': C:\Server\Kite Development\BackTrader>python signal.py Traceback (most recent call last): File "signal.py", line 2, in <module> @Ramesh-Gangappa said in AttributeError: module 'backtrader' has no attribute 'Cerebro': File "C:\Program Files\Python37\lib\multiprocessing\process.py", line 18, in <module> import signal Your module has the same name as something which Python internals are trying to import. I actually fear you have misplaced things, but the safest bet: rename your script.

PyFolio is deprecated and no longer something which is tracked. Assuming you have all the pieces in place to get it working, my own experiments with sub-daily data failed because the internals of PyFolio seem to work but actually don't.

Let's put it the other way round What are you missing? Do you need something which is in a perpetual beta (or even alpha) state like many Android/iPhone applications? (always promising and always failing to deliver)

@Nguyễn-Sơn said in Ask on the logic of notify order and next to execute strategy: Where is the wrong here? I think it is obvious. Your logic sells several times in a row. @Nguyễn-Sơn said in Ask on the logic of notify order and next to execute strategy: else: if self.ema[0] < self.sma[0]: self.order = self.sell_bracket(stopprice = self.dataopen[0] * (1 + 0.02), limitprice = self.dataopen[0] * (1 - 0.04)) @Nguyễn-Sơn said in Ask on the logic of notify order and next to execute strategy: How to solve this problem? Correct your logic to something which wont sell several times in a row.

Indicators in backtrader have lines which are the output and are not calculated on the spot as your execution suggests. All ta-lib indicators when integrated in backtrader behave as the regular indicators of backtrader. They won't work isolated (which seems to be how you use them) See the actual auto-generated STOCH indicator here and the lines naming Docs - ta-lib - STOCH - https://www.backtrader.com/docu/talibindautoref/#stoch You have to use it like any other indicator in backtrader by referencing the lines and the actual index [x] you want to address (or by using arithmetic/logic operations on the lines) Docs - Indicator Usage - https://www.backtrader.com/docu/induse/ Docs - Platform Concepts - Lines - https://www.backtrader.com/docu/concepts/#lines You may also have a look at the ta-lib comparison post here Docs - ta-lib integration - https://www.backtrader.com/blog/posts/2016-07-26-talib-integration/talib-integration/ As shown the values are exactly the same (actually backtrader is more accurate or de-facto-compliant in some cases, where ta-lib fails to define an initial seed for exponentially smoothed based indicators)

If a parent order expires, so do the children. As for closed you will have to define the meaning of that.

If you issued 2 sell in a row you would go short. But you probably want to use the size parameter of the methods or use close to close a position and then either sell or buy to go to the other side. Or use a Sizer or ... Docs - Strategy - https://www.backtrader.com/docu/strategy/ and see buy, sell and close @Nguyễn-Sơn said in How to make a run to both long and short?: Other question is asking for the buy_bracket as well as sell_bracket. What exactly the meaning of stopprice and limitprice. Is it the take profit target and stoploss threshold? The signature indicates that one of the brackets is a Stop order and the other one a Limit order. The prices correspond to those. But yes, they do correspond to those concepts.

Hi, I did as ab_trader suggested and can confirm that it works pretty well. I am actually loading a future as a chain in backtrader via a default_strategy that I inherit when I create a new strategy. So my code works on a specific future and I am able to trade the whole curve. Let me know if you have any questions.

Hi, this is what I have done and indeed it works. Mysteries of pickle! thanks a lot

@backtrader Thank you. I have other issue when cannot download the graph which is ploted

@backtrader said in Cumulative/Running mean implementation: Thanks, anyways I created my own indicator. Btw I meant Expanding mean ie. avg[i] = avg(df[0:i]).

Hey @vbs I observed the same error when creating my own calendar; I also feel this code might be erroneous. My data holds price information after the trading session. Once it gets outside of the session, it enters an infinite loop as the time is always after the end of the session.

@backtrader hi Sir, At first I want to sorry because I forgot to provide you the data as I thought it was not important Which of course clarifies why you need that for a value=1000. I don't understand this? Also I've tried to do as your suggestions and read the documentation, but it is really hard for me. But it is okay now, thank you so much. I will be more careful next time