it's impossible, use ta-lib ...

@vladisld Uh ... Thnx, I forgot about the lot size.

@run-out thnx, I understand

i misunderstood the order creation and interaction with a broker class

@run-out in github, i searched and found an example (https://github.com/bartosh/backtrader/blob/ccxt/backtrader/brokers/ccxtbroker.py)

    def _submit(self, owner, data, exectype, side, amount, price, params):
        order_type = self.order_types.get(exectype)
        _order = self.store.create_order(symbol=data.symbol, order_type=order_type, side=side,
                                         amount=amount, price=price, params=params)
        order = CCXTOrder(owner, data, amount, _order)
        self.notify(order)
        return order

Why is the order being re-created?

@run-out thnx, I understand

i misunderstood the order creation and interaction with a broker class

@run-out in github, i searched and found an example (https://github.com/bartosh/backtrader/blob/ccxt/backtrader/brokers/ccxtbroker.py)

    def _submit(self, owner, data, exectype, side, amount, price, params):
        order_type = self.order_types.get(exectype)
        _order = self.store.create_order(symbol=data.symbol, order_type=order_type, side=side,
                                         amount=amount, price=price, params=params)
        order = CCXTOrder(owner, data, amount, _order)
        self.notify(order)
        return order

Why is the order being re-created?

@run-out How to tracking position - understandable.

But, if implemented broker class is, how to correctly set position (size, upclosed, upopened)?

it's impossible, use ta-lib ...

@run-out thnx the answer :)

In documentation for the Strategy class, section Member Attributes has:

....
position: actually a property which gives the current position for data0.
Methods to retrieve all possitions are available (see the reference)

This attribute, instance of class Position.
When the broker (class broker) receives an order to open a position, this attribute must be correctly set.

Addition, notify_trade sending only after call order.execute method. Its correctly.

How correctly setting self.position (upopened, size ... etc)?
I'm still :) looking in source ...

I made my class broker, orders are sent and executed - ok.
But, how to set the value of attributes position correctly?
i see several methods:

1. using notify_trade, and thus set the value attributes position
2. using self.data[0], and thus set the value attributes position

But how do it right in the broker class?

@vladisld Uh ... Thnx, I forgot about the lot size.

@vladisld thnx, worked.

But I don't understand what the relationship is cash on the broker and not open BUY position ...

Could you tell me why this is so ?! Or what to read in the documentation?

Hi All,

Can someone comments, I don't understand behavior of the class CrossOver.
Simple strategy:
if EMA (30) Сrosses UP WMA(30) - buy
if EMA (30) Сrosses Down WMA(30) - sell

import sys
import backtrader as bt
from datetime import datetime, timedelta

import backtrader.indicators as btind
import backtrader.feeds as btfeed

from backtrader_plotting import Bokeh
from backtrader_plotting.schemes import Tradimo

import argparse

import logging

logger = logging.getLogger(__name__)
logging.basicConfig(stream=sys.stderr,
                    level=logging.DEBUG,
                    format='%(asctime)s [%(levelname)s][module: %(module)s.py] %(funcName)s: %(message)s',
                    datefmt='%Y-%m-%d %H:%M:%S'
)

class MyTradimo(Tradimo):
    def _set_params(self):
        super()._set_params()
        # Disable the output of Volume on the displayed chart
        self.volume=False

class CommFixed(bt.CommInfoBase):
    params = (
        ('stocklike', False),
        ('commtype', bt.CommInfoBase.COMM_FIXED),
    )

class TestedStrategy(bt.SignalStrategy):
    def __init__(self):
        self.order = None

        self.ema = btind.ExponentialMovingAverage(self.data.close)
        self.lwma = btind.WeightedMovingAverage(self.data.close)

        self.buysell = btind.CrossOver(self.lwma, self.ema)

    def next(self):
        if self.buysell < 0:
            self.sell(exectype=bt.Order.Market)
        elif self.buysell > 0:
            self.buy(exectype=bt.Order.Market)

def parse_args(pargs=None):

    parser = argparse.ArgumentParser(
        formatter_class=argparse.ArgumentDefaultsHelpFormatter,
        description='Sample for Signal concepts')

    parser.add_argument('--data', required=False,
                        default='csv/BTCUSD_M5_2021_01_28__21_35-2021_02_02__21_25.csv',
                        help='Specific data to be read in')

    parser.add_argument('--fromdate', required=False,
                        default=None,
                        help='Starting date in YYYY-MM-DD[THH:MM:SS] format')

    if pargs is not None:
        return parser.parse_args(pargs)

    return parser.parse_args()

def runstrat(args=None):
    args = parse_args(args)

    if args.fromdate:
        fmtstr = '%Y-%m-%d'

        dtsplit = args.fromdate.split('T')
        if len(dtsplit) > 1:
            fmtstr += 'T%H:%M:%S'

        start_date = datetime.strptime(args.fromdate, fmtstr)
    else:
        start_date = datetime(2021,1,31,23,00)

    end_date = datetime(2021,2,1,18,00)

    logger.info("Start: {} End: {}".format(start_date, end_date))

    data = btfeed.GenericCSVData(
        dataname='BTCUSD_M5_2021_01_28__21_40-2021_02_02__21_35.csv',
        dtformat=('%Y.%m.%d %H:%M:%S'),
        timeframe=bt.TimeFrame.Minutes,
        compression=5,
        fromdate=start_date,
        todate=end_date,
    )

    cerebro = bt.Cerebro(stdstats=False)
    cerebro.adddata(data)
    cerebro.addstrategy(TestedStrategy)
    cerebro.broker.setcash(1000)    
    cerebro.broker.addcommissioninfo(CommFixed, "CommFixed")
    cerebro.addobserver(bt.observers.BuySell)

    cerebro.run()

    b = Bokeh(style='bar', tabs='multi', scheme=MyTradimo())
    cerebro.plot(b)

if __name__ == '__main__':
    runstrat()

We are doing two tests:

1. BTC, M5, Start: 2021-01-31 22:35:00 End: 2021-02-01 18:00:00
2. BTC, M5, Start: 2021-01-31 22:40:00 End: 2021-02-01 18:00:00

We see such results:

1. OK
   

2. Where is BUY?!
   

It looked more complicated in the code - I simplified everything as much as possible for an example.