Navigator Trading System

Reiner · Thu, 23 Feb 2017 16:57:59 +0000

Hi guys, The tendency of the last and first days of the month is one of the most well known and reliable statistical setups. This pattern has been working for decades and in many indices. I would like to share a trading system that based on the TDOM (trading day of the month) idea from Larry Williams one of the famous statistical traders. The basic idea is to short the market around the first trading day of the month and to turn the position at the statistical monthly turning point. The system uses smart position management mechanism such as order cumulation, position sizing based on historical monthly behavior and maximal position size monitoring. Money management will be done by procentage stop loss and take profit orders. The system is very simple but amazing profitable and works without any technical indicators. The idea looks like a holy grail but of course it isn’t and I would like to point two issues: – the algo is relatively long time invested (in average 43% of the month) and this means significant overnight costs for cfd’s and potentialy a higher drawdown risk – wide stop loss value for long positions, backtest shows only moderate losses but also a significant max drawdown which mean higher volatility of running P&L which some may find hard to handle. The so-called Navigator trading system works in all intraday timeframes and the backtest was done with 200.000 4 hours candles. Please find attached the first version for the DAX. Reviews and suggestions for improvement are welcome. Best, Reiner Navigator strategy offers a statistical advantage but is of course not a holy grail and there is no guarantee to make money with it. The system is optimized for historical data and historical gains are not a garantee to be successful in the future. I strictly recommend to adjust the position sizes to your personal account size and try first in demo mode. The results in life trading will differ from the backtest results. I also recommend to start in life trading with a small account size. Anyone can use the programs for free and at their own risk.

This topic has 151 replies, 37 voices, and was last updated 6 years ago by stockdemon.

Viewing 15 posts - 106 through 120 (of 152 total)

← 1 2 3 4 5 6 7 8 9 10 11 →

04/02/2017 at 10:03 AM #30668

CN

@zilliq, that is still v2 as it looks…

CN
Participant

Topics: 1
Replies: 156
Been thanked: 17 times

04/02/2017 at 10:52 AM #30681

Pere

Zilliq: So you really think that V2 is better than V1, even when the “normal” results are better in V1? You really mean that results of V1 with WFE are worse than V2 with WFE?

I work with IG. Do you know how to start the WFE in PRT? Nicolas said that IG will put this WFE early, but I don’t know if they already did it.

Pere
Participant

Topics: 20
Replies: 97
Been thanked: 20 times

04/02/2017 at 10:59 AM #30683

zilliq

Yes I tried only on V2. I don’t understand why V2 would be worst than V1 ? Generally the last versions of a system are better

zilliq
Participant

Topics: 58
Replies: 683
Been thanked: 152 times

04/02/2017 at 11:00 AM #30686

zilliq

@Petrus

I tried on IG and it works. Do an optimisation and you will see

zilliq
Participant

Topics: 58
Replies: 683
Been thanked: 152 times

04/02/2017 at 11:03 AM #30691

GraHal

The question is … what value is the WF results using out of the optimised results for 5 periods (if using 5 periods)? It can only be the average out of the 5 results?

Whereas for example, two of my 5 periods show optimised value figures different from the other 3 periods and in those two periods the WF results are not as good. So I conclude that in the two In Sample ‘bad periods’ there was an exceptional event that gave the outlier value. This value then did not give produce good WF % for the corresponding Out of Sample period due to the exceptional event not occuring in the Out of Sample period.

So to go forward Live, I am going to use the figure shown in the 3 periods (it’s the same figure for 3 periods) which give good WF %.

Walk Forward requires interpretation of the results (in my opinion) not just using the average figure thrown out by the Walk Forward engine.

I hope above starts a discussion? Maybe I should start a new Thread?

GraHal

GraHal
Participant

Topics: 211
Replies: 8763
Been thanked: 2179 times

04/02/2017 at 11:11 AM #30692

GraHal

@winnie37 you asked in #30628 … how to proceed to do a walk forward … do you need what to do to get WF working? Or is my #30691 above what you are looking for?

Maybe I’d best read below to see if I can learn any more? 🙂 I got WF yesterday and just clicked the buttons to get it going (easy) and then thought about what the results are telling me.

https://www.prorealcode.com/blog/learning/strategy-optimisation-walk-analysis/

Also Nicolas has done the video below in French.

https://www.prorealcode.com/blog/video-tutorials/analyse-walk-forward-prorealtime/

Cheers
GraHal

GraHal
Participant

Topics: 211
Replies: 8763
Been thanked: 2179 times

04/02/2017 at 12:24 PM #30696

Pere

Well, the answer is simple:

V1 shows in demo the last trade with a gain of 3583€ during 21 days, while V2 shows only 904,80€ during 11 days.

V1 has only two trades this year, with gains of 1160€ and 3583€, while V2 has three trades in the same period, two of them with gains of 741€ and 904,80€ and one loss of 2165€, so a total loss of 519,20€.

Anyone has these results in life?

Pere
Participant

Topics: 20
Replies: 97
Been thanked: 20 times

04/02/2017 at 12:36 PM #30697

zilliq

@Grahal

Hi,

All optimized variables are good in fact.

WF IS NOT to determine optimized variable (Forget the variable he used in the results), as we do in an “normal” optimization=Non WF

WF is ONLY to determine if an optimization at equal interval can be profitable with the system

That’s why we need WFE>50 %, more than 3/5 period with >50 % , final in gain and so on…

Zilliq

zilliq
Participant

Topics: 58
Replies: 683
Been thanked: 152 times

04/02/2017 at 12:56 PM #30698

GraHal

Hi Zilliq, thanks

I’ll use an example …

WF produces 5 values, 1 each for 5 periods all different from each other (say 3, 5, 7, 12, 14)

AND this Value is

Different than the optimum value (say 10) produced during a nonWF / normal optimisation (for whole period from start to finish of 5 periods).

What use is the WF% (for values 3, 5, 7, 12, 14) if I then go and use value 10 in my Live System?

GraHal

GraHal
Participant

Topics: 211
Replies: 8763
Been thanked: 2179 times

04/02/2017 at 1:17 PM #30700

GraHal

Btw for my example above, assume all 5 periods are > 50% WF %.

I will anticipate your answer (maybe?) … I guess the 5 different WF values are telling me I need to optimise regularly?

So then, with value 14 being the most recent WF figure (and also shows a healthy WF% > 50%) do I use value 14 to go Live with?

OR

Do I use the average of all 5 periods for WF – value 8 – but keep optimising every x period (x = In Sample period for 5 periods in original WF)?

Thank You
GraHal

GraHal
Participant

Topics: 211
Replies: 8763
Been thanked: 2179 times

04/02/2017 at 4:53 PM #30730

zilliq

Simply use the last Optimize variable for the next period

Remember WF is ONLY to see if you can win in the next period when you optimize your system

We don’t see the future but if you have a WFE>50 % you have a much higher probability than if WFE<50 %

Another way : if you have a WFE>50 % you have a high chance than your optimization will be reproductible

I’m pretty sure than the vast majority of the system on PRT or others on the Net don’t pass the WF test…

zilliq
Participant

Topics: 58
Replies: 683
Been thanked: 152 times

04/02/2017 at 5:41 PM #30733

actionjackson

I wanted to know if the DAX tdom’s are stable for a longer period (20 years).

I migrated Navigator DAX H4 V1 version for daily timeframe and the result confirm the tdom of 1 (sell) and 8 (buy).

In the code the daily version buy tdom is 9 but in a daily timeframe this at 1.00 pm and it’s more or less the same as 8 and 17 CET in H4.

It’s interesting that this simply idea works so good for the last 20 years. Draw is unfortunately high.

2 users thanked author for this post.

Pfeiler, Bandido

actionjackson
Participant

Topics: 1
Replies: 8
Been thanked: 10 times

04/02/2017 at 8:24 PM #30751

GraHal

@zilliq you say … use the last Optimize variable for the next period … so in my example where the WF 5 periods In Sample Optimised values are 3, 5, 7, 12, 14, you are saying to use variable value 14 to go Live with? Accepting that the value 14 has a good WF% > 50%.

Thanks
GraHal

GraHal
Participant

Topics: 211
Replies: 8763
Been thanked: 2179 times

04/03/2017 at 7:09 PM #30878

kg6450

@Grahal

Just tried it now and it works thanks!

I was thinking that it was to do with the times not working with the 4H timeframe but didn’t notice the time on line 62, thanks!

kg6450
Participant

Topics: 13
Replies: 43
Been thanked: 4 times

04/13/2017 at 10:35 PM #31995

Paris

both systems have opened positions .

i would like to add to code system a line with “graph ” that will use to see the take profit and stop loss .

i ve tried different ideas but nothing works really .

does anybody have an idea

Paris
Participant

Topics: 6
Replies: 145
Been thanked: 11 times
Author

Posts

Viewing 15 posts - 106 through 120 (of 152 total)

← 1 2 3 4 5 6 7 8 9 10 11 →

Find exclusive trading pro-tools on

Navigator Trading System

2 users thanked author for this post.