Pathfinder _ zilliq research
-
Hi Arno,
For me, it’s against my religion to average my losses, sorry.
I think it can work ONLY if you have a very high % of winning trade (in the past, but who can predict the future)
The problem is what I said to Reiner before on my previous post : Pathfinder take more position than it could (depending on, the risk). If it works, no problem, you’re rich, but if you lose you cry
I think (it’s my opinion, not a critical), that this booster to 10 positions (if I remember) is too dangerous and against the Van Tharp theory.
This afternoon I had a little bit time to code different suggestions I spoke to Reiner
1/ Add position ONLY if you are after the trailing stop = pyramid trading
I code a simple strategy DAX ut15 Spread=1 and test it
As you will see on the pictures
With this system you have higher winning trade, higher ratio and you’re less in the market (and so less expensive) ! +8.34 % vs +7.91 %
It need to be confirm but seems promising
2/ Add an Alembert Martingale (to the basic code) (with maxposition depending on the risk)
It seems you win not more and with higher drawdown and you’re more in the market
I think it can be optimize (Trailing stop and so on)…But not sure it could be a good option +7.02 %
I will test other suggestions when I will have time and say to you the results
Have fun
Zilliq
Well it seems my idea was not so bad …
Instead to ponderate the number of position by saisonnality I wanted to ponderate the number of position with the number of call or short trade when there is a loosing trade (=less lag)
ZRatio :-)=winning/losing WHEN the trade is loosing
The purpose is to ponderate the winning or loosing trade, with this ratio. Ponderate the call trade when short trade are more often losing and vice versa
If there is less call trade who are loosing I will ponderate them and same with short
For this I create a very simple code (MM crossing) UT15 Spread 1 Capital 10000
DEFPARAM CumulateOrders=true DEFPARAM PRELOADBARS=2000 REM Achat once ordersize=1 once exitindex=-2 indicator1 = ExponentialAverage[a](close) indicator2 = Average[b](close) c1 = (indicator1 CROSSES OVER indicator2) IF c1 THEN BUY OrderSize contract ROUNDEDUP AT MARKET ENDIF REM Vente indicator3 = ExponentialAverage[c](close) indicator4 = Average[d](close) c2 = (indicator3 CROSSES UNDER indicator4) IF c2 THEN SELL AT MARKET ExitIndex=BarIndex ENDIF c3=(indicator1 CROSSES under indicator2) IF c3 THEN SELLSHORT OrderSize contract AT MARKET ENDIF c4=(indicator1 CROSSES over indicator2) IF c4 THEN EXITSHORT AT MARKET ExitIndex=BarIndex ENDIF // positions sizing IF Barindex= ExitIndex+1 THEN ExitIndex =0 IF PositionPerf(1)<0 THEN if countofposition[1]>0 then w=w+1 elsif countofposition[1]<0 then l=l+1 endif OrderSize = 1 //OrderSize+1 ELSIF PositionPerf(1)>=0 THEN OrderSize =1 //MAX(OrderSize-1,1) ENDIF ENDIF graph w coloured (10,204,10) graph lWithout the ponderation and the ratio we obtain with this simple strategy Gain +5.02 % VQN 1.06
With the ponderation AND the ratio we obtain Gain +20.56 % (X4!) VQN 1.56. Who is much better 🙂
The code with the “Zilliq Ponderation” 🙂
DEFPARAM CumulateOrders=true DEFPARAM PRELOADBARS=2000 REM Achat once ordersize=1 once exitindex=-2 if w<>0 and l<>0 then pondc=round(l/w) ponds=round(w/l) else pondc=0 ponds=0 endif indicator1 = ExponentialAverage[a](close) indicator2 = Average[b](close) c1 = (indicator1 CROSSES OVER indicator2) IF c1 THEN BUY OrderSize+pondc contract ROUNDEDUP AT MARKET ENDIF REM Vente indicator3 = ExponentialAverage[c](close) indicator4 = Average[d](close) c2 = (indicator3 CROSSES UNDER indicator4) IF c2 THEN SELL AT MARKET ExitIndex=BarIndex ENDIF c3=(indicator1 CROSSES under indicator2) IF c3 THEN SELLSHORT OrderSize+ponds contract AT MARKET ENDIF c4=(indicator1 CROSSES over indicator2) IF c4 THEN EXITSHORT AT MARKET ExitIndex=BarIndex ENDIF // positions sizing IF Barindex= ExitIndex+1 THEN ExitIndex =0 IF PositionPerf(1)<0 THEN if countofposition[1]>0 then w=w+1 elsif countofposition[1]<0 then l=l+1 endif OrderSize = 1 ELSIF PositionPerf(1)>=0 THEN OrderSize =1 ENDIF ENDIF graph w graph lAfter small Walk Forward Optimisation Gain +39.08 % VQN 1.02 / Bad VQN but not bad for a “simple” strategy
This is the previous results I will work on it to optimize and try the other ideas
Cheers
Zilliq
https://www.zilliqtradingresearch.fr/
Coding is not a crime 😉
Robin thanked this postJust for fun, if you do a Alembert Martingale (I use the code of Nicolas )
We obtain +49.42 % VQN 1.26
Remember it’s just a crossing average and a little bit of Magic
🙂
Zilliq
CN thanked this post
- You must be logged in to reply to this topic.
Summary
This topic contains 25 replies,
has 9 voices, and was last updated by 
8 years, 12 months ago.
Topic Details
| Forum: | ProOrder: Automated Strategies & Backtesting |
| Language: | English |
| Started: | 03/03/2017 |
| Status: | Active |
| Attachments: | No files |
