5 Min EUR/USD Strategy

Forums › ProRealTime English forum › ProOrder support › 5 Min EUR/USD Strategy

Viewing 15 posts - 166 through 180 (of 251 total)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
  • 02/21/2018 at 8:09 PM #63416
    Gertrade

    @ manel,

    I thank you for your interest in this line of the forum. I just read you and allow me to answer you. It is true that on the number of edited versions, there is little that passes ProOrder selection. This shows us that Backtests are insufficient to validate a strategy, even with a lot of history. The 5mn versions have not been profitable and they are no longer listed. There are still some 28s versions that are still profitable, the oldest is the V2 that runs on ProOrder since 05/02 (Graph and results below). If you want to test, I invite you to look at the latest versions (versions 6 and 6. +), they are less subject to drawdowns unlike the V2 and others.

    I take this post to deliver a version 6.7 whose code has been completely reorganized for a better reading. Its particularity is to integrate a double strategy to overcome the problems of drawdowns (in clear when one is no longer in the trend) that occur too often in the life of a strategy. This double strategy must make this version more durable. It is our situation with respect to the equity curve (above or below) that directs the execution of trades towards one or the other strategy.
    This version is even more profitable than the previous ones and the variable MaxOrders, which for me is the barometer of my strategies, has been optimized to a maximum of 76 lots, which is for me a record.

    1 user thanked author for this post.
    avatar manel
    02/21/2018 at 11:33 PM #63423
    GraHal

    I look forward to Manel’s findings and views etc

    I am sorry Gertrude, but I just can’t get into the 28 sec TF knowing that the bars I am looking at / testing are not the same as we all would be testing on … for the reasons I discussed in the post below …

    GraHal wrote:

    @gertrade … reason why I think 28 sec TF is flawed

    I would however love to be proved wrong and see consistent results going forward with Live trades.

    I will pop in, test and report back from time to time. I still can’t get my head around the big difference in results from CFDs to spreadbet noted here …

    GraHal wrote:

    PS The good results are on CFDs!!! So there is a big difference between CFD and spreadbet … this will be the same for all Algos, but probably more pronounced difference the lower the Timeframe

    Is it that the CFD results are well well curve fitted??  I meant to check a few other Algos on both CFD and spread bet, but sorry I haven’t yet.

    Anybody got any thoughts? Doesn’t have to be profound, but it keeps the convo flowing?? 🙂
    GraHal

    1 user thanked author for this post.
    avatar Gertrade
    02/22/2018 at 12:04 AM #63424
    GraHal

    Attached are results I get on the latest 28 Sec v6.7 … 1st image is CFD and 2nd image is on spreadbet.

    The 3rd image is the 28 sec v6.7 but run (with no changes!) on a 5 Min TF … Mmmm what is this telling us?

    Looks like your talents are wasted on the 28 Sec Gertrade (I’m sure I’m going to type Gertrude one day! 🙂 )  … get onto the 5 Min TF??????????

    All 3 attached are over 10,000 bars with the 28 sec that goes back to last Friday afternoon 16 Feb.

    The 5 min over 10K bars goes back to Mon 28 Jan 18.

    GraHal

    1 user thanked author for this post.
    avatar Gertrade
    02/22/2018 at 12:05 AM #63428
    Gertrade

    @ Grahal,

    I’m sorry but I have no solutions to offer you.
    I really appreciated your collaboration.
    The spread debate has been very rewarding for me (all my backtests are now with a spread of 2). I now have a nice visual indicator.
    The TF28s is very advantageous for me, which is why I continue in this direction.
    I would like to share this advantage with everyone. Especially since my latest version is very successful, it represents a lot of personal investment in number of hours since V1.0
    I tried versions TF30s TF5mn but I knew failure on failure with ProOrder.
    I do not understand everything about this debate on TF28s. I understood that I was outside the standards (1mn / 5mn / 15mn etc …) and that it could pose problem on other platforms to other users, but I do not see why?

    Greetings

    Gertrade

    02/22/2018 at 12:29 AM #63432
    GraHal
    Gertrade wrote:

    This version is even more profitable than the previous ones

    The image you show is for  v2 (see top right corner of your image) but the Algo version you discuss and provide is v6.7 ??

    I’m off to bed … confused!!! 🙂
    GraHal

    02/22/2018 at 12:34 AM #63434
    Gertrade

    @ Grahal,

    Below are the statistics and the graphics of the evening with V6.7 on ProOrder.

    02/22/2018 at 12:58 AM #63439
    Gertrade

    below the backtest of the V6.7 on 17,000 units

    02/22/2018 at 7:06 AM #63442
    Gertrade
    @ Grahal,
    Gertrade wrote:

    There are still some 28s versions that are still profitable, the oldest is the V2 that runs on ProOrder since 05/02 (Graph and results below)

    In the first part of the #post-63416 I talk about the V2 (Results and Graph), it is in the second part that I present the V6.7 (delivery of the stratum).

    So Grahal !!! we were already asleep !!!

    I put you the results and graphics of the V6.7 that you expected above.

    Let yourself be tempted to test ProOrder V6.7, she still made some positive trades this morning.

     

    02/22/2018 at 9:04 AM #63451
    Gertrade

    @ Grahal,

    In the advanced search table, I use the EURUSD from the first line.

    Below advanced search table

    02/22/2018 at 9:12 AM #63453
    Vonasi

    To confirm one way or the other the effect of a 28 second time frame you could just select the start date for a backtest and then change the start time and run multiple tests. This way you can test multiple times at one minute starting intervals. The 28 second bars will be different each time. If there are more positive tests than negative then the 28 second thing might be worth continuing with. With a time frame that cannot be evenly divided into 1 minute this is the only way to truly test a strategy.

    2 users thanked author for this post.
    avatar GraHal, avatar Gertrade
    02/22/2018 at 11:14 AM #63484
    GraHal

    Ha I tried it on CFDs made about 10 or 12 changes to start times and even closed down the eurusd window and opened again etc (but didn’t close the platform) … I confirm that the approximate profile of the equity curve is maintained!!!!

    On my last test I picked an individual trade that gave > $200 rise and made a note of the time. Then I changed the start time and the equity profile looked very different, but on further checking the profitable trade is still there. The equity profile looks different because it is missing a previous profitable trade.

    So my conclusion is that Gertrades conditions are such that they occur many times over several minutes and so there is more than chance to get in before the rise or fall etc??

    GraHal

    PS Mmm I just checked actual time of the profitable trade and it is exact same time even with a different starting time overall (I did close the eurusd window in between tests).  Mmm not sure now re my conclusion above … too much of a coincidence??

    Makes me think that PRT is storing the backtest and only pulling that part of it which is needed to satisfy test period??

    I think the Platform would need to be closed down between each test / change of start time? Somebody else have a go please?? (I’m losing / not making money as I’m not trading doing this … then again maybe that is a good thing and so I am making money by NOT trading! 🙂 🙂 )

    02/22/2018 at 1:11 PM #63515
    Gertrade

    I tried Vonasi’s method of changing the start time of the backtest.
    After many changes (a good twenty), I confirm that no change occurs on the launch time of the trades and that the equity curve is always positive.
    In conclusion, the TF28s has no influence on any malfunction of launching trades. (Verified on Backtest with spread of 2)
    I would like to thank everyone involved in this line of the forum in addition to their daily occupation.

    02/22/2018 at 2:58 PM #63524
    GraHal

    But it could be for the reason I suggested … PRT is storing the backtest and only pulling that part of it which is needed to satisfy the requested test period??

    Actual bars being viewed / triggered on could be different by as much as 14 seconds … which is significant on a 28 Second Timeframe??

    Come on Vonasi stop sipping that wine and soaking up the Greek sun on your boat … what you reckon?? 🙂 🙂

    I’ll do a few test later and I will close the platform between each test.

    1 user thanked author for this post.
    avatar Gertrade
    02/22/2018 at 3:27 PM #63534
    Gertrade

    The early afternoon was hectic on the EURUSD. Indeed, between 13:30 and 13:55, the V6.7 went in drawdown, and since then has caught up and even made a profit.

    Below are the results and graph of the V6.7 on ProOrder.

    02/22/2018 at 3:43 PM #63536
    GraHal

    Yep attached is a test done after closing the platform.

    You have to make sure you closed the eurusd window before exit / Quit and save, but make sure you don’t save the results of your last test else the platform will just open up again with the same results / data!

    As you can see I have zoomed in on exactly the same time slot Tue 20 Feb at 13:28:44 and the accompanying trades are significantly different. I’ve done that fairly quick in between / while manual trading on the  Dow so if I have made a mistake then please let me know.

    GraHal

    PS 2nd attachment shows previous test trades at that particular time slot … as a comparison.

  • Author
    Posts
Viewing 15 posts - 166 through 180 (of 251 total)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Find exclusive trading pro-tools on ProRealCode Marketplace

Similar topics:
end of day” forex trading strategy for usd/jpy
ndq mean reversion 5 min strategy
5m ustech 100 strategy
bollinger+fractal strategy idea
changes to weekly reversal fx strategy
classic macd 200ema strategy
coding of a trend following strategy with pullback counter
combining short/long strategies into 1 strategy issue
conversion of pinescript strategy (tradingview)
conversion of strategy tester from pinescript t.view

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