Hi everyone,

I published a little contribution in my Github. A example of usage of Backtrader with Binance integration. Using the CCXT lib for broker and feed.

The base strategy uses a simple RSI + SMA filter, with stop loss of 3%.

Feel free to contribute and improve the example: https://github.com/rodrigo-brito/backtrader-binance-bot

Best regards,
Rodrigo Brito

Hi @Noah-Bastola,

I'm using backtrader for bitcoin/usdt trades in production. The easy way is integrate with binance API through CCXT Store lib. A simple example is available below:

https://gist.github.com/rodrigo-brito/3b0fca2487c92ad97869247edd5fd852

It is an interesting feature. Sometimes I want to visualize critical points such as "stop loss".

Hi everyone.
About the problem with the position calculation in my simple example. I think it is a problem with bt-ccxt-store. I create a workarround, ignoring the position value and store this value manually inside the Strategy Class. Other customization is the buy() and sell() functions, the sizer does not work very well with binance. When the fee is charged in BNB, it does not include the wallet value. I'm using this class to create custom buy and sell operations. For now, I'm using backtrader to get the historical information and indictators. All other operations I'm doing manually.
Custom code: https://gist.github.com/rodrigo-brito/8c82020f04e946e3f0c39c7243cfe1ee

I implemented the MFI for Backtrader, but I think this can be improved:

#!/usr/bin/env python3

import math
import backtrader as bt

class MFI(bt.Indicator):
    lines = ('mfi', 'money_flow_raw', 'typical', 'money_flow_pos', 'money_flow_neg')

    plotlines = dict(
        money_flow_raw=dict(_plotskip=True),
        money_flow_pos=dict(_plotskip=True),
        money_flow_neg=dict(_plotskip=True),
        typical=dict(_plotskip=True),
    )

    params = (
        ('period', 14),
    )

    def next(self):
        typical_price = (self.data.close[0] + self.data.low[0] + self.data.high[0]) / 3
        money_flow_raw = typical_price * self.data.volume[0]
        
        self.lines.typical[0] = typical_price
        self.lines.money_flow_raw[0] = money_flow_raw

        self.lines.money_flow_pos[0] = money_flow_raw if self.lines.typical[0] >= self.lines.typical[-1] else 0
        self.lines.money_flow_neg[0] = money_flow_raw if self.lines.typical[0] <= self.lines.typical[-1] else 0

        pos_period = math.fsum(self.lines.money_flow_pos.get(size=self.p.period))
        neg_period = math.fsum(self.lines.money_flow_neg.get(size=self.p.period))
        
        if neg_period == 0:
            self.lines.mfi[0] = 100
            return

        self.lines.mfi[0] =  100 - 100 / (1 +  pos_period / neg_period)

I used two auxiliary lines to sum negative/positive values when typical_price change. There are some alternative to reduce it and compare two different lines given a period?

Talib MACD returns a different value than std MACD with re-sample data.
A report of @mr-m0nst3r

Using this example: https://gist.github.com/rodrigo-brito/4f779001647ce1ce7235a7d254b7e75b
With this dataset: https://github.com/rodrigo-brito/backtrader-binance-bot/blob/master/dataset/binance_nov_18_mar_19_btc.csv

Without re-sample, it is ok:

EMA12 = 6960.21 EMA26 = 6949.12 DIFF = 11.09
STD MACD = 11.09
TALIB MACD = 11.09
--------------
EMA12 = 6961.31 EMA26 = 6950.47 DIFF = 10.84
STD MACD = 10.84
TALIB MACD = 10.84
--------------
EMA12 = 6961.38 EMA26 = 6951.31 DIFF = 10.08
STD MACD = 10.08
TALIB MACD = 10.08
--------------

But, with a re-sample of 30 minutes, it returns a different value:

cerebro.resampledata(data, timeframe=bt.TimeFrame.Minutes, compression=30)

--------------
EMA12 = 7031.31 EMA26 = 7057.54 DIFF = -26.23
STD MACD = -26.23
TALIB MACD = -83.05
--------------
EMA12 = 7012.22 EMA26 = 7046.41 DIFF = -34.19
STD MACD = -34.19
TALIB MACD = -90.54
--------------
EMA12 = 7004.46 EMA26 = 7040.14 DIFF = -35.68
STD MACD = -35.68
TALIB MACD = -88.64
--------------

The graphic result is also different:

Hi @diamonddogs95, I used https://github.com/xFFFFF/Gekko-Datasets to generate a database in SQLite. Then I open it and run this query https://github.com/rodrigo-brito/backtrader-binance-bot/blob/master/dataset/queries.sql

Hi, @pfederra

1. I have the same issue in notify_order(), it works only in simulations to me.
2. I think the fork has a lot of issues, but work very well for the main features, like order and feed.
3. The get positions does not work for me very well, it has a precision issue. I'm using a function to get my wallet ballance and settting a custom sizer. I'm also defining a global variable with the last operation (BUY/SELL) to control the next operation. It is a working arrond, but working in live.

@coolernax I don't persist the feed or strategies positions. The feed loads a delayed history when started. About the trade position, I'm start always with no funds in crypto, the first operation should be a buy. Maybe it can be improved with a simple flag in the script.

It is an interesting feature. Sometimes I want to visualize critical points such as "stop loss".

I published a simple example with Binance integration: https://github.com/rodrigo-brito/backtrader-binance-bot

Hi everyone,

I published a little contribution in my Github. A example of usage of Backtrader with Binance integration. Using the CCXT lib for broker and feed.

The base strategy uses a simple RSI + SMA filter, with stop loss of 3%.

Feel free to contribute and improve the example: https://github.com/rodrigo-brito/backtrader-binance-bot

Best regards,
Rodrigo Brito

Hi @pfederra. I'm receiving zero values too. I think it is a issue related to bt-ccxt-store.