NAS 2m HULL-SAR trading system

Viewing 15 posts - 106 through 120 (of 344 total)
  • Author
    Posts
  • #155697 quote
    nonetheless
    Participant
    Master

    In the 2 weeks that I’ve been running it, the first trades were 12700, 12687, 12691 within 16 minutes between the first and last (Dec 21).

    The second entries were 12681, 12706, 12715 about 40 minutes apart (dec 24).

    Then one position opened on the 29th at 12914, still open.

    Is that what you’ve got?

    Judging from previous versions, this seems typical. If the entries were more widely spaced then either you’d be buying at a much worse price, or it’s fallen so far that the broader conditions are no longer met – in which case it may well be a losing trade and you’ll be pleased that there’s only one entry.

    Munshun thanked this post
    #155711 quote
    Munshun
    Participant
    Junior

    As I write I just started it a few days ago.

    It took 3 trades (all Long) on Dec 28 at 12828 (17:14), 12835 (17:20) and 12836 (17:28), those were all closed (with a profit) on Dec 29 at 07:55.

    Then it took 3 trades (all Long) on Dec 29 at 12914 (14:34), 12917,8 (14:42) and 12917,0 (14:52). Those were sold with a loss today at 15:38. Strange that it only took 1 trade for you, but 3 trades for me.

    At 15:38 it just immediately opened a Short after it closed the positions above.

    So you are correct. The trades are not taken immediately after eachother, they are spaced 6-10 minutes apart, or something like that. Does the behaviour seem accurate you think?

    #155714 quote
    nonetheless
    Participant
    Master

    I’m talking about

    1. NAS-2m-HULL-SAR-v5.3L.itf

    which is long only. Which algo are you running?

    Also, did you adjust the Ctime settings for Sweden?

    Munshun and Kovit thanked this post
    #155727 quote
    Munshun
    Participant
    Junior

    I am running HULL-SAR v3b.3, which seems to have shorts, right? I have not changed any time-setting as I am running ProRealTime with GMT time.

    Would you advice I change to the version you link above?

    #155730 quote
    nonetheless
    Participant
    Master

    v3b.3 was developed with just a 1 year backtest, so I wouldn’t recommend.

    v5.3L is built on 6 years of data 75/25 so should be more reliable (in theory) and making it Long only improved performance dramatically. If you’re on GMT then it should be good to go with no changes.

    Munshun thanked this post
    #155731 quote
    Munshun
    Participant
    Junior

    Thanks for this! Will change!

    #158465 quote
    nonetheless
    Participant
    Master

    Hi @Paul, I’ve been working on an ATR TS version of this in combination with your breakeven code but I’m finding that it doesn’t want to move the stop. This is the code I’m using:

    // break even stop incl. cumulative positions
    once enablebe = 1
    if enablebe then
    //====================
    once besg = 0.3 //% break even stop gain
    once besl = 0.04 //% break even stop level (+ or -)
    besensitivity = 2 // [1] default [2] hl [3] lh [4]typicalprice(not use once)
    //====================
    if not onmarket or ((longonmarket and shortonmarket[1]) or (longonmarket[1] and shortonmarket)) then
    benewsl=0
    mypositionpricebe = 0
    endif
    positioncountbe = abs(countofposition)
    if benewsl > 0 then
    if positioncountbe > positioncountbe[1] then
    if longonmarket then
    benewsl = max(benewsl,positionprice * benewsl / mypositionpricebe)
    else
    benewsl = min(benewsl,positionprice * benewsl / mypositionpricebe)
    endif
    endif
    endif
    if besensitivity=1 then
    besensitivitylong=close
    besensitivityshort=close
    elsif besensitivity=2 then
    besensitivitylong=high
    besensitivityshort=low
    elsif besensitivity=3 then
    besensitivitylong=low
    besensitivityshort=high
    elsif besensitivity=4 then
    besensitivitylong=typicalprice
    besensitivityshort=typicalprice
    endif
    if longonmarket then
    if besensitivitylong-positionprice>=((positionprice/100)*besg)*pointsize then
    benewsl=positionprice+((positionprice/100)*besl)*pointsize
    endif
    endif
    if shortonmarket then
    if positionprice-besensitivityshort>=((positionprice/100)*besg)*pointsize then
    benewsl=positionprice-((positionprice/100)*besl)*pointsize
    endif
    endif
    if barindex-tradeindex>1 then
    if longonmarket then
    if benewsl>0 then
    sell at benewsl stop
    endif
    if benewsl>0 then
    if low crosses under benewsl then
    sell at market
    endif
    endif
    endif
    if shortonmarket then
    if benewsl>0 then
    exitshort at benewsl stop
    endif
    if benewsl>0 then
    if high crosses over benewsl then
    exitshort at market
    endif
    endif
    endif
    endif
    mypositionpricebe = positionprice
    endif
    

    but if you look at the attached image, the stop should have moved when the price hit 12878.3, no?

    #158466 quote
    nonetheless
    Participant
    Master

    attachment

    #158476 quote
    Monobrow
    Participant
    Senior

    Not anything to do with IG min Ts distance? They are quite high i think, like 10 points in spreadbet. Idk about CFD.

    #158548 quote
    nonetheless
    Participant
    Master

    No, the min distance on the NAS is only 4. This should go to b/e after about 40 points, but it went over 90 and still nothing.

    I always feel better once i know a position is all locked up and can’t lose money.

    #158717 quote
    winnie37
    Participant
    Veteran

    thanks a lot for this algo! Did you tried to adapt it on DJ?

    #158762 quote
    nonetheless
    Participant
    Master

    Still working on it, but doesn’t look quite as promising as the NAS.

    winnie37 thanked this post
    #158861 quote
    Paul
    Participant
    Master

    hi nonetheless  , I didn’t saw your previous post with the question regarding the breakeven. I haven’t looked into this, but it’s probably not relevant anymore now you ‘ve the code from Roberto. I’am curious how it would perform on this strategy.

    #158908 quote
    nonetheless
    Participant
    Master

    Thanks Paul, I’ve been playing around with all different TS for this, and yes now Roberto’s is another option – still tinkering with that one.

    Curious about your breakeven + ATR TS, it seems to behave properly (ie positions do close as if the stop had moved) but it’s not indicated in the open positions window. Also on the IG web platform, the SL display never changes but then the position will mysteriously close, as if the stop had moved.

    Have you ever seen this?

    #159131 quote
    tundercut
    Participant
    Senior

    Hi guys, I did a small backtest on the Dax time frame 2 minutes with excellent results even if only for a month, but the real time strategy I tested on Friday 22 January did not open any positions, while the backtest shows 2 open positions. Can anyone explain to me why and if there is an error to correct? Thanks.

Viewing 15 posts - 106 through 120 (of 344 total)
  • You must be logged in to reply to this topic.

NAS 2m HULL-SAR trading system


ProOrder: Automated Strategies & Backtesting

New Reply
Author
Summary

This topic contains 343 replies,
has 42 voices, and was last updated by bege
3 years, 1 month ago.

Topic Details
Forum: ProOrder: Automated Strategies & Backtesting
Language: English
Started: 10/09/2020
Status: Active
Attachments: 129 files
Logo Logo
Loading...