Topic: FXT file for backtest
How do you use FXT file data history , instead of using csv in Forexsb?
- muplayer2000
- Licensed Member
- Offline
- Blaiserboy
- Junior Part-Time Aspiring Space Cadet
- Offline
Re: FXT file for backtest
I guess no one saw this post.....
Hopefully someone will reply to it.
- Popov
- Lead Developer
- Offline
Re: FXT file for backtest
I have no idea what is a FXT file. I do not use such files. Can you share more info for it? What program exports FXT files?
- Blaiserboy
- Junior Part-Time Aspiring Space Cadet
- Offline
Re: FXT file for backtest
FXT is created by MT4, it is what is used for backtesting......
Tickstory loads MT4 with FXT as well as CSV
- Popov
- Lead Developer
- Offline
Re: FXT file for backtest
I see now. Thank you.
MT4 uses the bar data files to generate ticks for backtesting. These ticks are completely useless in FSB. The interpolation method used by MT generates the ticks depending on the bar shape and you can find a "Holy Grail" expert that exploits that interpolation in order to show highly overestimated results.
FSB have used such approach in the late 2006. Many holy Grails was reported in that time. Ti fix the problem, I added dynamic interpolation methods that make much more reliable backtest. These methods reports "ambiguous" bars when the data do not provide enough information for 100% strict execution order.
From the current methods, the "Shortest" one is near to the MT interpolation. Eventually I may provide MT backtesting method in FSB in the future major releases. However I do not see any particular benefits from using it except eventually for receiving more close backtetsing results with MT.
- Blaiserboy
- Junior Part-Time Aspiring Space Cadet
- Offline
Re: FXT file for backtest
I seem to trust the data examination in FSB. Without knowing a lot about data structure but with knowing the mess of MT4, I prefer to use the data provided by the FSB data download as it is about as clean as we can find.
We do have to exp[ect there to be a discrepancy between MT4 and FSB from time to time because of the nature of the crap data in MT4. That is one good reason to stay away from a time frame lower than 60 minutes.
The best results that I get are 4 hours and 8 hours. Seemed weird to me at first to get better test results from the higher time frames but they give better data points.