Breakeven- en trailingstop on different securities & indexes & forex

Forums ProRealTime English forum ProOrder support Breakeven- en trailingstop on different securities & indexes & forex

  • This topic has 79 replies, 15 voices, and was last updated 3 years ago by avatarPaul.
Viewing 15 posts - 31 through 45 (of 80 total)
  • #126439

    I understand your interest in other markets. You will have to start excluding parameters to the point where it starts taking trades. But it breaks the coherence of the parameters. Not sure it’s worth the effort though.

    #126454

    >> meant to quote GraHal here but something happened…

    1st is for set up of what to use, the 2nd is the bread and butter if you dont make the home run, the third is the execution of it all if I can understand it correctly. I will try to apply this when standard deviation of range goes outside of a regular day. Usually there is a small profit to be collected there on both sides, but it is usually a quick one and needs a really good trailing to make it before a reversal. Maybe this can do it. 🙂

    #129380

    I am clearing the decks of loads of Systems … off PRT and onto my PC.

    I came across attached … good results so I thought I’d add to this Topic!  I did not optimise at all, just backtested today!

    Paul you put so much hard work … I hope this Topic might be revived.

    I will put on Forward Test on Monday and report back … and if I forget to report back, please remind me at some stage anybody!? 🙂

    3 users thanked author for this post.
    #131367

    @grahal im just here to remind you hehe 😀

     

    #131385

    Ha and I needed reminding  too … I had put it in my ‘4Test folder’ and then forgot to put it on test!

    I’ve been doing very well this week on manual trades and it’s taken my eye off the Auto-Systems!

    I’ve just set it going on Forward Test.

    Thanks

    #131573

    Hi @Paul, I’ve been playing with the ATR TS code you used on the new Vectorial DAX and the results are fantastic – just wanted to say a big thanks for that!

    Looking at this section of it,

    I tend to leave tsatrperiod  and tsminstop as they are, as those values seem to work on any instrument and TF. But changes to ts atr distance, tsminatrdist, and tsincrements all have a huge effect. I’m still a bit unclear as to how these 3 work together. A lower ts atr distance, 3 to 6 for example, def gives a higher win % and seems better for scalping.

    Do you have any basic rule of thumb as to how you select values for those on different instruments or TFs?

    At the moment I’m working on a 2m TF with atr distance =5, tsminatrdist=2 and increments=0, seems to get good results overall.

    5 users thanked author for this post.
    #131601

    what do you mean by ATR = 5
    I put the screen copy for more understanding

    #131614

    i mean trailingstoplong / short = 5

    ts atr distance is rem’d out

    1 user thanked author for this post.
    #131671

    Hi Nonetheless. Thanks for your comment!

    I do the same with the tsatrperiod, leaving it at 14 regardless instrument.

    the tsminstop is different. For this it’s best to have the correct value from IG. 

    It means if the current price is too close to the ts exit stop price, it will reject the exit when using the wrong value. (not visible in a backtest)

    Because of this I added an extra exit criteria, if index crosses over/under the tsnewsl it still exits next bar at open on market. My purpose was for fast timeframes, because if the stop is rejected, the next bar will be fast. On a hourly timeframe it doesn’t bring much.

    Then there’s tssensitivity. If you have often a spike and a close considerably away from the new high/low, the system originally only use the close. With some algo’s I thought might be interesting to calculate from the real high/low instead of close.

    The trailingstop long/short is used to set the atr distance when the gains are sufficient. I do not have a basic rule on this, but when the trailing stop starts working, it has to have gains which makes sense compared to a loss on stoploss or average losses.

    In the old situation, with tsincrements at 0 and tssensitivity at 0, it means a fixed (besides the round difference) number of points for the trailing stop. Those points are calculated from the highest/lowest close.

    In the other situation, it starts working when the atr distance trailingstoplong/short is reached.

    From that moment on it starts reducing the atr distance to the tsminatrdist with he increments set. Values which make sense normally are 0-0.25 with increments of 0.05. At the vectorial dax I’ve put them even higher.

    In other words, if the trailing stop value in points was originally +/- 80, it starts reducing it to +/- 40 points quickly once it kicks in (using values in vectorial dax)

     

    To have a better visual, try to use graph tgl/tgs and graphonprice tsnewsl and see the difference using increments at zero and higher.

    The trailingstop doesn’t do completely what I had in mind though, but for me it’s good.

    4 users thanked author for this post.
    #131681

    Thanks very much for the explanation, that makes a lot of things clearer. ATR TS is so much more sophisticated than the standard version where all you really have is the start and the step. Yours does a much better job of catching a bigger piece of the MFE for trades that don’t ever get to the TP.

    It’s a brilliant contribution, consider me a convert. 👍

    2 users thanked author for this post.
    #131809

    I have a “trailing challenge” for anyone up to it. I would guess that @Paul and @nonetheless might like this.

    Attached is the Kama Sma strategy originally posted by robertogozzi. It includes his sophisticated trail. However the gain/loss ratio is about the same (2.6-2.7) with and without the trail. The win-rate is better with trail. But the interesting part is the MFE/MAE ratio at 3.7 (!) meaning there is way more candy to steal. I have tried a lot of trails and breakevens from this forum but can’t find a good one boosting the G/L above 3. Any suggestions on a suitable trail? A side note is that the equity curve will never be straight so don’t get hung up on that.

     

    2 users thanked author for this post.
    #131816

    And of course thank you @robertogozzi for converting the strategy to PRT. It’s still awesome!

    1 user thanked author for this post.
    #131827

    I doubt that Roberto’s version can be bettered … if it could he would have done it!

    #131832

    There’s never an end to new ideas to improve things.

    I made it, but then I never updated it.

    #131875

    here’s a version with atr ts/be

     

Viewing 15 posts - 31 through 45 (of 80 total)

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