Best posts made by dasch

@shieldman the version in github can be used for live trading with oanda v20. We have it running since today, it is not that well tested, so any feedback and testing is welcome.

You could store the last cost in your strategy. You would watch for finished trades and then inspect the attributes of the closed trade:

commission (float): current accumulated commission
pnl (float): current profit and loss of the trade (gross pnl)
pnlcomm (float): current profit and loss of the trade minus commission (net pnl)

The notification can be accessed in notify_trade

You could also create an indicator the same way.

I found a workaround for this issue. When creating the data feed, i used the following params which generate the correct timestamps:

rightedge=True, boundoff=1

with live data from oanda and rightedge=False the data will get delivered when arriving instead of being resampled.

You can do it this way:

# -*- coding: utf-8 -*-

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

# Import the backtrader platform
import backtrader as bt
import backtrader.indicators as btind
from backtrader.observer import Observer as btobserver

import datetime
import time

'''
TD Sequential-Style indicator

// S = Sell Signal
// B = Buy Signal
// A = Sell (above bar) or Buy (below bar) Signal that is 'not ideal' - meaning the signal bar's or the previous bar's high or low is not higer or lower than one of the previous 2 bars.
// - This could indicate that waiting another bar will deliver a 'true' signal. (Useful at the beginning of the trend reversal signal, not so much at the end)
//
// Indicator and Rules based on indicator as seen on CoinSheet.org


    Ein TD Sequential Buy Setup ist erfüllt, wenn 9 aufeinanderfolgende Bars tiefer Schliessen als 4 Bars zuvor. Für den Sell ist es umgekehrt, dort müssen 9 aufeinanderfolgende Bars höher Schliessen als 4 Bars zuvor.

    Das Ziel der Zählung ist es, mögliche Wendepunkte anzuzeigen und somit Low Risk Entrys zu finden. Falls die Bedingungen nicht bis zum 9er Count erfüllt wird, ist das Setup gecancelt. Sobald ein TD Setup erfüllt ist, würde normalerweise die Countdown resp. Intersection-Phase beginnen. Beim Countdown werden danach 13 Bars die einen tiefern Close haben, als das Low des zwei Bars zurückliegenden Stabes. Die Countdown Bars müssen jedoch nicht aufeinanderfolgen, die Zählung läuft weiter bis zu Bar 13, sofern nicht das Hoch des vorangehenden TD Setups (bei einem Buy Setup...sonst natürlich das Tief bei einem Sell) überschritten wird.

    Das ist jetzt wirklich nur ein ganz grober Überblick der Standartparameter.???

    Beim angehängten Indi handelt es sich übrigens um den TD Combo und nicht den TD Sequential. Der Unterschied zwischen den beiden liegt einfach gesagt darin, dass die Countdown Zählung beim erreichen eines kompletten TD Setups, die Countdownzählung zürückrechnet und somit schon beim ersten Bar des Setups beginnen kann, anstatt erst nach der Beendigung des Setups.:P

    Grob gesagt kann beim erfüllen eines Setups mit einem kleinen Wendepunkt gerechnet werden und beim erfüllen eines Countdowns ein grosser Wendepunkt.
    
    Entry:
    There are three methods of entry:
    (a) Enter on the close of the day on which the countdown is completed;
    (b) Long Trades: Enter on the close if Close[i] > Close[i - 4];
        Short Trades: Enter on the close if Close[i] < Close[i - 4];
    (c) Long Trades: Enter on the close if Close[i] > High[i - 2];
        Short Trades: Enter on the close if Close[i] < Low[i - 2]. Index: i ~ Current Bar. In this test we apply the 2nd method.
    
    Exit:
    Pattern Exit: The current position is liquidated once the new setup is complete and price fails to exceed the furthest price recorded by the recent inactive setup.
    Stop Loss Exit: 
    Long Trades: The difference between the close at the entry bar and the low of the lowest day of the pattern is subtracted from the low of the lowest day. Short Trades: The difference between the high of the highest day of the pattern and the close at the entry bar is added to the high of the highest day.                
    Note: In order to activate a stop loss, the close must exceed the calculated levels.

'''

class strategy(bt.Strategy):
    params = dict(
      candles_past_to_compare = 4,
      stoploss_dist = 0.0,
    )

    ''' Logging function for strategy '''
    def log(self, txt, dt=None):
        dt = dt or self.dataprimary.datetime.datetime(0)
        print('%s, %s' % (dt.isoformat(), txt))
    
    
    '''Initialization '''    
    def __init__(self):
        # Keep a reference to the primary data feed
        self.dataprimary = self.datas[0]
        # Keep a reference to the "close" line in the primary dataseries
        self.dataclose = self.dataprimary.close
        # Control vars
        self.live = True
        self.order = None

        self.bearishFlip = False
        self.bullishFlip = False

        self.tdsl = 0 # TD sequence long
        self.tdss = 0 # TD sequence short
        self.buySetup = False # TD buy setup 
        self.sellSetup = False # TD sell setup
        self.buyCountdown = 0 # TD buy countdown
        self.sellCountdown = 0 # TD sell countdown
        self.buyVal = 0 # buy value at intersection
        self.sellVal = 0 # sell value at intersection
        
        self.buySig = False
        self.idealBuySig = False
        self.sellSig = False
        self.idealSellSig = False

        print('-' * 80)
        print('Strategy Created')
        print('-' * 80)


    ''' Data notification '''
    def notify_data(self, data, status, *args, **kwargs):
        print('*' * 5, 'DATA NOTIF:', data._getstatusname(status), *args)
        if status == data.LIVE:
            self.live = True
        elif status == data.DELAYED:
            self.live = False


    ''' Store notification '''
    def notify_store(self, msg, *args, **kwargs):
        print('*' * 5, 'STORE NOTIF:', msg)
        

    ''' Order notification '''
    def notify_order(self, order):
        if order.status in [order.Completed, order.Cancelled, order.Rejected]:
            self.order = None
        print('-' * 50, 'ORDER BEGIN', datetime.datetime.now())
        print(order)
        print('-' * 50, 'ORDER END')


    ''' Trade notification '''
    def notify_trade(self, trade):
        print('-' * 50, 'TRADE BEGIN', datetime.datetime.now())
        print(trade)
        print('-' * 50, 'TRADE END')
        
        if not trade.isclosed:
            return

        self.log('OPERATION PROFIT, GROSS %.2f, NET %.2f' %
            (trade.pnl, trade.pnlcomm))

    ''' Next '''        
    def next(self):
        self.buySig = False
        self.sellSig = False
        self.idealBuySig = False
        self.idealSellSig = False
        
        # Calculate td sequential values if enough bars
        if len(self.dataclose) > self.p.candles_past_to_compare:
            
            # begin of sequence, bullish or bearish flip
            if self.dataclose[0] < self.dataclose[-self.p.candles_past_to_compare] and self.dataclose[-1] > self.dataclose[-(self.p.candles_past_to_compare + 1)]:
                self.bearishFlip = True
                self.bullishFlip = False
                self.tdsl = 0
                self.tdss = 0
                self.sellSetup = False
                self.sellCountdown = 0
            elif self.dataclose[0] > self.dataclose[-self.p.candles_past_to_compare] and self.dataclose[-1] < self.dataclose[-(self.p.candles_past_to_compare + 1)]:
                self.bullishFlip = True
                self.bearishFlip = False
                self.tdss = 0
                self.tdsl = 0
                self.buySetup = False
                self.buyCountdown = 0

            # sequence start
            if self.dataclose[0] < self.dataclose[-self.p.candles_past_to_compare] and self.bearishFlip:
                self.tdsl += 1
            elif self.dataclose[0] > self.dataclose[-self.p.candles_past_to_compare] and self.bullishFlip:
                self.tdss += 1
            
            # sequence intersection
            if self.tdsl == 9:
                self.buySig = (self.dataprimary.low[0] < self.dataprimary.low[-3] and self.dataprimary.low[0] < self.dataprimary.low[-2]) or (self.dataprimary.low[-1] < self.dataprimary.low[-2] and self.dataprimary.low[-1] < self.dataprimary.low[-3])
                self.bearishFlip = False
                self.tdsl = 0
                self.buySetup = True
                self.buyCountdown = 0
            elif self.tdss == 9:
                self.sellSig = (self.dataprimary.high[0] > self.dataprimary.high[-2] and self.dataprimary.high[0] > self.dataprimary.high[-3]) or (self.dataprimary.high[-1] > self.dataprimary.high[-3] and self.dataprimary.high[-1] > self.dataprimary.high[-2])
                self.bullishFlip = False
                self.tdss = 0
                self.sellSetup = True
                self.sellCountdown = 0
                
            # sequence countdown
            if self.buySetup:
                if self.dataprimary.close[0] <= self.dataprimary.low[-2]:
                    self.buyCountdown += 1
                if self.buyCountdown == 8:
                    self.buyVal = self.dataclose[0]
                elif self.buyCountdown == 13:
                    if self.dataprimary.low[0] <= self.buyVal:
                        self.idealBuySig = True
                    self.buySetup = False
                    self.buyCountdown = 0
                    
            elif self.sellSetup:
                if self.dataprimary.close[0] >= self.dataprimary.high[-2]:
                    self.sellCountdown += 1
                if self.sellCountdown == 8:
                    self.sellVal = self.dataclose[0]
                elif self.sellCountdown == 13:
                    if self.dataprimary.high[0] >= self.sellVal:
                        self.idealSellSig = True
                    self.sellSetup = False
                    self.sellCountdown = 0
                    
        # Trade only if no pending order and live data
        if not self.order and self.live:
            if not self.position:
                if self.buySig:
                    if self.p.stoploss_dist > 0:
                        stoplossPrice = self.dataprimary.close[0] - self.p.stoploss_dist
                        self.buy_bracket(stopprice=stoplossPrice, exectype=bt.Order.Market)
                    else:
                        self.buy()
                elif self.sellSig:
                    if self.p.stoploss_dist > 0:
                        stoplossPrice = self.dataprimary.close[0] + self.p.stoploss_dist
                        self.sell_bracket(stopprice=stoplossPrice, exectype=bt.Order.Market)
                    else:
                        self.sell()
            else:
                if self.position.size > 0 and (self.sellSig or self.idealSellSig):
                    self.close()
                elif self.position.size < 0 and (self.buySig or self.idealBuySig):
                    self.close()
                    
        # Log the closing price of the series from the reference
        self.log("Close {} TD SeqShort {} TD SeqLong {} TD CDBuy {} TD CDSell {}".format(self.dataclose[0], self.tdss, self.tdsl, self.buyCountdown, self.sellCountdown))

class observer(btobserver):
    plotinfo = dict(plot=True, subplot=False, plotlegend=False, 
                    plotlinelegend=False, plotlinevalues=False, 
                    plotabove=True, plotname="TD Sequential",)
    lines = ('buy', 'buy_ideal', 'sell', 'sell_ideal', 'seq_up_1', 
            'seq_up_2', 'seq_up_3', 'seq_up_4', 'seq_up_5', 
            'seq_up_6', 'seq_up_7', 'seq_up_8', 'seq_up_9', 
            'seq_up_1_b', 'seq_up_2_b', 'seq_up_3_b', 'seq_up_4_b', 
            'seq_up_5_b', 'seq_up_6_b', 'seq_up_7_b', 'seq_up_8_b', 
            'seq_up_9_b', 'seq_up_10_b', 'seq_up_11_b', 'seq_up_12_b', 
            'seq_down_1', 'seq_down_2', 'seq_down_3', 'seq_down_4', 
            'seq_down_5', 'seq_down_6', 'seq_down_7', 'seq_down_8', 
            'seq_down_9',
            'seq_down_1_s', 'seq_down_2_s', 'seq_down_3_s', 'seq_down_4_s', 
            'seq_down_5_s', 'seq_down_6_s', 'seq_down_7_s', 'seq_down_8_s', 
            'seq_down_9_s', 'seq_down_10_s','seq_down_11_s', 'seq_down_12_s',)
            
    plotlines = dict(
        buy_ideal = dict(marker='^', markersize=10.0, color='green',),
        buy = dict(marker='2', markersize=10.0, color='green',),
        sell_ideal = dict(marker='v', markersize=10.0, color='maroon',),
        sell = dict(marker='1', markersize=10.0, color='maroon',),
        seq_up_1 = dict(marker='$1$', markersize=6.0, color='black', ls='',),
        seq_up_2 = dict(marker='$2$', markersize=6.0, color='black', ls='',),
        seq_up_3 = dict(marker='$3$', markersize=6.0, color='black', ls='',),
        seq_up_4 = dict(marker='$4$', markersize=6.0, color='black', ls='',),
        seq_up_5 = dict(marker='$5$', markersize=6.0, color='black', ls='',),
        seq_up_6 = dict(marker='$6$', markersize=6.0, color='black', ls='',),
        seq_up_7 = dict(marker='$7$', markersize=6.0, color='black', ls='',),
        seq_up_8 = dict(marker='$8$', markersize=6.0, color='black', ls='',),
        seq_up_9 = dict(marker='$9$', markersize=6.0, color='black', ls='',),
        seq_up_1_b = dict(marker='$1$', markersize=6.0, color='blue', ls='',),
        seq_up_2_b = dict(marker='$2$', markersize=6.0, color='blue', ls='',),
        seq_up_3_b = dict(marker='$3$', markersize=6.0, color='blue', ls='',),
        seq_up_4_b = dict(marker='$4$', markersize=6.0, color='blue', ls='',),
        seq_up_5_b = dict(marker='$5$', markersize=6.0, color='blue', ls='',),
        seq_up_6_b = dict(marker='$6$', markersize=6.0, color='blue', ls='',),
        seq_up_7_b = dict(marker='$7$', markersize=6.0, color='blue', ls='',),
        seq_up_8_b = dict(marker='$8$', markersize=6.0, color='blue', ls='',),
        seq_up_9_b = dict(marker='$9$', markersize=6.0, color='blue', ls='',),
        seq_up_10_b = dict(marker='$10$', markersize=6.0, color='blue', ls='',),
        seq_up_11_b = dict(marker='$11$', markersize=6.0, color='blue', ls='',),
        seq_up_12_b = dict(marker='$12$', markersize=6.0, color='blue', ls='',),
        seq_down_1 = dict(marker='$1$', markersize=6.0, color='black', ls='',),
        seq_down_2 = dict(marker='$2$', markersize=6.0, color='black', ls='',),
        seq_down_3 = dict(marker='$3$', markersize=6.0, color='black', ls='',),
        seq_down_4 = dict(marker='$4$', markersize=6.0, color='black', ls='',),
        seq_down_5 = dict(marker='$5$', markersize=6.0, color='black', ls='',),
        seq_down_6 = dict(marker='$6$', markersize=6.0, color='black', ls='',),
        seq_down_7 = dict(marker='$7$', markersize=6.0, color='black', ls='',),
        seq_down_8 = dict(marker='$8$', markersize=6.0, color='black', ls='',),
        seq_down_9 = dict(marker='$9$', markersize=6.0, color='black', ls='',),
        seq_down_1_s = dict(marker='$1$', markersize=6.0, color='blue', ls='',),
        seq_down_2_s = dict(marker='$2$', markersize=6.0, color='blue', ls='',),
        seq_down_3_s = dict(marker='$3$', markersize=6.0, color='blue', ls='',),
        seq_down_4_s = dict(marker='$4$', markersize=6.0, color='blue', ls='',),
        seq_down_5_s = dict(marker='$5$', markersize=6.0, color='blue', ls='',),
        seq_down_6_s = dict(marker='$6$', markersize=6.0, color='blue', ls='',),
        seq_down_7_s = dict(marker='$7$', markersize=6.0, color='blue', ls='',),
        seq_down_8_s = dict(marker='$8$', markersize=6.0, color='blue', ls='',),
        seq_down_9_s = dict(marker='$9$', markersize=6.0, color='blue', ls='',),
        seq_down_10_s = dict(marker='$10$', markersize=6.0, color='blue', ls='',),
        seq_down_11_s = dict(marker='$11$', markersize=6.0, color='blue', ls='',),
        seq_down_12_s = dict(marker='$12$', markersize=6.0, color='blue', ls='',),
    )
    
    
    params = (
        ('bardist', 0.00005),  
    )
    
    def __init__(self):
        pass
        
    def next(self):
        # begin of sequence for long position
        if self._owner.tdsl > 0 and self._owner.tdsl < 10:
            name = 'seq_up_%d' % self._owner.tdsl
            attr = getattr(self.lines, name, False)
            attr[0] = self.data_low[0] - self.p.bardist
        
        # begin of sequence of short position
        elif self._owner.tdss > 0 and self._owner.tdss < 10:
            name = 'seq_down_%d' % self._owner.tdss
            attr = getattr(self.lines, name, False)
            attr[0] = self.data_high[0] + self.p.bardist
            
        if self._owner.buySetup:
            if self._owner.buyCountdown > 0 and self._owner.buyCountdown < 13:
                name = 'seq_up_%d_b' % self._owner.buyCountdown
                attr = getattr(self.lines, name, False)
                attr[0] = self.data_low[0] - (2 * self.p.bardist)
        elif self._owner.sellSetup:
            if self._owner.sellCountdown > 0 and self._owner.sellCountdown < 13 and self._owner.sellSetup:
                name = 'seq_down_%d_s' % self._owner.sellCountdown
                attr = getattr(self.lines, name, False)
                attr[0] = self.data_high[0] + (2 * self.p.bardist)
        
        if self._owner.buySig:
            self.lines.buy[0] = self.data_low[0] - (2 * self.p.bardist)
        elif self._owner.sellSig:
            self.lines.sell[0] = self.data_high[0] + (2 * self.p.bardist)
        elif self._owner.idealBuySig:
            self.lines.buy_ideal[0] = self.data_low[0] - (3 * self.p.bardist)
        elif self._owner.idealSellSig:
            self.lines.sell_ideal[0] = self.data_high[0] + (3 * self.p.bardist)
            

To plot this, you need to add the observer

Hi

I would like to tell you my opinion about your considerstions since i had the same questions a few weeks ago and my decision went against zipline, quantconnect and others for backtrader.

After understanding the basics it was very easy to get it working and the use of data is really great in backtrader.

This is very subjective but backtrader seems for me very logical, flexibel, extendible and easy to start with. Easier than other compareable solutions.

At the end i will be mostly work with strategies, indicators, data, optimization and trades. All this is working in backtrader. The code of backtrader is nice, the way it is working is very logical and it is easy to adjust it for your needs.

So maybe try to look for your needs and compare them with the available solutions.

For the long term thinking, its not said, the companies will follow the same path, they are going at the moment so it was more important for me to get the software that fits more the needs at the moment. Also it is not that complicated to rewrite your strategies from backtrader to others, if it would be needed.