@vladisld said in How to calculate indicator line using prior period values?:

Market Direction Indicator

MDI = 100 * (self.CP2[-2] - self.CP2[-1]) / ((self.data.close[-1] + self.data.close[-1]) / 2)

# Create lines
self.l.CP2[0] = self.CP2[-1]
self.l.MDI[0] = MDI

print("CP2: ", self.CP2[-1], " MDI: ", MDI)

Amazing, thanks @vladisld!

I'm attempting to write an indicator based off a projected moving average formula (CP2) which also has Lamberts Marker Direction Formula (MDI) in there as well.

My issue is that I need to use prior period values of CP2 to calculate the MDI formula but I'm not sure of the best approach as I understand that most of the calculations should be done in the init() function.

Here is what I have so far and as you can see I have attempted to use an array to store the CP2 values which is probably not best practice but getting the 'IndexError: list index out of range' error because it can't access the values. I guess I mostly need help in how to write this in the best practice format.

from __future__ import (absolute_import, division, print_function,
                        unicode_literals)

# Import the backtrader platform
import backtrader as bt
import math
from decimal import *

class ProjectedMovingAverageIndicator(bt.Indicator):

  lines = ('CP2','MDI',)
  params = (('periodSlow', 100),('periodFast', 25),('movav', bt.indicators.MovAv.Simple),)
  plotinfo = dict(subplot=True)
  plotlines = dict(
    CP2=dict(_samecolor=False),
    MDI=dict(_samecolor=False),
    )

  def __init__(self):
    self.addminperiod(self.params.periodSlow)

    # Calculate moving averages
    self.movavSlow = self.p.movav(self.data.close, period=self.p.periodSlow)
    self.movavFast = self.p.movav(self.data.close, period=self.p.periodFast)
    self.CP2 = []

  def next(self):

    # Perry Kaufman project moving average
    self.CP2.append(self.p.periodSlow * ((self.movavSlow[0] - self.data.close[-1]) / self.p.periodSlow) - ((self.movavFast[0] - self.data.close[-1]) / self.p.periodFast))

    # Market Direction Indicator
    MDI = 100 * (self.CP2[-2] - self.CP2[-1]) / ((self.data.close[-1] + self.data.close[-1]) / 2)

    # Create lines
    self.l.CP2[0] = self.CP2[-1]
    self.l.MDI[0] = MDI

    print("CP2: ", CP2[-1], " MDI: ", MDI)

Thanks

Ah that looks like a good place to start thanks!

Ok I played around a bit and got it working:

from __future__ import (absolute_import, division, print_function,
                        unicode_literals)

# Import the backtrader platform
import backtrader as bt
import math

class ProjectedMovingAverageIndicator(bt.Indicator):

  lines = ('smaSlow','smaFast',)
  params = (('periodSlow', 200),('periodFast', 100),('movav', bt.indicators.MovAv.Simple),)
  plotinfo = dict(subplot=False)
  plotlines = dict(
    smaSlow=dict(_samecolor=True),
    smaFast=dict(_samecolor=False),
    )

  def __init__(self):
    self.addminperiod(self.params.periodSlow)
    movavSlow = self.p.movav(self.data, period=self.p.periodSlow)
    movavFast = self.p.movav(self.data, period=self.p.periodFast)
    self.l.smaSlow = movavSlow #bt.Cmp(movav, self.data)
    self.l.smaFast = movavFast #bt.Cmp(movav, self.data)

  def signal(self):
    if self.lines.smaFast[-1] < self.lines.smaSlow[-1] and self.lines.smaFast[0] > self.lines.smaSlow[0]:
        signal = True
    else:
        signal = False
    #print(signal)
    return signal

Thanks

Hmm image link didn't work, here is another [https://prnt.sc/vd8mek]

Thanks for the speedy reply @ab_trader. That's what I thought should happen, but as you can see in the attached it's not happening. I must be doing something wrong, I looked at the Bollinger band indicator code on Github and can see that the lines are being declared in the init function as opposed to next:

def __init__(self):
        self.lines.mid = ma = self.p.movav(self.data, period=self.p.period)
        stddev = self.p.devfactor * StdDev(self.data, ma, period=self.p.period,
                                           movav=self.p.movav)
        self.lines.top = ma + stddev
        self.lines.bot = ma - stddev

        super(BollingerBands, self).__init__()

Should I be doing that instead?

Thanks


How do you plot more than one line on a chart with an indicator?

The documentation mentions the LinePlotterIndicator but it doesn't explain how to call it.

For testing I'm just creating a simple Indicator with two moving averages which I would like to plot on the chart:

# Import the backtrader platform
import backtrader as bt
import math

class ProjectedMovingAverageIndicator(bt.Indicator):
  lines = ('smaSlow','smaFast',)
  params = (('periodSlow', 200),('periodFast', 100),)
  plotinfo = dict(subplot=False)

  def __init__(self):
      self.addminperiod(self.params.periodSlow)
      
  def next(self):
      datasum = math.fsum(self.data.close.get(size=self.p.periodSlow))
      self.lines.smaSlow[0] = datasum / self.p.periodSlow
      self.lines.smaFast[0] = datasum / self.p.periodFast
      print(self.lines.smaSlow[0],' ', self.lines.smaFast[0])

  def signal(self):
    if self.lines.smaFast[-1] < self.lines.smaSlow[-1] and self.lines.smaFast[0] > self.lines.smaSlow[0]:
        signal = True
    else:
        signal = False
    #print(signal)
    return signal

Thanks

I don't have a background in maths or data science and I've been reading Perry Kauffman's Trading Systems and Methods and there are a lot of mathematical notation/formulas in the book which I would like to understand. Can someone point me in the direction of a good resource where I can learn about this please? There is a lot out there and I'm not sure where to start.

Many thanks!

I am trying to replicate the behaviour of a full script via the command line with btrun and have the following command:

btrun --csvformat btcsv \
      --data data/gbpusd-1m.csv \
      --strategy strategies/TemplateStrategy.py \
      --fromdate 2003-1-1 \
      --todate 2003-1-5 \
      --timeframe minutes \
      --commission 0 \
      --cash 10000 \
      --compression 1

How would I pass parameters to GenericCSVData such as open, high, low, close, openinterest as below?

    data = bt.feeds.GenericCSVData(
        timeframe=bt.TimeFrame.Minutes,
        compression=1,
        dataname=datafile,
        fromdate=datetime.datetime(2003, 1, 1),
        todate=datetime.datetime(2003, 1, 5),
        dtformat=('%Y-%m-%d %H:%M:%S'),
        datetime=0,
        high=2,
        low=3,
        open=1,
        close=4,
        volume=5,
        openinterest=-1)

Thanks

Ah well now I feel stupid lol. Thanks for pointing that out, I thought I had bought the entire years worth of data but it would appear not!