Reverse Strategy Ideas?
Page 1 of 843 123 LastLast
Results 1 to 10 of 21

Thread: Reverse Strategy Ideas?

  1. #1
    Hi, I just have some ideas that I would like to put forth and my logic behind them and the results from testing them that I would like to talk that are currently baffling me to no end and would like to hear you all of your ideas on why this is or perhaps how I can address this problem and determine if its even worth pursuing. So this is not a matter of my skill or ability merely am not understanding this is not working when every object of logic states that it should, I'm a developer. So heres the problem...

    I'm sure for those of us that have coded many EAs more often than not they blow off the account pretty quickly the first two or three runs and should not most of the EAs perform similarly even after months of programming. I know this is a function of the test quality in mt4, but with 90% quality from your brokers feed should provide you for the most part consistent results whether it be losing or winning. So here is what I suggest

    Given a Strategy that merely enters the market in a way based on a set of entrance principles. It's SL: 10pips, TP: 5pips and the egy only lets the transaction run till either of these levels are hit

    The Results:
    Trades: 679. Wins: 254 (37 percent ) Losses: 425 ( 63 percent )

    Obviously this egy would blow off an account pretty fast. So my thought is that. Why not when we get a signal, simply do the opposite of everything in the egy. And instead of 10pip SL we use a 5pip stop and instead of 5pip TP we use a 10pip TP so today those prior stop outs are cut in half pip wise. And all those wins are bigger and occur more often since we are currently entering in the contrary direction. So this was coded by me. Exact same entrance rules, identical timeframe and dates, identical data and backtest quality even checked that it was entering exact same prices and times .

    The Results:
    Trades: 679. Wins: 31 (4.5%) Losses: 648 (95.5%)

    This does not make any sense whatsoever to me. I know 90% quality is not 100% so its hard to demonstrate this exactly, but these are pretty drastic differences When technically all the preceding losses should now be wins and vice versa thus our new triumph% ought to have been (63 percent ) and this ought to be a very profitable egy. Any thoughts or ideas would be appreciated. Thanks!

  2. #2
    I've come across similar problems.


    The principal reason is that the market is in a specific state. It only happens to work for that setup. Was that test EU during a few of its trending periods that are longer?

  3. #3
    Yep it was EU, Jan 01 2009 - Today. I've completed it over several phases under different market conditions and the exact same outcomes. I understand the market could possibly be in a condition at the point of the set up. But still logically by performing the reverse should not the losses now be wins? If everything else is a continuous (dates, market requirements in those dates, entry price etc)
    I understand this may not work on another period. However, in an isolated test with everything staying constant except for changing the direction one passes and switching the SL and TP the outcomes should be precise reverse. Why this isn't true, Have you any idea?

  4. #4
    I don't have any way to refute it if that is what you're saying. Your concept as I see it's accurate.

    The point where this concept falls apart is exactly what happens ahead.

  5. #5
    You might have a situation which has an element of curve-fitting to it. Its hard to say without knowing more. 679 trades may (or may not) be a lot of transactions (over time, I mean) based on your own egy. It doesnt strike me.

  6. #6
    679 trades in 22 days? (Less weekends, 15). ?! What makes a sample size that is fantastic? Or sample density for that matter.

  7. #7
    In response to Post 4:

    Right, and I know that but am currently just worried about why it does not function when used within precisely the exact same period, not forward testing. For that it should do precisely as anticipated but it does not. Do you have any clue why this would not work using the very same dates and information? Perhaps you have tried it before? Or you would know of a way to backtest, given all the EAs you've coded? I currently can get 90-92% backtest from IBFX information centre. And of course the 90 percent 1m data from alpari. Is right only a matter of this quality even though it is 90% .

  8. #8
    Actually your result makes sense. Do you account for spread in your stop loss? If not... this is readily explained.

    Stop Loss

    supposing spread is fixed at 2 pips during. Before reversing. Your stop loss was 10 pips that means price needs to move 8 pips from one to strike.

    Once you reversed, the stop loss became 5 pips, meaning price only need to move 3 pips from you for one to get quit out.

    Take Profit

    before reversing. TP was put at 5 pips. After accounting for spread, price needs to transfer 7 pips in your favor to hit TP.

    After reversing, TP is now 10 pips. That means 12 pips is required to move in your favor in order to reach TP.

    Summary

    after reversing, you've increased your chances of hitting stop with a lower stoploss. And diminished your chances of hitting at TP with a TP target.
    Get the picture?

    Respect,
    Zen

  9. #9
    Something that is not symmetric can't be reversed by you. Period.

  10. #10
    Quote Originally Posted by ;
    679 transactions in 22 days? (Less weekends, 15). ?! What then makes a fantastic sample size? Or sample density for this issue.
    Sample dimensions should be measured in time, not necessarily transactions. Market conditions (and volatility) go past 22 days

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
This website uses cookies
We use cookies to store session information to facilitate remembering your login information, to allow you to save website preferences, to personalise content and ads, to provide social media features and to analyse our traffic. We also share information about your use of our site with our social media, advertising and analytics partners.