Topic: Big Big Dilemma

hi Popov and everyone.

I am generating some portoflios and i am able to generate them only in USDJPY. I test them in EA studio and then on MT4 tester.I saw the results are the same so i usually start with real trading and so far i generate a ton of profit.

Now come the problems. With all the other crosses, every strategy is not good. They are really bad! To be precise on EA studio they looks good but on MT4 tester and on real account they lose money.

So today i downloaded (for the first time) Petko’s EAs for EURUSD i put them on EA studio with my broker’s data (obviously) and they were all losing. Then i place them on MT4 Tester and they were all in profit.



So my question is: is possibile that EA studio is using some indicators that my broker hasn’t or has different? And again is Mt4 tester more reialable than EA studio.

Thank you

Re: Big Big Dilemma

The indicators in EA Studio and in MT work in the same way.

When the strategy uses SL and TP, it may happen that both exits are possible in the same bar. In those cases, EA Studio tries to protect your account and always touches the SL. We call that "Account protective algorithm". The purpose of that model is to prevent showing overestimated results because there is no sure way to determine the actual result. on the other hand, MT literally fabricates ticks from a thin air and shows a random result.

So my question is: is possible that EA studio is using some indicators that my broker hasn’t or has different?

The exported Expert Advisors contains all the necessary components in order to work best in your account.

And again is Mt4 tester more reliable than EA studio.

The brokers pay a ton of money to MetaQuotes for that tester. It must be good enough for them. On the other hand, EA Studio is designed to be liked by traders.

Re: Big Big Dilemma

hello Popov,

sorry to disturb you again.

So it is possible that the same strategy gives you completely different results in EA Studio from Mt4 tester, right?

On USDJPY it never happens

thank you!!

Re: Big Big Dilemma

The particular symbol doesn't play any role in the backtesting. The formulas are absolutely the same. The different performance comes from different data range, spread, volume, swap, commissions...

If you have your data correctly imported in EA Studio, the strategy must show very near results in EA Studio and in MT. You may have a difference if you use near SL and TP.

If you find a difference between MT and EA Studio, please examine the data and the deals. If there is a difference, please report. Please attach the expert and screenshots that show the exact deal that differs. I'm sure we will find the reason and if it is a bug, we will fix it immediately.

5 (edited by erndu 2019-05-04 15:31:55)

Re: Big Big Dilemma

here it is

Before Ea Studio

https://i.ibb.co/7VBq40j/popov-1.png

6 (edited by erndu 2019-05-04 15:32:24)

Re: Big Big Dilemma

https://i.ibb.co/8M4jjzg/popov3.png

Re: Big Big Dilemma

here it is the Portfolio: https://www.dropbox.com/s/z9puk45u7h5i9er/Portfolio%20Expert%20EURUSD%20M15%20%281%29.mq4?dl=0

here is my data:
https://www.dropbox.com/s/vleffc7hx5av8xn/USDJPY15.json?dl=0

thank you Popov

Re: Big Big Dilemma

After Mt4 Tester

https://i.ibb.co/khmvgxW/popov-2.png

sorry but the forum doesn't allow to post too many links in one message

Re: Big Big Dilemma

Popov wrote:

The particular symbol doesn't play any role in the backtesting. The formulas are absolutely the same. The different performance comes from different data range, spread, volume, swap, commissions...

If you have your data correctly imported in EA Studio, the strategy must show very near results in EA Studio and in MT. You may have a difference if you use near SL and TP.

If you find a difference between MT and EA Studio, please examine the data and the deals. If there is a difference, please report. Please attach the expert and screenshots that show the exact deal that differs. I'm sure we will find the reason and if it is a bug, we will fix it immediately.

I too, have a difference with live trade results and EA Studio backtest, they dont correlate sometimes.
"You may have a difference if you use near SL and TP."
What is near in this case?

Re: Big Big Dilemma

I mean that there are many bars where the SL and the TP are in the same bar range. It depends on the data period and the SL, TP.

EA Studio always touches the SL in that case - "Account protective algorythm". The purpose is to prevent showing overestimated results.
EA Studio shows Backtest Quality lower than 100% in such cases.

When you test / trade such strategy, it will always show higher result because it will touch TP in some cases.

Re: Big Big Dilemma

Hello Popov,

i sent you a private message also. I have checked all the big cross. The results of EA Studio and of Mt4 T. are the same for every cross with JPY. EURJPY, USDJPY and AUDJPY.

They are different instead for every timeframe for all the others cross.

Re: Big Big Dilemma

Popov wrote:

I mean that there are many bars where the SL and the TP are in the same bar range. It depends on the data period and the SL, TP.

EA Studio always touches the SL in that case - "Account protective algorythm". The purpose is to prevent showing overestimated results.
EA Studio shows Backtest Quality lower than 100% in such cases.

When you test / trade such strategy, it will always show higher result because it will touch TP in some cases.

ok I understand..
If you would guess, for eg EURUSD M5, how much apart should the TP and SL be without messing up the EA Studio backtest?

Re: Big Big Dilemma

I wouldn't call it a mess-up. There's no hard rule how much apart they should be, it depends. If you really want to know exact number, then construct such a strategy and use your own data and start making the gap between TP and SL smaller and smaller incrementally. While doing so, pay attention to the backtest quality metric. If you have a significant drop in that number, you can formulate your exact measurement. But that applies only in those given conditions, only in that data set. New data needs new calibration, if you wish to do so.

Re: Big Big Dilemma

Footon is right!

I would only add that the TP is not your best friend.

Re: Big Big Dilemma

If you find such a big discrepancy for a particular strategy and it has high Backtest Quality in EA Studio, please report.
If we find a bug, we will fix it immediately.  A possible place for a mistake is the exact setting of the MQL indicator parameters.

I say that this is possible because we cannot test it within EA Studio. Such a bug can be found only if the MT indicator shows different results on the same data set and settings.