Trailing stop fails to move

Viewing 15 posts - 16 through 30 (of 30 total)
1 2
  • Author
    Posts
  • 10/28/2020 at 2:46 PM #148790 quote
    nonetheless
    Participant
    Master

    Actually I’ve got another problem with that TS (as posted above). It moves to breakeven then stops, doesn’t trail with the price movement. Obviously it’s working correctly in backtest, but in live trading it gets stuck. Have you seen this as well?

    10/28/2020 at 3:25 PM #148793 quote
    Paul
    Participant
    Master

    no not yet, it never stops 😉 Your dj 5m mod v5 stopped suddenly with a strange message but I think unrelated to this.

    On the 10s frame I use 2 trailingstops in demo and they seem to work alright, but because of using 2 I don’t know which was triggered. I will check.

    Can you provide a pic?

    Screenshot-2020-10-28-at-15.24.33.jpg Screenshot-2020-10-28-at-15.24.33.jpg
    10/28/2020 at 3:30 PM #148795 quote
    Paul
    Participant
    Master

    those exits were because the atr trialingstop. I will run it without.

    10/28/2020 at 3:41 PM #148796 quote
    nonetheless
    Participant
    Master

    I haven’t got a pic, when I realised it was stuck I closed the position manually. That was with NAS 5m Mod v4S

    10/28/2020 at 3:57 PM #148798 quote
    Paul
    Participant
    Master

    Interested in testing your mod with the atr trailingstop instead to see if it behaves the same live?

    I updated that one, and the breakeven too, both to use cumulative positions, it’s included in de maex strategy

    10/28/2020 at 6:22 PM #148806 quote
    nonetheless
    Participant
    Master

    yeah, I’ll def check it out when I get a minute, it’s either that or go back to the old %TS I had been using. It’s a bit basic but does the job. The big advantage to yours is the separate values for long/short, so maybe I can build that in to the other one (of Nicolas)

    10/28/2020 at 6:37 PM #148809 quote
    Paul
    Participant
    Master

    Thought of another way if you take that trailingstop (it’s all based on the original)

    Duplicate the code and make sure both not overlap each other.

    1st one for long and remove all short code, and the second one for short with all long removed.

    Should be the same effect (besides the accelerator and sensitivity) but still the advantage for separate values.

    josef1604 thanked this post
    10/28/2020 at 6:52 PM #148811 quote
    nonetheless
    Participant
    Master

    My thoughts exactly. Does this look right to you? Seems to test ok

    //%trailing stop function
trailingpercentlong  = tst // %
trailingpercentshort = tss // %
stepPercentlong = st
stepPercentshort = st2
if onmarket then
trailingstartlong = tradeprice(1)*(trailingpercentlong/100) //trailing will start @trailingstart points profit
trailingstartshort = tradeprice(1)*(trailingpercentshort/100) //trailing will start @trailingstart points profit
trailingsteplong = tradeprice(1)*(stepPercentlong/100) //% step to move the stoploss
trailingstepshort = tradeprice(1)*(stepPercentshort/100) //% step to move the stoploss
endif

//reset the stoploss value
IF NOT ONMARKET THEN
newSL=0
ENDIF
 
//manage long positions
IF LONGONMARKET THEN
//first move (breakeven)
IF newSL=0 AND close-tradeprice(1)>=trailingstartlong THEN
newSL = tradeprice(1)+trailingsteplong
ENDIF
//next moves
IF newSL>0 AND close-newSL>trailingsteplong THEN
newSL = newSL+trailingsteplong
ENDIF
ENDIF
 
//manage short positions
IF SHORTONMARKET THEN
//first move (breakeven)
IF newSL=0 AND tradeprice(1)-close>=trailingstartshort THEN
newSL = tradeprice(1)-trailingstepshort
ENDIF
//next moves
IF newSL>0 AND newSL-close>trailingstepshort THEN
newSL = newSL-trailingstepshort
ENDIF
ENDIF
 
//stop order to exit the positions
IF newSL>0 THEN
SELL AT newSL STOP
EXITSHORT AT newSL STOP
ENDIF

    Not sure about tradeprice(1), whereas you have positionprice. I know the difference, it’s more the (1) I’m unsure of

    josef1604 thanked this post
    10/28/2020 at 7:03 PM #148814 quote
    Paul
    Participant
    Master

    looks good. Tradeprice as written is correct.  Only the reset is off when reversing a position, but most of the time the difference is smal to nothing.

    nonetheless and josef1604 thanked this post
    10/28/2020 at 7:26 PM #148821 quote
    nonetheless
    Participant
    Master

    Thanks Paul. I added

    sensitivity = (low+high+close)/3

    instead of close and the performance is back to where it was using your TS – at least in backtest it is. Let’s see how it runs live.

    josef1604 thanked this post
    11/11/2020 at 6:31 PM #150172 quote
    nonetheless
    Participant
    Master

    Hi @Paul, I’ve got more problems with this cumulative orders TS. This is the code I’m using:

    // %trailing stop function incl. cumulative positions

once trailingstoptype = 1
 
if trailingstoptype then
//====================
trailingpercentlong  = 0.26 // %
trailingpercentshort = 0.26 // %
once acceleratorlong = 0.003 // [1] default; always > 0 (i.e. 0.5-3)
once acceleratorshort= 0.023 // 1 = default; always > 0 (i.e. 0.5-3)
ts2sensitivity  = 4 // [1] default [2] hl [3] lh [4] typicalprice (not use once)
//====================
once steppercentlong  = (trailingpercentlong/10)*acceleratorlong
once steppercentshort = (trailingpercentshort/10)*acceleratorshort
if onmarket then
trailingstartlong = positionprice*(trailingpercentlong/100)
trailingstartshort = positionprice*(trailingpercentshort/100)
 
trailingsteplong = positionprice*(steppercentlong/100)
trailingstepshort = positionprice*(steppercentshort/100)
endif
 
if not onmarket or ((longonmarket and shortonmarket[1]) or (longonmarket[1] and shortonmarket)) then
newsl           = 0
mypositionprice = 0
endif
positioncount = abs(countofposition)
if newsl > 0 then
if positioncount > positioncount[1] then
if longonmarket then
newsl = max(newsl,positionprice * newsl / mypositionprice)
else
newsl = min(newsl,positionprice * newsl / mypositionprice)
endif
endif
endif
if ts2sensitivity=1 then
ts2sensitivitylong=close
ts2sensitivityshort=close
elsif ts2sensitivity=2 then
ts2sensitivitylong=high
ts2sensitivityshort=low
elsif ts2sensitivity=3 then
ts2sensitivitylong=low
ts2sensitivityshort=high
elsif ts2sensitivity=4 then
ts2sensitivitylong=(typicalprice)
ts2sensitivityshort=(typicalprice)
endif
if longonmarket then
if newsl=0 and ts2sensitivitylong-positionprice>=trailingstartlong*pipsize then
newsl = positionprice+trailingsteplong*pipsize
endif
if newsl>0 and ts2sensitivitylong-newsl>=trailingsteplong*pipsize then
newsl = newsl+trailingsteplong*pipsize
endif
endif
if shortonmarket then
if newsl=0 and positionprice-ts2sensitivityshort>=trailingstartshort*pipsize then
newsl = positionprice-trailingstepshort*pipsize
endif
if newsl>0 and newsl-ts2sensitivityshort>=trailingstepshort*pipsize then
newsl = newsl-trailingstepshort*pipsize
endif
endif
if barindex-tradeindex>1 then
if longonmarket then
if newsl>0 then
sell at newsl stop
endif
if newsl>0 then
if low crosses under newsl then
sell at market
endif
endif
endif
if shortonmarket then
if newsl>0 then
exitshort at newsl stop
endif
if newsl>0 then
if high crosses over newsl then
exitshort at market
endif
endif
endif
endif
mypositionprice = positionprice
endif

    It goes to breakeven exactly as expected but then stops, doesn’t trail. In the attached image, the breakeven was at the circled candle. 6 candles later the stop is still at 11835.2. the step is .003% so it should move .35 per candle = 11837. Any thoughts on this?

    ts.jpg ts.jpg
    11/11/2020 at 6:57 PM #150176 quote
    nonetheless
    Participant
    Master

    Ok, never mind – I just realized that acceleratorlong = .003 does not mean .003%

    I forgot because the other TS I use has a straight % value for trailingstep

    This one does trail but sooooo slowly – effectively just holds the position open with the hope of getting to the target. With a setting of .003 I might as well just use the breakeven code.

    11/11/2020 at 7:29 PM #150180 quote
    Paul
    Participant
    Master

    Hi Nonetheless good you found the reason.

    I might still mention a few things.

    Probably it misses *pointsize at the end, this is the part I changed

    trailingstartlong = (tradeprice(countpos)*(trailingpercentlong/100))*pointsize
trailingstartshort = (tradeprice(countpos)*(trailingpercentshort/100))*pointsize
trailingsteplong = (tradeprice(countpos)*(steppercentlong/10000))*pointsize
trailingstepshort = (tradeprice(countpos)*(steppercentshort/10000))*pointsize

    I also changed /100 to /10000 so steps are better manageable.  Doesn’t change results, but the *pointsize might give different results on i.e 1 euro or 2 dollar DJ charts. (the countpos needs to be changed).

    In the past I noticed that using if newsl>0 without defining longonmarket or shortonmarket sometimes, but not always, gave  a few strange results, that why I at the end of the code had it specified.

    Anyway for this part you could test 2 variations. Should have the same results.

    if (shortonmarket and newsl > 0) or (longonmarket and newsl>0) then
if positioncount > positioncount[1] then
if longonmarket then
newsl = max(newsl,positionprice * newsl / mypositionprice)
endif
if shortonmarket then
newsl = min(newsl,positionprice * newsl / mypositionprice)
endif
endif
endif

    or

    positioncount = abs(countofposition)

if longonmarket and newsl > 0 then
if positioncount > positioncount[1] then
newsl = max(newsl,positionprice * newsl / mypositionprice)
endif
endif

if shortonmarket and newsl > 0 then
if positioncount > positioncount[1] then
newsl = min(newsl,positionprice * newsl / mypositionprice)
endif
endif

    Maybe you can keep above as backup if you still find strange things trading live.

    nonetheless thanked this post
    11/11/2020 at 8:32 PM #150189 quote
    nonetheless
    Participant
    Master

    That’s great, I’ll try both those options – thanks again!

    09/17/2021 at 6:04 PM #177823 quote
    murre87
    Participant
    Senior

    Do we have a final version of this.

    Add to the snippet library?

  • Author
    Posts
Viewing 15 posts - 16 through 30 (of 30 total)
1 2
  • You must be logged in to reply to this topic.

Trailing stop fails to move


ProOrder: Automated Strategies & Backtesting

New Reply
Author
author-avatar
nonetheless @nonetheless Participant
Summary

This topic contains 29 replies,
has 4 voices, and was last updated by murre87
4 years, 5 months ago.

Topic Details
Forum: ProOrder: Automated Strategies & Backtesting
Language: English
Started: 10/26/2020
Status: Active
Attachments: 3 files
Logo Logo
Loading...
Do you like cookies? 🍪 We use cookies to ensure you get the best experience on our website. (Learn more)
©2026 ProRealCode - Trading & Coding with ProRealTime™