Different Tester Results for Two Computer Using the Same EA input Parameters

New comment
 

Complement of the new year.


I need help to make sense of the tester results.

I have developed several EA for Mt5. When I optimize the EA using the same range of parameters with two different laptops, I get different set of results. Further, when I take optimized results/parameters from one Laptop and use the same parameters for the two different laptops, I get different results.

This makes it difficult to comprehend and make decision on which are reasonable parameters to use. And if it gives different results with the two laptops, how confident will I be that when I deploy the EA, I will get the results as expected from testing.

Please any one to assist in making sense of this.

Looking to hear from you.

Regards

 

You have not provided sufficient information to analyse your issue.

  • Specify the type of modelling used.
  • Specify if this is the exact same broker and account, or not.

If you are using the exact same broker, account, modelling type, test period, input parameter, etc. then describe in detail the types of differences you notice.

If everything is the same, except the results, then the may be data corruption or bugs in EA's code.

So make sure the quality measurement in the test results are good and are the same in both cases.

 
Fernando Carreiro #:

You have not provided sufficient information to analyse your issue.

  • Specify the type of modelling used.
  • Specify if this is the exact same broker and account, or not.

If you are using the exact same broker, account, modelling type, test period, input parameter, etc. then describe in detail the types of differences you notice.

If everything is the same, then the may be data corruption or bugs in EA's code.

Also make sure the quality measurement in the test results are good and the same in both cases.

Thanks

The testing are done under the same conditions.

1. I am testing using data from Meta Quotes-Demo account for both two laptops.

2. Same durations when optimizing and testing.

3. Modelling using-Every tick

4. Same optimization process (being Fast Generic based algorithm or Slow Complete algorithm)


The results are different in terms of

1. Net Profit (in some cases a difference of more than 20 times)

2. Gros Profit

3. Gross Loss

4. Draw Downs,

5. Profit factor

6. Recovery factor

7. Sometimes in the number of trades

8. Same Delays


You indicate that I need to ensure the quality measurement in the test results are good, I am not sure what you mean by this and how do I go about this?

Regards

 
Niyoshaka Stanley Lameck #: 1. I am testing using data from Meta Quotes-Demo account for both two laptops.

Don't use MetaQuotes demo accounts! They are for beta testing and can change without notice. You will also end up updated with beta builds of MetaTrader which can have bugs.

Always use demo accounts from a regulated and reputable broker, and make sure you only use the officially released MetaTrader builds to maintain stability. The current official build is 4755 (13 Dec 2024).


Niyoshaka Stanley Lameck #: You indicate that I need to ensure the quality measurement in the test results are good, I am not sure what you mean by this and how do I go about this?

Ideally it should be 100% ... The "red" parts indicate inconsistent or missing data ...


 

Thanks

I see the data quality and it has been green all along during the testing.

I will recheck again using the demo account of another broker with the two laptops.

I had compared the results of of two laptops with data from Meta Quotes-Demo account as well as comparison of Meta Quote-Demo account and another Broker demo account on the same laptop but I had not tested the other broker with the two different laptops which was also different.

Regards

 
Niyoshaka Stanley Lameck #: I see the data quality and it has been green all along during the testing. I will recheck again using the demo account of another broker with the two laptops. I had compared the results of of two laptops with data from Meta Quotes-Demo account as well as comparison of Meta Quote-Demo account and another Broker demo account on the same laptop but I had not tested the other broker with the two different laptops which was also different.

You cannot compare with different accounts or different brokers. They will not have the same data feed.

Forex and CFD's do not have centralised exchanges, and different brokers and even different account types, will have different data feeds.

New comment