Fractal breakout intraday Strategy EUR/USD 1H –

Forums › ProRealTime English forum › ProOrder support › Fractal breakout intraday Strategy EUR/USD 1H –

Viewing 15 posts - 91 through 105 (of 360 total)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 … 20 21 22 23 24
  • 04/21/2017 at 7:09 PM #32911
    Elsborgtrading

    ALE, for the first post and the pic of the WF results. Which sample did you choose for variables, if you were to run it live? I am just trying to figure out how to choose the best variables. The WF video guide is still only in French, which I don’t understand, and information about this is very little. more or less it’s down to WFE>50%

    Cheers Kasper

     

    04/21/2017 at 7:14 PM #32912
    Nicolas

    @Elsborgtrading

    I answered to this question in this post : https://www.prorealcode.com/topic/fractal-breakout-intraday-strategy-eurusd-1h/page/6/#post-32855

    The last one because it refers to the recent market behavior, assuming this period is close of the current day of course.

    Walk Forward optimisation is the same as a classic one except that it divides the history into 2 periods, the In Sample one (which is the tested period where variables are optimized) and the Out Of Sample one (the period where these optimized settings are tested). This test can be repeated X times you want over the whole history. Please tell me if it is now clear 🙂

    1 user thanked author for this post.
    avatar Elsborgtrading
    04/21/2017 at 7:46 PM #32914
    Elsborgtrading

    Hi Nicolas, yes I read the post, but obviously it’s not an exact science- So what I got out of your posts is that:

    • It should be the most recent WF sample if WFE>50%
    • WFE >50% but if larger that 100% we cannot expect this to be normal behaviour.
    • if sample 1-5 is very different we need optimize again after x month
    • if sample 1-5 is more or less a like, we have a very solid strategy

    is this the rule of thumb?

    Your WF video on youtube only has 295 hits- it would be a lot more if it was in English.. *wink *wink 😀

    Have a nice weekend

    Kasper

    3 users thanked author for this post.
    avatar PrinceMyshkin, avatar GraHal
    04/21/2017 at 8:16 PM #32916
    Henrik

    Hi!

    If you test your strategy with different setups for your WF :

    ancored/not ancored

    repetitions

    in sample%

    out of sample%

    and then take the best WF and do a regular opti on the same amount of in sample data as your most robust WF .

    Then you atleast do the same as your test.

    just an idea.

     

    Regards

    Henrik

     

     

     

    1 user thanked author for this post.
    avatar GraHal
    04/21/2017 at 10:40 PM #32920
    GraHal

    Just an idea Nicolas … why don’t you get a French to text converter and then a french text to english translator and then  english text to audio … hey presto english WF video and any other language you wish! 🙂

    Or even French to English audio / any other language?

    I know ‘it can get lost in the translation’, but you know brain storming and all that, it might trigger a less crazy idea??

    GraHal

    1 user thanked author for this post.
    avatar Elsborgtrading
    04/21/2017 at 11:02 PM #32922
    ALE

    @all can we register in mother language walk farward video for the whole community to help Nicolas  ?

    Nicolas can you give us in English test  your explanation of WF, to translate and register it to add to your video?

    2 users thanked author for this post.
    avatar Elsborgtrading, avatar GraHal
    04/22/2017 at 8:59 AM #32935
    Nicolas

    Ok guys, I’ll do an English video this week. I promised to do one!

    A recent discussion with the PRT development team (last week or so) and they told me that a more complete explanation of the walk forward module is coming in a documentation update soon.

    04/22/2017 at 10:34 AM #32937
    GraHal

    General:

    Just a few thoughts I’ve woke up with (sorry guys :)) but first off I think we should move Walk Forward discussion to it’s own Thread?

    The good things that are already on this Thread re WF … Nicolas gems, good summary by Kasper & Henrik (bullet points are easy to remember) need moving to the Thread below else other members will miss them / not read them when searching for Walk Forward?

    Maybe you could move all questions / answers / comments Ale as a Mod to the Thread below?

    https://www.prorealcode.com/topic/how-to-use-prt-walk-foward-test/

    Maybe we can save Nicolas his valuable time? How to use Walk Forward is easy and intuitive and anyway it’s all in the Link below no need for a video on how to use??

    https://www.prorealcode.com/blog/learning/strategy-optimisation-walk-analysis/

    What we need is for us to ask our questions on the 1st Link above and for others to answer … that way we all increase our understanding? It’s the interpretation of the results that is the difficult bit?

    Specific:

    For me it’s this notion that we use the last set of optimised variables (provided OOS WF eff > 50%) to move forward to Live Trading?  Surely this is only recommended if very similar to previous sets?

    If the last set of optimised variables are not very similar then we are moving forward and using the bias of the last IS period … example IS may be UP period (Live may be Down period) or an exception  in the last IS period (NF Payroll etc).

    Then the next issue is … if we have to keep going back over until, say at least 3 x IS optimised variables are similar (and give OOS WF eff >50%) then it’s the time it takes …. PRT Server is soooooo slow!??

    Thanks for Listening
    GraHal

     

     

    04/24/2017 at 11:19 PM #33256
    ALE

    HELLO,

    Following is a rough version of the strategy. It seems to work better in different markets. I found that in order to find the CP value appropriate to the market, the search can start from a crude code, I removed the trailing stop and the Donchian stop, then I put a simple stop and a take profit, and I have startan automatic search of the values: CP,SL,TP.  The dax version that follows it is rough because it seems to work better than with the trailing stop and Donchian stop. In the coming days, I’ll used the rough version to look for other markets and other time frames. Thank you all for your attention

    A greeting, Ale

    6 users thanked author for this post.
    avatar dajvop, avatar Nicolas, avatar O-jay8, avatar reb, avatar victormork
    04/24/2017 at 11:40 PM #33262
    ALE

    …FBS_DAX_1H_V2

    6 users thanked author for this post.
    avatar dajvop, avatar Nicolas, avatar O-jay8, avatar reb, avatar victormork
    04/25/2017 at 7:35 AM #33276
    deleted_account

    Looks really promising! Thank you very much!

    1 user thanked author for this post.
    avatar ALE
    04/25/2017 at 5:08 PM #33359
    ALE

    HELLO,

    here below results of Rough strategy Bund 1h

    Regards

    1 user thanked author for this post.
    avatar victormork
    04/26/2017 at 11:04 AM #33464
    Nicolas

    FYI, I compiled a lot of useful informations about Walk Forward tool in this new blog article: https://www.prorealcode.com/blog/prorealtime-walk-analysis-tool/

    How it helps everyone..

    9 users thanked author for this post.
    avatar PrinceMyshkin, avatar torkelab, avatar Elsborgtrading, avatar manel, avatar Despair, avatar GraHal, avatar victormork, avatar ALE
    04/26/2017 at 10:57 PM #33599
    ALE

    Hello

    Below the EUR/USD  versions for the TF 30m, having lowered the TF I decided to add filters  to raise efficiency. If you want to try changing the filter values, it will be fun.
    In the first version there are two filters, while in the second I have eliminated the 1st filter and I kept the 2nd
    1° filt: It is composed of two moving averages with a period of 20 and 50 that control each other
    2°filt: The candle breaking the level must be 30 pips above or below the previous closure.
    Thanks

    3 users thanked author for this post.
    avatar Despair, avatar Nicolas, avatar GraHal
    04/27/2017 at 10:17 AM #33651
    GraHal

    I’ve just set FBS_EURUSD_30M_V1.itf Live with below variable values.

    Feels like a ‘leap of faith’ using the final value out of 5 as per Nicolas guide on WF … see attached, difference in result from WF to non-WF / regular BT (with CP = 4).

    As you say Ale … it will be fun … only set at £1 per point, so no big deal.  I’ll let you know how it goes.

    GraHal

     

  • Author
    Posts
Viewing 15 posts - 91 through 105 (of 360 total)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 … 20 21 22 23 24

Find exclusive trading pro-tools on ProRealCode Marketplace

Similar topics:
bollinger+fractal strategy idea
entre 20h30 et 21h00, acheter 1 contrat au marché
eur/usd backtest spread

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