Download MetaTrader 5
To add comments, please log in or register
Publish your article and get a reward!
Vener Garayev
Vener Garayev 2014.06.13 21:47 


I used to test my data on my old computer and everything was fine. When I bought a new computer and installed MT4 on it, the results I get on the new computer differ drastically from the ones on the old computer.

I transferred the same history data for EURUSD from old computer to new one, so there is no difference in data. Same broker data, same parity tested, same testing dates, same testing parameters, same spread. NOTHING is different. However, the results are different.

The only difference is the build of the MT4. On my old computer the MT4 is build 509, while on my new computer it is build 649 (I believe the latest). 

So my question is, why would two different builds of MT4 produce different results, while all other things are same?



SYED NAUFAL GADDAFI 2014.06.13 22:45  
To get your answer, you need to upgrade MT4 509 to latest inside the old computer.
whroeder1 2014.06.13 23:32  
veneqara: why would two different builds of MT4 produce different results, while all other things are same?
Are you using the exact same spread?
Ian Venner
Ian Venner 2014.06.14 03:47  

If the spread is not the difference I would also check the chart history data again. I did 4 year, 1 minute backtest comparisons on builds 509 vs 604, 605, 606, that showed with identical chart data the results were identical except that the new builds were faster.

Vener Garayev
Vener Garayev 2014.06.14 12:35  

deysmacro, I don't want to lose my build 509 in which my tests are realistic. Something is problematic with the new build, and I would like to know what. Bringing my different MT4 platforms to same build would not give me anything.

WHRoeder, yes, even the spreads are same.

SDC, the chart history data is exactly same, because I exported it from the old computer and imported to new computer. I checked every single possible issue so that everything is same. The only difference, as you mentioned, is that the new computer completes testing much faster. In fact, it should be otherwise, because what I call "my old computer" is much faster in terms of its processor, speed, hard drive, etc. So, I normally expect the "new computer" to be slower, while it completes testing much faster. So, I suspect that the new build does not read all data or something like that. Because at the same time the results differ in combinations. For example, in my "old computer" my EA produces results with large numbers of TOTAL TRADES produced, while in my "new computer" the results produce only 4-5 TOTAL TRADES for almost all combinations for same period!

Again, let me repeat the problem. I am using the same expert advisor on both computers, the only difference I found so far is the build of the MT4. My testing period is same (December 2012 - May 2014), my parity is same (EURUSD), my testing period is same (H1), my all optimization parameters are same, my optimization spread is same, my historical data is same. My only suspicion is about how the new build of MT4 processes testing. Otherwise, there is no way that everything else is same, the results would differ so much.


Ian Venner
Ian Venner 2014.06.15 06:15  

I think either your EA is inconsistant or your chart data is corrupted. Examine these two strategy tester reports. It is the same EA coded, compiled and tested on 509 then recreated, compiled and tested on the latest build 660.

build 509 report 


 build 660 report

To add comments, please log in or register