Hello, I have an algorithm ready for paper trading before going real, I have tested it with BTCUSD as ticker, BTC, and USD in the portfolio cashbook. However, following the recommendations on the documentation I have to use TESTUSD as account currency to paper trade on Bitfinex. For the algorithm to work (I mean the only combination of Symbols that allowed me to run a backtest without conversion issues), I had to change the ticker and cashbook symbols to the following:

    account_currency = 'TESTUSDT' # or 'TESTUSD'
   cashbook_btc = 'TESTBTC'
   cashbook_usd = 'TESTUSDT' # or 'TESTUSD'
   myticker = 'TESTBTCTESTUSDT' # or 'TESTBTCTESTUSD'

Nevertheless, both options show some signs of erratic data on the backtest, so I wonder if using that convention will give me a meaningful approximation of how my model would perform in reality. I am sorry for not being able to provide a backtest but attached are the figures for the two cases. All with the BrokerageName.Bitfinex:

TESTUSD:

197487_1663364272.jpg

TESTUSDT:

197487_1663364146.jpg

USD (As long as I know this one doesn't have issues):

197487_1663364251.jpg

(I) Has anyone used this ticker?

(ii) Did that give you meaningful insights? 

(iii) Is it supposed to be used only in paper live trade and not in backtests?

(iv) Should I use the QuantConnectBrokerage model to test the algorithm?

Thanks in advance. I already reported the issue with the TESTBTC here.

 

 

Author