Simple Moving Average Crossover Strategy

Forums ProRealTime English forum ProOrder support Simple Moving Average Crossover Strategy

  • This topic has 95 replies, 12 voices, and was last updated 1 year ago by avatarCRISRJ.
Viewing 15 posts - 16 through 30 (of 96 total)
  • #187308

    Have added a volatility filter to the original 15 min version, which was optimized over the last 50k bars.  Reduces draw down and increases performance so net happy with this minor addition.  Will rerun the Walk Forward process but expect little change.

    Am also testing the 5 min version which nonetheless kindly posted as this too looks very promising.

    #187341

    I advice you don’t surprise yourself with backtest, or maybe to be sure of the robustness of your code, test Long on some sample down trends and see how it reacts or maybe try with Walk Forward strategy, and I totally agree with you the simpler the better

    this answer is for phoentzs, sorry

    #187343

    on the Backtest you only buy half a point ?

    I do backtests at the minimum position size, in this case 0.5 pp

    This is partly because the M0ney Management is based on the max drawdown @ min position size (line 26)

    As for your further comments, my basic attitude is that no matter how good a backtest might look, I assume that it’s completely wrong and totally unreliable, will probably crash and burn.

    Run it for months in demo, study the entries and exits – see if it behave the way you expected.

    This one got off to a good start yesterday, 4 winning trades, ~400 pts.

    maybe you can try to optimise with the Walk Forward method, it’s maybe will help you, I will study if in the future deeply maybe, now I’m not at this level

    #187395

    @samsampop

    I haven’t a chance to run your new version but one thing you should change is line 69, should be:

    PriceDistance = 4 * pipsize

    this is the IG minimum stop distance, so it will change from one instrument to the next.

    1 user thanked author for this post.
    #188786

    Quick update – strategy has stayed out of the market since Jan 20th and good job too for a long only strategy.

    #189149

    here’s a revised 5m version – better long term performance with lower drawdown, positionsize = 1

    altered the 4h filter and added StochasticRSI

    took one loss in Feb, but otherwise managed the Ukraine mess fairly well (so far)

    5 users thanked author for this post.
    #189251

    Hi nonetheless

    Quick question if I may – how often would you re-optimize the variables for this latest strategy that you have kindly posted?

    Thanks very much

    #189257

    I don’t have a fixed policy on that. I rework things as and when I have an idea that might improve some aspect. Obviously, if you’re optimizing on max data on a 12 year backtest you’d hope that those numbers would be worth keeping. Someone else might keep the basic structure of the algo but optimize for maybe 6 years where most of the profit has been made –  I guess then you’d want to refresh it more often. Who knows … 🤔

    1 user thanked author for this post.
    #189295

    thank you very much for your contribution nonetheless. I quickly looked at the code and tested to remove all code after trailing. was basically the same result.

    #189303

    the 3 functions after the trail are standard bits of code that i add to most algos, the effect varies from one to the next. In this case they collectively add about 3%

    If you’d rather not have an extra 3% then I suggest you switch them off immediately 😁

    Attached are the results with a constant exposure value of €10k, gives a better visual on the histogram and the curve.

    3 users thanked author for this post.
    #190505

    Hello @nonetheless, than you for your work. I am really confused, i’am not having same backtest as you, I am using V4.1 L, and i have add 1h at h1 and h2 because i’m in France. Can you share with us the last itf you have. Thank you very much.

    #190511

    I haven’t done any other work on that one, so v4.1L is the only itf I have. How much discrepancy are you seeing?

    1 user thanked author for this post.
    #190518

    @nonetheless Here is my backtest for the same period and Position size 1 MM off so you can see the huge difference :’D . I added also the parametres. Thanks for you help.

    #190523

    the only thing i can think of is the time settings in the DSD section. Try it like this, allowing for the Euro time difference

     

    1 user thanked author for this post.
    #190525

    Thank you, it ads a bit of amelioration but still not that great performance that you have.
    This is my code, no spread for this backtest, no MM, parametres are the same.
    I’m ready to change the time zone for PRT if it will improve my results.

    If you can’t see any solution. I will not be a problem. I am already grateful for your contribution.
    Thanks

     

Viewing 15 posts - 16 through 30 (of 96 total)

Create your free account now and post your request to benefit from the help of the community
Register or Login