MT5 History Quality only 48%

New comment
 

Hi folks,

So tonight I decided to do some back testing and optimisation of one of my EA's. I've been working on this one for a few weeks now

Up until tonight I had always got 100% History Quality with my tick data when back testing, however for some reason tonight I only got 48% history quality

This was on several forex instruments, not just one. My broker is a very well known large broker based in Australia and up until tonight I'd never had a problem with the history data in MT5

Any suggestions what it could be? could it be my broker? or maybe something wromg with my terminal

I have gone through and checked everything, however I'm still only getting 48% history quality where as I would normally get 100% history quality

Many thanks, Paul  

 

I would ask you to consider, that we, on the other side of the screen, have no idea what you are doing.

We cannot see your computer nor can we read your mind. So, you will have to help us out, with a "little" more info, right?

Things like ...

  • What mode are you testing — open prices, M1, tick, or real tick?
  • What date range are you using for your back-test?
  • How far back does your historical tick data go?
That would be the "bare minimum" information, but the more you provide, the more informed our answer will be.

 
Fernando Carreiro #:

I would ask you to consider, that we, on the other side of the screen, have no idea what you are doing.

We cannot see your computer nor can we read your mind. So, you will have to help us out, with a "little" more info, right?

Things like ...

  • What mode are you testing — open prices, M1, tick, or real tick?
  • What date range are you using for your back-test?
  • How far back does your historical tick data go?
That would be the "bare minimum" information, but the more you provide, the more informed our answer will be.

Thanks for your reply


I am currently back testing using 'every tick based on real ticks' mode

I am testing from mid June 2022 to yesterday

Previously I have never had an issue with back testing for say the last year with 100% history quality. Now it seems to be a problem


*update*

I have carried out some more back testing since my original post using different date ranges and ended up with different results

Fron June 22 to June 23, 48% history quality

December 2022 to June 2023 95% history quality

January 2023 to June 2023 100% history quality 


I must admit that I am starting to wonder if this is a broker issue and not an MT5 issue


Thanks for your help

 
Bollinger68 #:

Thanks for your reply


I am currently back testing using 'every tick based on real ticks' mode

I am testing from mid June 2022 to yesterday

Previously I have never had an issue with back testing for say the last year with 100% history quality. Now it seems to be a problem


*update*

I have carried out some more back testing since my original post using different date ranges and ended up with different results

Fron June 22 to June 23, 48% history quality

December 2022 to June 2023 95% history quality

January 2023 to June 2023 100% history quality 


I must admit that I am starting to wonder if this is a broker issue and not an MT5 issue


Thanks for your help

History quality in MT5 Strategy Tester is always a broker issue, since MT5 downloads tick data from the broker server, so your broker has good quality real tick data for the recent history but not so good for further back in the past periods.

You can see details about the downloaded ticks in the Strategy Tester Journal.

You should try with another broker too, to improve your results.

 
Eleni Anna Branou #:

History quality in MT5 Strategy Tester is always a broker issue, since MT5 downloads tick data from the broker server, so your broker has good quality real tick data for the recent history but not so good for further back in the past periods.

You can see details about the downloaded ticks in the Strategy Tester Journal.

You should try with another broker too, to improve your results.

Thanks for your reply...

I have just received this email below from my broker which still doesn't explain to me why I had absolutely no problem with my historical tick data last week, and then suddenly after a break of a few days I now have a problem with the historical data. Is it my broker? is it Metaquotes?

Thank you very much for contacting XX XXXXXXX Customer Service!
 
Please note we do not upload the tick data if you are backtesting Tick by Tick you won't be able to get 100% of the data, the data we provide are the ones you can download directly from your Metatrader platform
 
The historical data from the platform are provided by Metaquotes, hence XX XXXXXXX cannot provide this data.



And yet in their previous email they said they only keep the data for a few days, however last week I could download data for the last year without any issues. Even after this problem started, I am still able to download data for all of 2023, just not for the last year like I was able to do up to last week

Hope this all makes sense 

Any suggestions?

Thanks

 
Bollinger68 #:

Thanks for your reply...

I have just received this email below from my broker which still doesn't explain to me why I had absolutely no problem with my historical tick data last week, and then suddenly after a break of a few days I now have a problem with the historical data. Is it my broker? is it Metaquotes?

Thank you very much for contacting XX XXXXXXX Customer Service!
 
Please note we do not upload the tick data if you are backtesting Tick by Tick you won't be able to get 100% of the data, the data we provide are the ones you can download directly from your Metatrader platform
 
The historical data from the platform are provided by Metaquotes, hence XX XXXXXXX cannot provide this data.

And yet in their previous email they said they only keep the data for a few days, however last week I could download data for the last year without any issues. Even after this problem started, I am still able to download data for all of 2023, just not for the last year like I was able to do up to last week

Hope this all makes sense 

Any suggestions?

Thanks

  

I am not sure they understood that you were talking about the MT5 strategy tester, but in any case you can check from where the tick data is downloaded in the Journal of the strategy tester.

If indeed the real tick data is downloaded from the Metaquotes demo server, then look for another broker to perform your tests.

 
Eleni Anna Branou #:

I am not sure they understood that you were talking about the MT5 strategy tester, but in any case you can check from where the tick data is downloaded in the Journal of the strategy tester.

If indeed the real tick data is downloaded from the Metaquotes demo server, then look for another broker to perform your tests.

Thanks again for your reply...

I must admit that I also got the feeling that they didn't understand the question too. I had to email them again as they didn't exactly answer my first question, even though I went to great lengths to explain my problem.

It may be that they were assuming that MT4 and MT5 historical data is handled the same way when I believe it isn't. It just seems strange that something has clearly happened in the last few days to give me very different results, either at my broker or Metaquotes. 

With regard to the journal, the earliest start date was 2022/12/20, even though I had selected a date in June 2022.

I also noticed that the log file address was... 

log file "C:\Users\pauld\AppData\Roaming\MetaQuotes\Tester\   

So does this mean that Metaquotes are responsible for the data? Is this then a Metaquotes issue?

It now looks like my MT5 terminal is only able to download historical tick data for the last 6 months, whereas last week I could download historical data for the last year. What has changed in the space of a week?

Thanks again



 
Bollinger68 #:

Thanks again for your reply...

I must admit that I also got the feeling that they didn't understand the question too. I had to email them again as they didn't exactly answer my first question, even though I went to great lengths to explain my problem.

It may be that they were assuming that MT4 and MT5 historical data is handled the same way when I believe it isn't. It just seems strange that something has clearly happened in the last few days to give me very different results, either at my broker or Metaquotes. 

With regard to the journal, the earliest start date was 2022/12/20, even though I had selected a date in June 2022.

I also noticed that the log file address was... 

log file "C:\Users\pauld\AppData\Roaming\MetaQuotes\Tester\   

So does this mean that Metaquotes are responsible for the data? Is this then a Metaquotes issue?

Of course not, it's just the name of the company who provide the software, as you will find folders with Microsoft or other similar names.

The broker is ALWAYS responsible to provide the data. Though beside the software (MT5), I suppose MetaQuotes can also provide some historical data on the server side.

It now looks like my MT5 terminal is only able to download historical tick data for the last 6 months, whereas last week I could download historical data for the last year. What has changed in the space of a week?

Thanks again

Post the log if you want useful help. all the rest is just guessing.

 
Bollinger68 #:

I have carried out some more back testing since my original post using different date ranges and ended up with different results

Fron June 22 to June 23, 48% history quality

December 2022 to June 2023 95% history quality

January 2023 to June 2023 100% history quality 

I must admit that I am starting to wonder if this is a broker issue and not an MT5 issue

As asked previously ... "How far back does your historical tick data go?"

In another thread of yours, I explained how to do that, by opening up Symbols (Ctrl-U), going to the Ticks tab, requesting the data and seeing how far back it goes.

From your tests, it seems it only go back to the start of the year.

EDIT:

Please note that if your EA is multi symbol, or if the symbol in question needs to be converted into the account currency, then extra tick data is necessary. If any one of the required symbols has less tick data then the rest, then the quality drops.

For example, lets say you are testing EURUSD which has 1 year's worth of tick data, but your account currency is AUD, and AUDUSD only has 6 months of tick data, then you will only be able to test at 100% for a 6 months range and not for a full year.

 
Fernando Carreiro #:

As asked previously ... "How far back does your historical tick data go?"

In another thread of yours, I explained how to do that, by opening up Symbols (Ctrl-U), going to the Ticks tab, requesting the data and seeing how far back it goes.

From your tests, it seems it only go back to the start of the year.

EDIT:

Please note that if your EA is multi symbol, or if the symbol in question needs to be converted into the account currency, then extra tick data is necessary. If any one of the required symbols has less tick data then the rest, then the quality drops.

For example, lets say you are testing EURUSD which has 1 year's worth of tick data, but your account currency is AUD, and AUDUSD only has 6 months of tick data, then you will only be able to test at 100% for a 6 months range and not for a full year.

Thanks for your reply...

After looking at the journal and after testing several different currency pairs, it's safe to say that the tester is now only able to test from January 2023, despite the fact that I selected a date in June 2022 (roughly 1 year) this alone explains why I am getting the 48% history quality in the backtest tab.

I am using the real tick mode and as I said earlier, up to last week everything was working ok and there were no problems. When I hit ctrl+u, I select a date range from June 2022 to June 2023, however it looks like I am now only receiving historical tick data from January 2023 despite the fact that I haven't changed anything    

So my question is, why is this now suddenly a problem? Last week this wasn't a problem and I was getting 100% history quality tick data on my backtesting with no issues (1 year). I have not changed anything on my computer or the MT5 terminal itself

*UPDATE*

I have just tried to download historical data for several other (new) instruments and regardless of the date range I selected from last year, the tester is now only downloading historical data from January 03 2023, and yet last week I was able to download historical data for all of 2022 as well as 2023

Hope this all makes sense

Thanks for your help 

 
Bollinger68 #:

Thanks for your reply...

After looking at the journal and after testing several different currency pairs, it's safe to say that the tester is now only able to test from January 2023, despite the fact that I selected a date in June 2022 (roughly 1 year) this alone explains why I am getting the 48% history quality in the backtest tab.

I am using the real tick mode and as I said earlier, up to last week everything was working ok and there were no problems. When I hit ctrl+u, I select a date range from June 2022 to June 2023, however it looks like I am now only receiving historical tick data from January 2023 despite the fact that I haven't changed anything    

So my question is, why is this now suddenly a problem? Last week this wasn't a problem and I was getting 100% history quality tick data on my backtesting with no issues (1 year). I have not changed anything on my computer or the MT5 terminal itself

*UPDATE*

I have just tried to download historical data for several other (new) instruments and regardless of the date range I selected from last year, the tester is now only downloading historical data from January 03 2023, and yet last week I was able to download historical data for all of 2022 as well as 2023

Hope this all makes sense

Thanks for your help 

The answer is simple. Your broker has reduced the amount of available tick data. Ask them about it.
12
New comment