wasn’t this a test version?
Maybe that was why I put ?? in the title.
Wide spread overnight does not necessarily have to be bad thing provided the wide spread is not accompanied by illogical volatility which a System has no chance of predicting?
Spread at 6 overnight may be better compared to 40 to 80 points movement on a 1 min bar when spread is 2.4 and the professionals are tricking / faking us and our Systems into taking the wrong direction!? 🙁
Then 5 mins or so later … oh sorry you’re going the wrong way again now, us pro’s have reversed our positions … you must try and keep up with us if you want to make money!!?? 🙂
PaulParticipant
Master
after posting ofcouse I see the first line of the code… 🙂
// test trailingstop and breakeven on the dax
indeed high spread doesn’t need to be a bad thing. Looking at barhunter it acts only when the spread is the highest and still it preforms!
About tricking, something to use with barhunter? Maybe there are special times where the markets makes a move to do the opposite? Like before closing or opening, or of other markets? I don’t give it much chance but if it can be coded quick it would be fun to test on a higher timeframe.
PaulParticipant
Master
@grahal, there’s a quick try on 15min dax
It searches for the bar(=time) of the crosses from the moving average.
It can then go 1x long and 1x short within the next xx (=4) bars. Max trades is 2 a day. I didn’t look at sl/pt/ts and just took some averages and tested only 15m.
//-------------------------------------------------------------------------
// hoofd code : BarHunter v4.5p 15m dax trick
//-------------------------------------------------------------------------
// dax 1 hour timeframe backtest / dax 1 minute live
// spread 4
// test
defparam cumulateorders = false
defparam preloadbars = 1000
timeframe (default)
once positionsize = 1
once mode = 1 // use [1] for 1 hour timeframe, [0] for 1 minute timeframe
once tds = 0 // off when optimising [trend detection system]
// separate long/short or go both
once longtrading =1
once shorttrading =1
once holiday =1
once closefriday =1
once displaydim =1 // displays the number of days in market (activated graph)
once maxdim =5 // maximum days in market (first day = 0)
// select which intradaybar should be analysed (depends on timeframe settings)
once barnumberlong =13 //long (timezone dependent)
once barnumbershort=13 //short (timezone dependent)
sm=3
lm=15
// reset low timeframe
if intradaybarindex=0 then
tradecounter=0
tradecounterlong=0
tradecountershort=0
tradeday=1
endif
// holiday
if holiday then
if (month = 5 and day = 1) or (month = 12 and day >=24) then
tradeday=0
else
tradeday=1
endif
endif
// set high / low points to break
if longtrading or (longtrading and shorttrading) then
longma = average[lm](close)
shortma = average[sm](close)
endif
if shorttrading or (longtrading and shorttrading) then
longma = average[lm](close)
shortma = average[sm](close)
endif
// trend detection system
if tds=0 then
trendup=1
trenddown=1
else
if tds=1 then
trendup=(average[10](close)>average[10](close)[1])
trenddown=(average[10](close)<average[10](close)[1])
else
if tds=2 then
period= 2
inner = 2*weightedaverage[round( period/2)](typicalprice)-weightedaverage[period](typicalprice)
hull = weightedaverage[round(sqrt(period))](inner)
trendup = hull > hull[1]
trenddown = hull < hull[1]
else
if tds=3 then
period= 2
inner = 2*weightedaverage[round( period/2)](totalprice)-weightedaverage[period](totalprice)
hull = weightedaverage[round(sqrt(period))](inner)
trendup = hull > hull[1]
trenddown = hull < hull[1]
endif
endif
endif
endif
// conditions
condbuy=intradaybarindex >= barnumberlong
condbuy=condbuy and shortma crosses over longma
condbuy=condbuy and trendup
condsell=intradaybarindex >= barnumbershort
condsell=condsell and shortma crosses under longma
condsell=condsell and trenddown
timeframe (default)
// entry criteria
if mode then // mode[1] backtesting on 1 hour timeframe
if tradeday and tradecounter < 100 then
if (longtrading and not shorttrading) or (longtrading and shorttrading) then
if condbuy and tradecounterlong < 1 and (tradecounter<2 and (intradaybarindex < barnumberlong+4)) then
buy positionsize contract at market
tradecounter=tradecounter+1
endif
endif
if (shorttrading and not longtrading) or (longtrading and shorttrading) then
if condsell and tradecountershort < 1 and (tradecounter<2 and (intradaybarindex < barnumbershort+4)) then
sellshort positionsize contract at market
tradecounter=tradecounter+1
endif
endif
endif
else // mode[0] running demo / live on 1 minute timeframe
if tradeday and tradecounter < 1 then
if (longtrading and not shorttrading) or (longtrading and shorttrading) then
if condbuy and tradecounterlong < 1 then
buy positionsize contract at market
tradecounter=tradecounter+1
endif
endif
if (shorttrading and not longtrading) or (longtrading and shorttrading) then
if condsell and tradecountershort < 1 then
sellshort positionsize contract at market
tradecounter=tradecounter+1
endif
endif
endif
endif
//timeframe (1 hour, updateonclose)
// trailing atr stop on high timeframe
once trailingstoptype = 1 // trailing stop - 0 off, 1 on
once trailingstoplong = 5 // trailing stop atr relative distance
once trailingstopshort = 5 // trailing stop atr relative distance
once atrtrailingperiod = 14 // atr parameter value
once minstop = 10 // minimum trailing stop distance
//----------------------------------------------
atrtrail = averagetruerange[atrtrailingperiod]((close/10)*pipsize)/1000
trailingstartl = round(atrtrail*trailingstoplong)
trailingstarts = round(atrtrail*trailingstopshort)
if trailingstoptype = 1 then
tgl =trailingstartl
tgs=trailingstarts
if not onmarket or ((longonmarket and shortonmarket[1]) or (longonmarket[1] and shortonmarket)) then
maxprice = 0
minprice = close
newsl = 0
endif
if longonmarket then
maxprice = max(maxprice,close)
if maxprice-tradeprice(1)>=tgl*pointsize then
if maxprice-tradeprice(1)>=minstop then
newsl = maxprice-tgl*pointsize
else
newsl = maxprice - minstop*pointsize
endif
endif
endif
if shortonmarket then
minprice = min(minprice,close)
if tradeprice(1)-minprice>=tgs*pointsize then
if tradeprice(1)-minprice>=minstop then
newsl = minprice+tgs*pointsize
else
newsl = minprice + minstop*pointsize
endif
endif
endif
endif
timeframe (default)
// trailing atr stop exits on low timeframe
if longonmarket then
if newsl>0 then
sell at newsl stop
endif
endif
if shortonmarket then
if newsl>0 then
exitshort at newsl stop
endif
endif
// close to reduce risk in the weekend
if closefriday then
if onmarket then
if (dayofweek=5 and hour=22) then
sell at market
exitshort at market
endif
endif
endif
// stoploss & profit target
set target %profit 2
set stop %loss 2
// display days in market
if displaydim then
if (not onmarket or ((longonmarket and shortonmarket[1]) or (longonmarket[1] and shortonmarket))) then
dim=0
endif
if not ( dayofweek=1 and hour <= 1) then
if onmarket then
if openday <> openday[1] then
dim = dim + 1
endif
endif
endif
if onmarket and dayofweek=1 and hour=1 then
//dim=-1 // shows when position is active on monday first hour
endif
if onmarket and dim>=maxdim then
sell at market
exitshort at market
endif
endif
//graph dim // display days in market
//graphonprice newsl coloured(0,0,255,255) as "trailingstop atr"
//graphonprice breakvaluelong coloured(121,141,35,255) as "breakpoint"
//graphonprice breakvalueshort coloured(121,141,35,255) as "breakpoint"
//graph breakvaluelong coloured(255,0,0,255) as "breakvaluelong"
//graph breakvalueshort coloured(255,0,0,255) as "breakvalueshort"
//graph barindex-tradeindex
//graph intradaybarindex
graphonprice longma
graphonprice shortma
PaulParticipant
Master
well didn’t expect that!
here’s a new one , same code, change in moving average.
1 month on this one is still doing very good … see attached.
I put the equity curve on 15 mins TF so you could see the steadily increasing equity curve!
@Paul what lines of code cause most trades to open between 230000 and 240000 (but there a few after 240000 and before 020000)? See Paul 3 to see trade open times.
I did search for time, hour etc. It can’t be coincidence … the time band over which trades open is too regular to be a random occurence??
PaulParticipant
Master
@GraHal That there’s sometimes a trade between 23h-24h could be because of the weekend and how it handles the first hour on monday, perhaps in combination with uk settings? Because the trade is taking on sunday 23h on 31 may & 7 june. Using extra criteria like dayofweek should work to skip those trades on sunday. Nice find btw. Didn’t expect that was necessary!
Hi @GraHal, every time I say ‘no more’ to sub-1m TFs I seem to get drawn back in. It’s all your fault.
Anyway, I did a quick and dirty optimisation and you might want to try the following values:
|
|
SL = 0.75 // % stop loss
PT = 1.50 // % profit target
TS = .3 // % trailing stop
BESG = .15 // % break even stop gain
BESL = 0.00 // % break even stop level
sm = 20
lm = 40
|
Or for a higher win rate, v2
|
|
SL = 0.75 // % stop loss
PT = 1.50 // % profit target
TS = .2 // % trailing stop
BESG = .15 // % break even stop gain
BESL = 0.00 // % break even stop level
sm = 20
lm = 30
|
For it not to have blown up in 14 days of forward testing is good going – maybe it’s a cash cow?
Hi @nonetheless !
Did you continue this strategy ?
I would like to do some test with it but I did’nt find the initial post of it. Can you share these 2 versions please ? 🙂
Thanks in advance,
For it not to have blown up in 14 days of forward testing
Isn’t it 30 full days trading since 12 May?
Enthusiasm gets sapped by keep trying and trying with the same old code / strategy? Variety is the spice of life and sometimes the weirdest stuff works on odd timeframes and / or Instruments?? 🙂
Yeah thanks … I’ll try your settings also.
Hi Guys,
As we can’t do partial exits with IG (Who will be better for trailingstop), did you compare if it’s better to have a trailing stop / Breakeven or to exit all in one time with a R/R of 1/1 or 2/1 for example (No offense of course)
In my experience actually, I don’t have better results with TS/BE but may be I’m wrong
What’s your experience ?
SeeU
PaulParticipant
Master
breakeven I agree fully. Giving higher winchance, but often really small wins. Could better be used i.e. to lower the stoploss in half. Trailingstop is a must for me. Always open for improvements, so let’s see your exit criteria & calculation so I can compare in a strategy!
Hi @Paul
I think we need to use a classical/simple algos code and compare results with and without TS/BE.
If I have time tomorrow I willl do that 😉
SeeU
PaulParticipant
Master
hi zilliq
meantime I wanted to try to have a dynamic pt but stumbled on a problem.
goal is to log the highest mfe for each (long) trade.
second is to have a tradecounterlong
both above are done.
But then I need an average MFE runup based on the highest mfe from each trade. I can’t get through that part.
With average there could be a factor x as profittarget. But how to get an average?
That average could also be used as a dynamic level to activate the trailingstop.
Here’s what I have.
if barindex=0 then
tradecounterlong=0
endif
if not longonmarket then
lbc=0 // long bar counter
endif
if longonmarket then
lbc=lbc+1
hh=highest[lbc](high)
mfe=(round((hh-tradeprice(1)))*pipvalue)
endif
//added tradecounterlong=tradecounterlong + 1 when bought.
//graph tradecounterlong coloured(255,0,0,255)
//graph mfe coloured(0,0,255,255)
to code above could also be added to reset mfe to zero if there short or not onmarket.
any advise here on the average part?
PaulParticipant
Master
test setup dow or dax 1m tf
defparam cumulateorders = false
defparam preloadbars=0
c1=average[60](close)
c2=average[80](close)
condbuy=barindex>100 and c1 crosses over c2
timeframe (5 minutes,updateonclose)
st=supertrend[4,2]
timeframe (default)
condbuy = condbuy and close>st
if barindex=0 then
tradecounterlong=0
endif
if not longonmarket then
lbc=0 // long bar counter
endif
if longonmarket then
lbc=lbc+1
hh=highest[lbc](high)
mfeUP=round((hh-tradeprice(1)))*pipvalue
if mfeUP>0 then
mfelong=mfeUP
else
mfelong=0
endif
else
mfelong=0
endif
//
if condbuy and not longonmarket then
buy 1 contract at market
tradecounterlong=tradecounterlong + 1
endif
set stop ploss 100
set target pprofit 100
mfelong=mfelong
//graph tradecounterlong coloured(255,0,0,255)
graph mfelong coloured(0,0,255,255)
Hi Paul
Why do you want to a high MFE ? (And not a small MAE, or small Max drawdown an so on ) ?
PaulParticipant
Master
Hey Zilliq, This is the first part but goal is to get mfe & mae for long & short separate working. A.t.m. just want to have the calculation right and then later look what’s preferable.
Does it make sense and can it improve a strategy? I don’t know.
But the problem, it needs to take the highest mfe from each long trade, added together and divide that by the amount of long trades. Any ideas? Sounds simple but I can’t find the right way yet!
