Trade initiated on Live account is not also initiated on Strategy tester

 

Hi,

While reviewing the trades executed last week I've noticed that (apparently) one of my EAs initiated a trade on the Live account (at hours 3.00 on 12.07.18)  but it didn't also re-initiated it on the Strategy tester (that I ran today 15.07.18).

It seems that one of the criteria was met on the Live account ( MFI[1]<25 ) but not also on the Strategy tester ("Every tick based on real ticks"), where the value for this parameter was quite different ( MFI[1] = 29.76 ).

MFI[1] it's Money Flow Index (iMFI) for the last bar. MFI is computed for the last 12 periods.

It's like the indicator was computed in Live account on certain price and volume info, which are different (?) from the ones available later on the Strategy tester. Still, history quality for period 11-12.07.2018 is 100%..

Can anyone please tell me whether they encountered a similar situation ?

Is it possible that the data from Live is different from data available in Strategy tester ? ? ?

Thank you,

Florin

PS: the other option is that the broker initiated a deal "for" me.. but, so far, I would exclude this option

 

I suppose you are sure you used the same parameters ?

Which broker/symbol it was ?

MFI[1] is the value of the last closed candle ?

MFI second parameter is volume, did you used "Tick" or "Real" ?

 
Alain Verleyen:

I suppose you are sure you used the same parameters ?

Which broker/symbol it was ?

MFI[1] is the value of the last closed candle ?

MFI second parameter is volume, did you used "Tick" or "Real" ?

Yes, I use the same parameters
Broker - Admiral Markets 
Symbol - USDCAD
Time-frame - M12
MFI [1] is for the last closed candle (that as at 02.48 in this case) 
Tick is the second parameter 

Please advise. 
 
florinm77:
Yes, I use the same parameters. 
Broker - Admiral Markets 
Symbol - USDCAD
Time-frame - M12
MFI [1] is for the last closed candle (that as at 02.48 in this case) 
Tick is the second parameter 

Please advise. 

You are right it should be the same in theory. So I can only advice you to reproduce it, output all values in the log, run it on a live chart. Then compare with a Strategy Tester run.

Once done, either you will find why it's different, or you will find an MT5 bug and you can report it to Metaquotes.

 

Also check the journal for history integrity.

For example:

which *could* explain the difference.

[edit] The backtest report states 100% data quality, while in reality it isn't [/edit]
 
Enrique Dangeroux:

Also check the journal for history integrity.

For example:

which *could* explain the difference.

[edit] The backtest report states 100% data quality, while in reality it isn't [/edit]

I have checked the journal as advised but I've found no integrity problem :(

Thank you for your suggestion, anyway - I will definitely consider it in the future.

 
Alain Verleyen:

You are right it should be the same in theory. So I can only advice you to reproduce it, output all values in the log, run it on a live chart. Then compare with a Strategy Tester run.

Once done, either you will find why it's different, or you will find an MT5 bug and you can report it to Metaquotes.

Well, I'm afraid I do not really understand..

I can run the EA in the Strategy Tester but - with what I should compare this ?

What do you mean by "run it on a live chart" ?

 
florinm77:

Well, I'm afraid I do not really understand..

I can run the EA in the Strategy Tester but - with what I should compare this ?

What do you mean by "run it on a live chart" ?

There are 2 possibilities :

1° It's an issue on your side because you missed something.

2° It's an MT5 bug.

If you want to know which it is, you need to check it, because if you report a bug to MT5 without any proof they will probably don't take care.

 
Alain Verleyen:

There are 2 possibilities :

1° It's an issue on your side because you missed something.

2° It's an MT5 bug.

If you want to know which it is, you need to check it, because if you report a bug to MT5 without any proof they will probably don't take care.

Thank you. 
I will see what it can be done. 

One way to deal with this, I think, would be to send an email at the deal time that contains all the relevant data (values for parameters). But, unfortunately, so far, SendMail doesn't work on my EAs, though it does when I test it on the Metatrader menu. So I will open another thread with this issue. 

 
florinm77:
Thank you. 
I will see what it can be done. 

One way to deal with this, I think, would be to send an email at the deal time that contains all the relevant data (values for parameters). But, unfortunately, so far, SendMail doesn't work on my EAs, though it does when I test it on the Metatrader menu. So I will open another thread with this issue. 

Why an email ? Use Print statement to output values in the log.
 
Alain Verleyen:
Why an email ? Use Print statement to output values in the log.
Indeed I think this would be the fastest solution for troubleshooting this. Thank you very much for your suggestion!