Is it possible/useful to cumulate the trades only if the position is positive?
Reiner should confirm, but it would completely deserve the strategy since cumulating orders while in negative profit increases order basket capacity to exit faster in gain.
Hi Flo,
The strategy will also works in daily timeframe but daily timeframe has the disadvantage that you can only react at 1 am because all PRT daily timeframes are synchronized to that fix timestamp.
The results are much worse if you change the averaging down strategy (average profit <0) for long positions to an averaging up strategy (profit >0). The simple message is: accumulate low and sell high when markets recover easy to say but hard to manage for most of the traders!
If you are interested to understand why this simple setup is so reliable please read the following valueable study.
Best, Reiner
CNParticipant
Senior
Paintrade atm, anyone else having this in live?
With the Trump speech ahead, I canceled it yesterday.
10 short positions is just too much risk for my budget at the moment.
I really hope I was wrong, since I want to rely on this algo.
Thank you for the study! very interesting!
I have not read it in detail yet, but since I have no experience in coding, I will try to implement this idea as a first learning exercise.
Another thought I had… probably stupid, but what about combining pathfinder 4h and navigator?
only go long if both systems are long and short if both of them are short?
CNParticipant
Senior
Pathfinder just closed a short -10 pos with a HUGE loss, live-result attached bellow.
Massive blow, 2,100 euro loss.
Hi Reiner,
thanks a lot for this system. I’m amazed by your creativity and generosity. And I love that you are fascinated with repeating time patterns.
If you have time, I have two questions for you re: Navigator
- The supporting research you attached (btw, I love it that there is supporting research to a statistical model, this is VERY reassuring!) seems to suggest that the best moment to go long is T-3 to T+3, and the best moment to go short is T-8 to T-4 (see attached page). As these are not the dates you have considered in your algo (unless I got something wrong), I wonder if you have tested the above intervals and why you came up with day 1 and day 8 ?
- My second question has to do with risk and MM. Your algo works with number of contracts in absolute terms, meaning that in the beginning your 10x contracts max can mean a 10x leverage (even more so if we start today), while on the long run it can also mean that you are underleveraged. In a nutshell: too much risk at the beginning and not enough towards the end. So have you thought about reasoning in terms of leverage ? For instance a constant 2 to 4x maximum number ? (also thinking in leverage terms means that you can more easily understand the risk you are taking with each position. If you have a 4x leverage and you face a 1% negative position you know that your equity is 4% down and that you cannot be wiped out unless the market is 25% against you)
Just my two cents on a fantastic idea, looking forward to hearing your comments,
Best,
B
Jesus .. highest loss in the backtest shows 1.600€ loss, and the first trade of the algo results in 2.100€ loss?
Looks like the backtest is not accurate for this one. I will move it to the demo account for more results.
Why doesnt this losing short trade show when I run a backtest today?
Not a great first trade but that’s part of the process unfortunately. The results of a longer backtest (10yrs+) are actually consistent with this loss (see earlier post for the outcomes) so I see no issue with it at the moment. The max single loss in backtest was £4k for v1 and £3k for v2. As I mentioned before, the draw is high on this and this is a new system and still in progress so we should give it a chance for people to develop it further (as we already are doing). Hence, I agree that following it in demo is the best way at the moment, I would not advise people to just jump in with live trading just yet before they are fully aware of the risk parameters.
We see the limits of the optimisation. When a big loss increase the drawdown, the month is “descativated” with 0 position; but the statistical princip is just a belief for the future, and the “good” months in the past might be the worst in the future. So as the statistics are the only indicator for that system, I am not enough confident to run it live. It is very different of Pathfinder, that combines seasonality with several indicators (levels, averages, etc). Anyway, it worths elaborating and testing, so great thanks again to Reiner.
Hi CN,
That are really bad news. You mean Navigator and not Pathfinder.
If I get it right you have taken the version out of my discussion with @manel about the high drawdown and his suggestion how we could avoid this problem. The original version didn’t take this trade but this is of course no consolation for you. I have mentioned that I still working on it and have raised the issue with the high drawdown but maybe it was my fault to make it not clear enough.
I don’t trade this system life but before life trading I would test it in demo mode first and afterwards I would start with a small account size to verify the backtest as we all have learned that the results will differ in life trading. Maybe I should have made this point clearer.
The outcome of a trading system is the result of hundreds of trades and losers are unfortunately something you have to bear.
Best, Reiner
@Pfeiler
Totally agree with you, 2100 euros is too much lost for a single operation. Maybe we should adjust the risk a bit based on the size of the positions, and continue testing it in demo, but I’m agree with @manel, it’s part of the process.
Hi Reiner,
About athfinder did you see my message on the file, who can explain the loss of CN ?
Regards
Yes I agree 2100 euros is too much lost for a single operation (and sad if it’s the first trade !) but I note that the MaxDD for Navigator is around 4000 . So loose 2100 is regular and timing is the stats ‘fantasy : stats win only in the long run ….
Thanks Reiner for sharing and working on Navigator.
