2 {{ upvoteCount | shortNum }}
2 {{ upvoteCount | shortNum }}

Accuracy of Forex Robots Testing???

by robottop
1 replies
Before I start this topic would like to apologize about my English, it's not perfect and I don't want hire any editor, would like to write my self...
I would like to start this thread about most important subject in Forex Automated Systems. There is hundreds of robots on market, everybody showing incredible results with back test performance, question???
How I can be sure and confidence in back testing? Can I trust it? Is it accurate?
I have lots of experience in Forex Automated Systems. When I create my first robot and it shows me incredible result, I was shock, exited, thinking wow, I will start making hundred of thousands dollars. I connected to my real account and realize that after one month statement on my real account and performance on tester are different...I would like to make story short. It took eight month to find out and resolve the problem. If you don't have Tick Data uploaded to your MT4, you will have WRONG RESULTS...for example: there is up to 200 - one minute candle gaps each month and I am not talking about big gaps...I purchased most popular Expert Advisers selling on market with only one reason, run them on my tester and proof to their owners that result they show - WRONG...I do not want to blame them, but would like to warn any one who want to buy or create automated software that biggest issue they will have - ACCURATE BACK TESTING...
#accuracy #forex #robots
  • Profile picture of the author Alexa Smith
    Banned
    Originally Posted by robottop View Post

    How I can be sure and confidence in back testing? Can I trust it? Is it accurate?
    Actually these aren't even the most important questions, really, you know?

    All the backtesting could be 100% reliable and accurate, and it could still be a useless piece of junk because it's backfitted. Using computer-power to come up with indicator-based parameters which turned out to have traded successfully on the basis of historical price observations isn't too difficult at all.

    Unfortunately, it doesn't mean that they have any predictive value, though.

    Correlation is not causation.
    {{ DiscussionBoard.errors[3346256].message }}

Trending Topics