Indicators/Strategies/Analyzers

• 

  Indicators & Machine Learning
  • Sahli Simo

  2
  0
  Votes
  2
  Posts
  557
  Views

  B

  You cannot put values into the future, which is what you are trying to do. Keep the values in a ring buffer (use for example collections.deque with a maxlen during creation), push the last predicted value and pop the value from the left which is the current value.
• Y

  PyFolio performance issues on Pandas Data Frame
  • Yousuf

  3
  0
  Votes
  3
  Posts
  457
  Views

  Y

  Apologies for not articulating the problem well enough in the first post. However, you have correctly inferred that I am facing issues with not being able to see intraday positions and returns from PyFolio analyzer. My run strat function looks like this: def runstrat1(): args = parse_args() # Create a cerebro entity cerebro = bt.Cerebro(stdstats=True) # Add a strategy cerebro.addstrategy(strategy1.TestStrategy) # Get a pandas dataframe datapath = ('./dataset/data2.csv') dataframe = pd.read_csv(datapath,parse_dates=True,index_col=0) # Pass it to the backtrader datafeed and add it to the cerebro data = PandasDataExpanded(dataname=dataframe) cerebro.adddata(data) # Add PyFolio Analyzer cerebro.addanalyzer(bt.analyzers.PyFolio, _name='pyfolio') # Run over everything strats = cerebro.run() returns, positions, transactions, gross_lev = pyfoliozer.get_pf_items() for each in strat0.analyzers: each.print() print(positions['cash']) # Plot the result if not args.noplot: cerebro.plot(style='bar') return positions My class for pandas data feed look like this: class PandasDataExpanded(btfeeds.PandasData): lines = ('barCount',) params = (('barCount', -1),)
• S

  Unable to get LinePlotterIndicator to display
  • sfkiwi

  4
  0
  Votes
  4
  Posts
  455
  Views

  B

  The lines of all other indicators do actually get plotted, there is no reason why above and below to not get plotted. @sfkiwi said in Unable to get LinePlotterIndicator to display: But I am not able to get these values to display on the cerebro.plot(). I read in the docs that because of the use of > and < the lines won't just plot automatically This is when you make a comparison and is not assigned to a line. In that case one uses LinePlotterIndicator to fake an indicator that holds the resulting operation of the comparison. Don't you think that an actual plot (you posted an example in the other thread) could help?
• S

  Has anyone got LinePlotterIndicator to work?
  • sfkiwi

  2
  0
  Votes
  2
  Posts
  326
  Views

  B

  See the answer and corresponding question in your other thread: Community - Unable to get LinePlotterIndicator to display
• A

  How to calculate Chaikin Money Flow
  • Anil Bhatt

  2
  0
  Votes
  2
  Posts
  289
  Views

  B

  @anil-bhatt said in How to calculate Chaikin Money Flow: # Let the indicator get enough data self.addminperiod(self.p.len) That's not needed and it will probably hurt. But without knowing how and when the values differ ...
• J

  Indicator using multi datafeed: operate only on overlapping bar
  • Jacob

  20
  0
  Votes
  20
  Posts
  1513
  Views

  B

  @jacob said in Indicator using multi datafeed: operate only on overlapping bar: this is a great workaround for this issue/bug. This is not a workaround. It is clearly stated in the FAQ: Community - FAQ The system cannot know which timeframe/compression your data feed has unless you tell the system.
• R

  Period params in indicators
  • Rstrong

  2
  0
  Votes
  2
  Posts
  259
  Views

  B

  Indicators are designed to know nothing about datetime. The period they operate is measured in price bars (groups of OHLC price components) This in turn allows to create indicators on indicators. You may of course design a custom indicators which require the source data to carry a datetime payload. The period starts counting at the current bar and goes backwards (it cannot obviously go forward)
• M

  Why a simple user defined indicator failed to find the self.data.high
  • mothee

  5
  0
  Votes
  5
  Posts
  465
  Views

  B

  See, I do understand that your code is yours, but that is far away from being any test script. Furthermore, it is for sure not part of any test script because the indentation is wrong which would immediately generate a SyntaxError. The easiest way, also for you, to reproduce the error and really narrow it down to the lack of a params definition is to define a very simple script which actually reproduces the error. See for example this in your script cerebro.addindicator(func_ind, **kwargs_i) What's func_ind? Without ruling out a problem in backtrader itself, you may be passing the wrong thing to addindicator. Just a long shot.
• 

  It's here - a beta you can use RIGHT NOW - Essential trade statistics - all in one place
  • Richard O'Regan

  11
  5
  Votes
  11
  Posts
  2602
  Views

  T

  Is it possible to use this without working from your branch @Richard-O-Regan ? It's quite behind the master backtrader branch now, but the stats here are very very useful!
• H

  How to code a trading system that works with Backtrader?
  • hlad

  11
  0
  Votes
  11
  Posts
  1317
  Views

  B

  @hlad said in How to code a trading system that works with Backtrader?: Is each feed capable of storing such state variables? Why shouldn't? The statement of problem is really unclear. You can store things in additional lines defined for the data feed and you store and access the entire history. You can add your own attributes and store as many as you like in for example a list or just the last value in a simple instance attribute.
• S

  Live feed inplace resampling
  • Søren Pallesen

  8
  0
  Votes
  8
  Posts
  1292
  Views

  D

  @dan-yang said in Live feed inplace resampling: I actually have the same issue, and I tested with the same code. It runs well for the history part (10 next) of the data, but fails right after that. Anybody has an idea? import sys import backtrader as bt class TestStrategy(bt.Strategy): def notify_data(self, data, status, *args, **kwargs): print('*' * 5, 'DATA NOTIF:', data._getstatusname(status)) def next(self): msg = '***** NEXT: {} '.format(bt.num2date(self.data.datetime[0])) for d in self.datas: msg += '{} {} {} '.format(d._name, len(d), d.close[0]) print(msg) def run(argv): cerebro = bt.Cerebro() data = bt.feeds.CCXT(exchange='coinbasepro', symbol='BTC/USD', timeframe=bt.TimeFrame.Minutes, compression=1, ohlcv_limit=10) cerebro.adddata(data, name='1M') cerebro.resampledata(data, timeframe=bt.TimeFrame.Minutes, compression=2, name='2M') cerebro.addstrategy(TestStrategy) cerebro.run() if __name__ == '__main__': run(sys.argv) here is the error ***** DATA NOTIF: LIVE ***** NEXT: 2018-06-27 06:34:00 1M 1 6084.04 2M 1 6084.04 ***** NEXT: 2018-06-27 06:35:00 1M 2 6084.05 2M 1 6084.04 ***** NEXT: 2018-06-27 06:36:00 1M 3 6084.05 2M 2 6084.05 ***** NEXT: 2018-06-27 06:37:00 1M 4 6084.05 2M 2 6084.05 ***** NEXT: 2018-06-27 06:38:00 1M 5 6084.04 2M 3 6084.04 ***** NEXT: 2018-06-27 06:39:00 1M 6 6080.75 2M 3 6084.04 ***** NEXT: 2018-06-27 06:40:00 1M 7 6080.74 2M 4 6080.74 ***** NEXT: 2018-06-27 06:41:00 1M 8 6080.74 2M 4 6080.74 ***** NEXT: 2018-06-27 06:42:00 1M 9 6080.75 2M 5 6080.75 ***** NEXT: 2018-06-27 06:43:00 1M 10 6080.74 2M 5 6080.75 ***** NEXT: 2018-06-27 06:44:00 1M 11 6080.75 2M 6 6080.75 ***** NEXT: 2018-06-27 06:45:00 1M 12 6080.75 2M 6 6080.75 Traceback (most recent call last): File "test.py", line 80, in <module> run(sys.argv) File "test.py", line 76, in run cerebro.run() File "c:\Anaconda3\lib\site-packages\backtrader-1.9.64.122-py3.6.egg\backtrader\cerebro.py", line 1127, in run File "c:\Anaconda3\lib\site-packages\backtrader-1.9.64.122-py3.6.egg\backtrader\cerebro.py", line 1298, in runstrategies File "c:\Anaconda3\lib\site-packages\backtrader-1.9.64.122-py3.6.egg\backtrader\cerebro.py", line 1557, in _runnext ValueError: min() arg is an empty sequence Never mind, I found the issue is with my exchange 'coinbasepro'. It works well on other exchange like 'bitmex'. I will leave the post here just in case other people run into the same issue and can learn from my case.
• B

  How to get the current value for the fund, not just in notify fcn?
  • Bingchen Liu

  5
  0
  Votes
  5
  Posts
  385
  Views

  B

  Got it, thank you !
• B

  Short the dip, long the correction
  • brettelliot

  3
  0
  Votes
  3
  Posts
  587
  Views

  B

  Ah good point. I was (over)thinking of it as a momentum strategy.
• R

  Referencing the first day in a Data
  • Rstrong

  3
  0
  Votes
  3
  Posts
  405
  Views

  R

  @backtrader wow that seems so obvious now haha Thanks
• I

  Height of indicator
  • imartynov

  2
  0
  Votes
  2
  Posts
  287
  Views

  B

  You need to play with rowmajor and rowminor. See Docs - Plotting
• A

  Compute indicator in _next_ process ?
  • aaggmm

  2
  0
  Votes
  2
  Posts
  295
  Views

  B

  Indicators are computed in their own next, so it isn't clear what your problem is. Docs - Indicator Development With regards to dynamic parameters, nothing prevents you from changing the parameters and altering the calculations you make during next
• A

  Extend a pandas datafeed
  • aaggmm

  3
  0
  Votes
  3
  Posts
  531
  Views

  B

  Any lines object can be extended and is documented. Specifically for data feeds: see Docs - Extending a data feed
• A

  How to get max close price for the past n days
  • aaggmm

  3
  0
  Votes
  3
  Posts
  401
  Views

  A

  I see. Thank you!
• M

  how to get the indicator data alone?
  • mothee

  8
  0
  Votes
  8
  Posts
  883
  Views

  B

  @mothee said in how to get the indicator data alone?: cerebro.addindicator(indcls=btind.RSI) That adds an indicator into a sample strategy class. It is meant for experimentation purposes. You should read the top of the forum. It will help you posting and coding: For code/output blocks: Use ``` (aka backtick or grave accent) in a single line before and after the block. See: http://commonmark.org/help/ The Indicator class it not complex or simple. It is meant to be used withing the backtrader echosystem. See: Docs - Platform Concepts - Getting a Slice It shows how to get the data (from the current point backwards). You may also access the attribute array of each line of the indicator if you want it all. data = indicator_instance.lines[0].array
• S

  2 Resample datas in one single strategy.
  • satheesh7134

  2
  0
  Votes
  2
  Posts
  330
  Views

  B

  See: Docs - Data Resampling for the general functionality about resampling Docs - Data - Multiple Timeframe for documentation and code where the data is resampled from the lower timeframe. You simply need to change the timeframe/compression parameters