LynyParticipant
Junior
@Raul, Grahal, manel, i will stop Live trade and just let the Demo run on V3. let continue test it then.
Hi,
at the end of all I think that there are some problems of backtest tick by tick and the function internal of trailing. Until PRT don’t solve these problems each strategy with back test with candle with much value zero there are high risk that that strategy don’t perform well. PRT must confirm when the problem will be solved because these our test confirmed that there are some problem in PRT. Thanks
Does anybody backtest the systems with real account ?
because i know the demo account of IG and PRT in demo account is full og bugs .
Some reasons to hope :
some systems are positive .
Some systems are pretty close to their backtest
and some systems loses less than their backtests .
of course all backtests are very good .
Paris- I´m backtesting with 10.3 live. Look at my latest post https://www.prorealcode.com/topic/5-min-mini1e-dax-spread-1/page/26/#post-23517
i mean backtest in demo account and backtest in real account . In My oppinion those are not the same . The backtest in real ( money …) is more accurate . My broker is IG .
Can anyone confirm this by posting a backtest in a long period to compare both ?
@Joachim I’m confused, am I missing something? They are not the same candles in the left and right pics on the post you refer to above?
Cheers
GraHal
Paris-
I’ve 10.3 both live and demo. Here´s a backtest from 2016-07. Pink is live and blue is demo. Same result.
GraHal-
One is normal candlestick and the other one is HA. But thats not the thing. The picture on the left is real result from live trading and the picture on the right is backtest during the same period.
I still think that you should not throw in the towel, you have to follow it in demo to see how it acts for a while. The fact that when changing the hour leaves a nefarious money curve. You may gradually gain, you have to wait for PRT to improve the backtest.
@Raul
hi,
all type of system are running in test in order to evaluate all. I wait for PRT to improve the backtest. Thanks
I am running the strategies in live (with greatly reduced risk of course) as that’s the only real way we can ascertain the issues. I tend to agree with Paris above that running in Demo accounts is fine to a certain degree but I also think that the demo account modules/software is not the same as Live as I have found discrepancies in both in the past so the only way is to actually test in Live (if you can afford the capital).
But I have amended all of code in these live strategies with the manually coded trailing stop which provides more accurate results in backtesting (I’ve also compared live account backtesting with live actual trading results of the past week for 7 strategies to confirm that this code is indeed more accurate than the inbuilt PRT trailing function). The equity curves are of course no way near as great as we first encountered when we started this journey but we now know that those results were fake and the tick by tick engine/trailing stop is not working properly in PRT.
However, with exception of a few strategies, almost all of them still show an overall profit (albeit smaller) over the course of a long period (ie 6-12mths) so like Raul mentioned, it’s not the time yet to give up. There will be some work to do in analysing and tweaking as there are long periods of small drawdowns/returns but this is part of the process. I will compare these live trading results over the next week or two and let you know if anything interesting comes up.
In the meantime hopefully PRT is working on solving the tick by tick calculation and trailing stop code which are currently both operating incorrectly.
**-NEWS-**
The problem has been spot by the PRT dev team. Seems that an issue exist when backtesting in tick/tick mode on the current day only. The problem don’t exist on days prior the current day.
A fix will be deployed in the next few days.
Thanks very much for the update Nicolas. Good to see some progress.
@ nicholas
i m not sure to understand well .
if i backtest the day of today for a system in tick tick it will a diferent result than if i backtest the day oftoday but monday?
but does the result of monday will it be the exact live of friday ? i m not sure …..
can you confirm ?
@Paris you understood well.
Hello, The system works in real on my account.
It opened a sales position on the dax at 090528 at 11,640.4
He closed the position at 090818 to 11640.4
No gains or losses.
However, the dax decreased to 090601 to 11 630.9.
The trailing stop of 5 should therefore have dropped to 11 635.9.
So I had to earn 4 points.
The stop trailing does not work in real.
We can not win.
I sent a report to PRT CFD. I’m waiting for the return.
Same to 091541. purchase 11 654.2.
The dax rises to 11,661.7 to 091653.
But it closes my position at the purchase price at 091753.
The cards are skewed. It’s a scam.
