• Registered: 2011-12-21
  • Posts: 7

Topic: Reliability of strategy results

Hey,

after reading through the post about the 1 minute EURUSD 65,000 pips strategy (forexsb.com/forum/topic/2544/65000-pips … down-2311/) and the comments that the results of the FSB aren't reliable because the Forex Strategy Trader doesn't behave like the Forex Strategy Builder I am wondering whether I can trust in the results of a strategy developed with the FSB.

Do these problems that the results are inconsistent only occur when using strategies for shorter timeframes? Has someone recognized similar problems with higher timeframes ?

In the post they used the HTF indicator which is told to be not working correctly. So is it also a problem of using non-standard-installation indicators with the FSB??
I am testing a strategy with the Aroon Histogram and the CCI. Are there any known problems with these indicators ??

Best regards,
Peter.

  • krog
  • Licensed Member
  • Offline
  • Registered: 2010-02-26
  • Posts: 407

Re: Reliability of strategy results

The HTF  have a bug where they have values in FST in live trading, but then different values when you get the same data and do a backtest with FSB. The HTF indicators were a good example of this. The values depended on the time of the first bar of data. Since it is very unlikely you would turn on your FST-MT4 autotrading at the same time as the first bar of your downloaded data, the 2 would have very widely different values. Use the WTF indicators instead:
http://forexsb.com/forum/topic/2438/wid … ndicators/
These are more reliable, I haven't heard many complaints about different values. (There will be a version 4 coming soon, when the forum gets fixed and allows attachments.)

A lot of the custom indicators are usually contributed by users (like me), and they might miss some areas of testing. It is very difficult to test every possible use before posting, usually someone uses it in some way I did not test and they find a bug, then have to post a new version to fix it. I try to put a "beta" or something in the name so future people who download it will see. But then "v1" or "v2" which should be working could still have bugs. I try to remember to go back and remove older bug versions, save future users some trouble.

Always best to run forward in a demo account for a while to test first. A bug like the HTF bug will show up very quickly, some bugs might take a larger set of demo trades to show up.

  • Registered: 2011-12-21
  • Posts: 7

Re: Reliability of strategy results

So in general, it isn't common that the FS Builder shows different entries and exits for trades than the FS Trader does in demo or live accounts when using the same data source?

It's just a problem of the custom indicators?

What about the problems with timeframes like M1 ?

  • Registered: 2012-03-11
  • Posts: 1

Re: Reliability of strategy results

Peter83 wrote:

So in general, it isn't common that the FS Builder shows different entries and exits for trades than the FS Trader does in demo or live accounts when using the same data source?

It's just a problem of the custom indicators?

What about the problems with timeframes like M1 ?




any update?

  • From: Australia
  • Registered: 2011-07-01
  • Posts: 431

Re: Reliability of strategy results

any update?

Update of what?

  • From: Brazil
  • Registered: 2012-04-12
  • Posts: 15

Re: Reliability of strategy results

I believe that it is of extreme importance a greater quality in modeling tests. As for example in MT4, with a 99% modeling, already here in FSB my biggest modeling was Scanned MQ 57.72%. Downloaded the historical data by TrueFX and still can't get a better modeling.

Unless the Scanned MQ, differs from MT4 modeling, wanting to believe that the first is better than the second.