Fractal breakout intraday Strategy EUR/USD 1H –

ALE · 2017-04-16T08:44:03+00:00

Hello guys We can discuss about the strategy here, we can help together to develop the strategy . This is the strategy posted in the Library there : https://www.prorealcode.com/prorealtime-trading-strategies/fractal-breakout-intraday-strategy-eurusd-1h/ Below you can find 3 screenshot with explanation of the stop, and levels used in this strategy. // EURUSD(-) - IG MARKET // TIME FRAME 1H // PROBACKTEST TICK by TICK - 200.000 bars // SPREAD 0.6 PIP // ALE DEFPARAM CumulateOrders = false ///BILL WILLIAM FRACTAL INDICATOR //CP=PERIOD CP=113 if close[cp] >= highest[2*cp+1](close) then LH = 1 else LH=0 endif if close[cp] =100000 and time < 230000) then C1 = (close CROSSES OVER HIL) D1 = (close CROSSES UNDER LOL) IF c1 and not shortonmarket THEN BUY positionsize CONTRACT AT MARKET ENDIF IF D1 and not longonmarket THEN SELLSHORT positionsize CONTRACT AT MARKET ENDIF ENDIF //TRAILING STOP TGL =5 TGS=5 if not onmarket then MAXPRICE = 0 MINPRICE = close PREZZOUSCITA = 0 ENDIF if longonmarket then MAXPRICE = MAX(MAXPRICE,close) if MAXPRICE-tradeprice(1)>=TGL*pointsize then PREZZOUSCITA = MAXPRICE-TGL*pointsize ENDIF ENDIF if shortonmarket then MINPRICE = MIN(MINPRICE,close) if tradeprice(1)-MINPRICE>=TGS*pointsize then PREZZOUSCITA = MINPRICE+TGS*pointsize ENDIF ENDIF if onmarket and PREZZOUSCITA>0 then EXITSHORT AT PREZZOUSCITA STOP SELL AT PREZZOUSCITA STOP ENDIF // DONCHIAN STOP DC=20 e= Highest[DC](high) f=Lowest[DC](low) if longonmarket then laststop = f[1] endif if shortonmarket then laststop = e[1] endif if onmarket then sell at laststop stop exitshort at laststop stop endif set target pprofit 30 This trading sistem can be used with other currency pairs. To find the values of variables suitable for other currency pairs, you must use WF, to avoid an overfitted values: if the WF will be done on bars 200,000 may be divided into two halves, the first test will be carried out with a large range value starting with the value in the attached picture below, the second test will have a range around the variables choices during the first test. IN THE LAST YEAR THE OPTIMIZATION OF TRAILING STOP SUGGEST THIS VALUE: TGL=9 TGS=10 … THIS MEAN ABOUT 10 PIP FOR BOTH

Viewing 15 posts - 241 through 255 (of 360 total)

← 1 2 3 … 16 17 18 … 22 23 24 →

Author

Posts
09/26/2017 at 3:24 PM #47413 quote

CKW
Participant

Veteran

OK. re-attached
Capture-1.jpg

09/27/2017 at 8:41 AM #47462 quote

Nicolas
Keymaster

Legend

“0 bar” mean that the trade has been opened and closed in the same bar, with tick mode enable, each tick of the bar is tested, so the backtest should be accurate.

09/27/2017 at 3:56 PM #47540 quote
stefou102
Participant

Veteran
Hello, I have taken the version from CKW and tried to improve, so that the code works also with other currencies (with different optimized parameters). I added:
- a closure of all opened positions before WE (avoiding Sunday gap)
- an additional filter based on Stochastics.
- additional filter for the short side (same as long side already introduced by CKW)
- I replaced the trailing stop by the one developed by Nicolas
Compared to previous version on EURUSD, backtest takes even less positions (so lower total return), but shows superior Gain/Loss ratio (>4!!). Other currencies works also, even if the KPI are a bit less impressive, although good enough to trade imo. Maybe the code are too optimized, I only have 3 years of 1H historical data. For those who have more, they can maybe run some more robust backtests? What is your opinion, readers?
victormork thanked this post

USDJPY.jpg EURUSD.jpg EURCAD.jpg EURGBP.jpg EURUSD_1H_Fractal.itf EUR_GBP_1H_Fractal.itf USDJPY_1H_Fractal.itf EURCAD_1H_Fractal.itf
10/02/2017 at 8:56 AM #47967 quote

victormork
Participant

Veteran

Hi Stefou, Nice work with trying some additional pairs. I my opinion is EURGBP the best addition. If I could make one suggestion it would be to try changing all targets and “pip based” commands to be ATR based which would avoid fitting values to current market conditions.

10/02/2017 at 9:19 AM #47970 quote

ALE
Moderator

Master

Hello as we could see in the last months , many results are different from backtest. we should find dome different way to avoid this problem some idea?

10/02/2017 at 9:28 AM #47974 quote

victormork
Participant

Veteran

Hi, I think it would be good to try a greater distance for trailing stops and profit targets. You might get a lower win rate but that’s irrelevant because you will be able to ride winners in a much more efficient way. Maybe also try using ATR to determine the distance?

10/02/2017 at 10:16 AM #47978 quote

Despair
Blocked

Master

@ALE: Which kind of difference between live and backtest did you discover?

10/02/2017 at 11:28 AM #47984 quote

ALE
Moderator

Master

X example this is an immane of last month backtest vs real
IMG_7087.png

10/02/2017 at 12:25 PM #47995 quote

Despair
Blocked

Master

Wow, this IS a difference! Did you have a look at single trades? What happens there exactly? It looks like for the live version the trades are stopped out pretty early. Are you sure both versions have the same stop loss?

10/02/2017 at 3:15 PM #48015 quote

CKW
Participant

Veteran

My edited version don’t have major significant difference. Except trade on 5-Sep triggered in live but not in backtest.
Capture.jpg

10/02/2017 at 3:32 PM #48020 quote

CKW
Participant

Veteran

Hello Stefou, I like Your “improved version”, both GL for short & buy more than 2.5. That’s great. I started yours in demo live:) thanks for your sharing!

10/02/2017 at 8:37 PM #48061 quote

stefou102
Participant

Veteran

Hello CKW you may want to edit the code where we exit all positions before WE . I know use this, to adjust when daylight changes (a copy from the forum): // quit all positions before we if (currentmonth = 3 and day >= 15) or (currentmonth = 11 and day < 8) then //daylight-saving corrections dlc = 1 else dlc =0 endif //no pos before WE If dayofweek = 5 and hour >= 22 -dlc Then If longonmarket Then Sell at market ElsIf shortonmarket Then Exitshort at market EndIf EndIf

10/04/2017 at 5:43 PM #48213 quote

Vonasi
Moderator

Master

I was trying out a strategy idea last week using the universal/standard Bill Williams fractal code found here in the library but I stopped development on the strategy when I noticed that the fractal levels calculated in the strategy code were not the same as those shown by an indicator with identical fractal code applied on the chart. I’ve just applied the same indicator with ALES fractal breakout strategy and noticed the levels also do not match. Anyone else noticed this?

10/04/2017 at 6:15 PM #48215 quote

ALE
Moderator

Master

@ Vonasi I’ve change level from High/low to close, because I want to consider close level breakout

//BILL WILLIAM FRACTAL INDICATOR

//CP=PERIOD

CP=113

if close[cp] >= highest[2*cp+1](close) then

LH= 1

else

LH=0

endif

if close[cp] <= lowest[2*cp+1](close) then

LL= –1

else

LL=0

endif

if LH=1 then

HIL= close[cp]

endif

if LL = –1 then

LOL=close[cp]

endif

10/04/2017 at 7:07 PM #48217 quote

Vonasi
Moderator

Master

Aaahh – that explains the difference. Any particular logic behind using CLOSE rather than LOW/HIGH? I believe that you ‘d get more false breakouts but then again also get earlier entry. Have you tried it with TYPICALPRICE rather than CLOSE?
Author

Posts