Is Metatrader Strategy Tester really that useless? - page 2

 
Jose Francisco Casado Fernandez:

Have you tried using the "every tick" mode in the Strategy Tester ?? If the answer is "Yes", then change your broker or import your own historical data from external sources. For example in Alpari I get a quality between 89% and 100% depending on the selected period:




It also depends on the Symbol.  Regards.

Thank you. 

No, I haven't, but I will. By the way, how can I use Alpari data? Would you please give me a link or something like that?
 
josephcom:
Thank you. 

No, I haven't, but I will. By the way, how can I use Alpari data? Would you please give me a link or something like that?
Open a demo account with them.  Regards.
 
Jose Francisco Casado Fernandez:
Open a demo account with them.  Regards.
Thank you Jose Francisco Casado Fernandez. 

... and to all the members who seriously care about their analysis quality: I'll let you know what I do and whatever the results.


 
Jose Francisco Casado Fernandez:
Open a demo account with them.  Regards.
By the way, I've already checked the chart candles ONE BY ONE for the last two months to see if any one of them is missing, and the answer is NO.

So, the question is why history quality is still so poor? What exactly a "poor history quality" mean?
 
josephcom:
I've already found the solution?! What is it? My History Quality is still 29% and you're telling me I have the solution? No, I don't.

... and don't take me wrong. I am still very fond of Metatrader and I really like it. I am talking about its Strategy Tester only...

I was intentionally unclear because experience show that posts about this issue get censored. Read your post scriptum again.

 
alphatrading:

I was intentionally unclear because experience show that posts about this issue get censored. Read your post scriptum again.

I really have no clue what you're telling me (and I guess no one else does). So, would you not post anything at all please? You're just making this topic complicated.

Thank you
 
alphatrading: I was intentionally unclear because experience show that posts about this issue get censored. Read your post scriptum again.

For clarification, the OP's post scriptum is and I quote:

P.S. I am using Metatrader "5" (and not "4") in which there is no "History Center"...

However, in MetaTrader there is a "History Center", just not by that name, and @Jose Francisco Casado Fernandez has already explained how to go to the "Symbols" and then download the "History" for Bars and Ticks.

So for the OP's sake, I ask what exactly do you mean? Please be more specific! I don't see what it is that would cause "censorship", except for discussing brokers.

Personally, I get 100% in my case no matter what the test type - be it Open Prices, M1 Bars, Tick Date or Real Tick Data. But I am curious as to why the OP would get 29%/30%.

josephcom: By the way, I've already checked the chart candles ONE BY ONE for the last two months to see if any one of them is missing, and the answer is NO. So, the question is why history quality is still so poor? What exactly a "poor history quality" mean?

@josephcom, personally, I get 100% in my case no matter what the test type (Open Prices, M1 Bars, Tick Date or Real Tick Data), but I do have a suggestion in case you have not tried it yet.

Just in case the data files are corrupt in some way, have your tried, closing MetaTrader and deleting all the history/tick data files so that it will start with a "clean slate"?

The history and tick data is stored in the following locations:

  • Bases/<name of broker and account type>/history/
  • Bases/<name of broker and account type>/ticks/
  • Tester/bases/<name of broker and account type>/history/
  • Tester/bases/<name of broker and account type>/ticks/

In each of the above paths there will be folders for each symbol. Just delete all the content within the above folders (but make a backup) and then restart MetaTrader.

I don't know if it will solve the problem or not, but it is a suggestion just in case the data is corrupt.

 
Fernando Carreiro:

For clarification, the OP's post scriptum is and I quote:

However, in MetaTrader there is a "History Center", just not by that name, and @Jose Francisco Casado Fernandez has already explained how to go to the "Symbols" and then download the "History" for Bars and Ticks.

So for the OP's sake, I ask what exactly do you mean? Please be more specific! I don't see what it is that would cause "censorship", except for discussing brokers.

Personally, I get 100% in my case no matter what the test type - be it Open Prices, M1 Bars, Tick Date or Real Tick Data. But I am curious as to why the OP would get 29%/30%.

@josephcom, personally, I get 100% in my case no matter what the test type (Open Prices, M1 Bars, Tick Date or Real Tick Data), but I do have a suggestion in case you have not tried it yet.

Just in case the data files are corrupt in some way, have your tried, closing MetaTrader and deleting all the history/tick data files so that it will start with a "clean slate"?

The history and tick data is stored in the following locations:

  • Bases/<name of broker and account type>/history/
  • Bases/<name of broker and account type>/ticks/
  • Tester/bases/<name of broker and account type>/history/
  • Tester/bases/<name of broker and account type>/ticks/

In each of the above paths there will be folders for each symbol. Just delete all the content within the above folders (but make a backup) and then restart MetaTrader.

I don't know if it will solve the problem or not, but it is a suggestion just in case the data is corrupt.

Thank you Fernando, I appreciate your reply.

Yes, I did. I tried what you just suggested and the result was still the same.

I am still seriously working on the issue and I'll tell you about anything new I possibly discover.

So far, I've found out that "History Quality" is not just about a candlestick's missing; but ALSO about incorrect spreads and volumes.

This may be a proof that this problem has nothing to do with Metatrader or its Strategy Tester, but with brokers. 


 
josephcom:
Thank you Fernando, I appreciate your reply.

Yes, I did. I tried what you just suggested and the result was still the same.

I am still seriously working on the issue and I'll tell you about anything new I possibly discover.

So far, I've found out that "History Quality" is not just about a candlestick's missing; but ALSO about incorrect spreads and volumes.

This may be a proof that this problem has nothing to do with Metatrader or its Strategy Tester, but with brokers.

One more thing - have you checked the Journal log to see if there are any message that could explain the problem?

The following is from normal 100% tests. Maybe your journal logs have some extra messages that could explain the problem ...

2017.11.27 11:31:05.733 Core 1  EURUSD.m,M15: 55725 ticks, 14352 bars generated. Environment synchronized in 0:00:00.515. Test passed in 0:00:00.438 (including ticks preprocessing 0:00:00.062).
2017.11.27 11:31:05.733 Core 1  EURUSD.m,M15: total time from login to stop testing 0:00:00.953 (including 0:00:00.515 for history data synchronization)
2017.11.27 11:31:05.733 Core 1  55801 total ticks for all symbols
2017.11.27 11:31:05.733 Core 1  EURUSD.m: generate 28987 ticks in 0:00:00.062, passed to tester 55801 ticks

2017.11.27 11:56:32.542 Core 1  EURUSD.m,M15: 17637142 ticks, 14352 bars generated. Environment synchronized in 0:00:01.953. Test passed in 0:00:15.406 (including ticks preprocessing 0:00:02.532).
2017.11.27 11:56:32.542 Core 1  EURUSD.m,M15: total time from login to stop testing 0:00:17.359 (including 0:00:01.953 for history data synchronization)
2017.11.27 11:56:32.542 Core 1  556 Mb memory used including 2 Mb of history data, 320 Mb of cached tick data (total memory for tick data 383 Mb)
 
Fernando Carreiro:

One more thing - have you checked the Journal log to see if there are any message that could explain the problem?

The following is from normal 100% tests. Maybe your journal logs have some extra messages that could explain the problem ...

That sounds a good idea. I'll check this out and let you know the outcome in a few minutes...