I’ve been running a backtest in Portfolio Trader after the session close to tune my strategy and compare with what actually occurred during the session and periodically the backtest shows an entry that wasn’t made in realtime, and I can confirm the signal conditions were met. I assume the same risk capital rules are applied in backtesting and realtime, so what are possible reasons for this? The logs show nothing unusual and I can confirm orders weren’t generated and rejected by the broker or not filled.
Many thanks
Portfolio Trader discrepancies [SOLVED]
Re: Portfolio Trader discrepancies [SOLVED]
As this phenomenon continues I'm thinking it may be that historical bars sometimes print differently than the real-time ticks, enough so that portfolio trader will show trades in backtest that didn't happen and vice versa, because it's not always several orders at once. Does this idea make any sense?
-
- Posts: 19
- Joined: 19 Apr 2013
- Location: Bucharest
- Has thanked: 1 time
- Been thanked: 1 time
- Contact:
Re: Portfolio Trader discrepancies
Hello ,
i spend more than 3 years with portfolio ..is just a nice tool for testing but is not the one i recommend to use in live (ex. in case of one single error >100 times ..portfolio will be switch of ..from MC .. this is just one ex..others ..never you will have same results as in the charts .
AI
i spend more than 3 years with portfolio ..is just a nice tool for testing but is not the one i recommend to use in live (ex. in case of one single error >100 times ..portfolio will be switch of ..from MC .. this is just one ex..others ..never you will have same results as in the charts .
AI
Re: Portfolio Trader discrepancies
Hi Al, sorry to hear that, I hope my experience is different. So far I have no problem with execution, it just frequently doesn't match backtest. I think I just have to treat backtest as theoretical. BTW I think you can de-activate the cancel autotrade after errors in properties if you want to though it's probably not a great idea.
Cheers
Cheers