MT4-Tester VS MT5-Tester - page 10

New comment
 
Alexander Laur:

Tester 4 is a counting tester.

Tester 5 is part of the intelligent system.

The testers are compared that are converters of the initial price series into equity series. I.e. two functions are compared
double MT4Tester( const double &Prices[], double &Equity[], void* TradeSystem );
double MT5Tester( const double &Prices[], double &Equity[], void* TradeSystem );

As the branch has shown, both functions return the same result (with the same input values). So it was decided to compare these functions in terms of speed performance. And also it was noticed that the first argument (const double &Prices[]) in MT5Tester is hardcoded and there is no such a limitation in another similar function.

 
Alexander Laur:


OK.

The 4- tester is a calculator.

The tester 5 is part of an intelligent system.

It is NOT CORRECT from the point of view of pure experimentation to rip one of its components out of the system and compare it with a specially sharpened calculator.

Is that clearer?


And what kind of intelligent system? Just curious.
 
Alexander Laur:


If I'm not mistaken, the opening of the data format somehow helped the bad guys to break into the data transfer protocol. All sorts of features emerged that allowed third-party firms, without the MK's permission, to make money from their software. This was particularly hotly debated when MK introduced its Signals service.

Notice I answered your request, you didn't answer mine.

You are confused. These problems were due to the open API to the terminal. This allowed third-party firms to build their own, sometimes very good applications based on this API. Naturally, the developers didn't like it and it was quickly shut down.

 
Alexander Laur:


"The testers being compared are the converters of the original price series into a series of equities. That is, two functions are being compared"

This is exactly the catch phrase. In order to compare something, it has to be prepared. A 5 tester does a lot more preparatory work because it is part of a SYSTEM, while a 4 tester is a simple calculator

The time minus accounting for preparatory work was compared.
 

Forum on trading, automated trading systems and strategy tester

MT4-Tester VS MT5-Tester

fxsaber, 2017.05.10 17:40

The testers being compared are the converters of the original price series into an equity series. I.e. two functions are compared
double MT4Tester( const double &Prices[], double &Equity[], void* TradeSystem );
double MT5Tester( const double &Prices[], double &Equity[], void* TradeSystem );

As the branch showed, both functions return the same result (with the same input values). So it was decided to compare these functions in terms of speed performance. I also noticed that the first argument (const double &Prices[]) is rigidly specified in MT5Tester and there is no such limitation in another similar function.


There is also a third argument (void* TradeSystem). The set of all TS that are suitable for MT4Tester is a subset of all TS that are suitable for MT5Tester. This is the advantage of the second function. So the advantages and disadvantages are stated objectively. All results are reproducible, which is actually the highlight of this comparison.


It is interesting that if the result 1598 had not been displayed, defenders (it is not clear from whom) of MT5 would continue claiming that MT5 loses so much, because it is not a calculator. However, the developers have tweaked and greatly improved the performance. There was plenty to tweak. And they will tweak more. There is a lot to compare it with.

 
fxsaber:

There is also a third argument (void* TradeSystem). The set of all TS, which are suitable for MT4Tester, is a subset of all TS, which are suitable for MT5Tester. This is the advantage of the second function. So the advantages and disadvantages are stated objectively. All results are reproducible, which is actually the highlight of this comparison.


It is interesting that if the result 1598 had not been displayed, defenders (it is not clear from whom) of MT5 would continue claiming that MT5 loses so much, because it is not a calculator. However, the developers have tweaked and significantly improved the performance. There was plenty to tweak. And they will tweak more. There is a lot to compare it with.

Useful topic. It's a good job.

Respect

 
Alexander Laur:

Compare the single passes of a MULTIVATE Expert Advisor in MT4 and MT5 testers.

Forum on trading, automated trading systems and strategy testing

MT4-Tester VS MT5-Tester

fxsaber, 2017.05.10 19:11

There is also a third argument (void* TradeSystem). The set of all the TCs that are suitable for MT4Tester is a subset of all the TCs that are suitable for MT5Tester. This is the advantage of the second function. So the advantages and disadvantages are stated objectively. All results are reproducible, that is actually the highlight of this comparison.
 
Alexander Laur:
History trimming refers to the preparation of the data used for testing, but not to algorithmic optimisation. Algorithmic optimisation optimises the algorithm that processes the data.

This is the same algorithmic optimisation as calculating the signal once per bar, unless indicator readings from bar 0 are used. It is the most accessible and the most effective.

There is no point in calculating not only the code of the Expert Advisor and indicators, but the entire trading environment (equity, margin requirements, quotes for other instruments) on every tick. And castrum history allows us to do this.

Of course, things get a bit more complicated when multicurrency is added. Perhaps that is why it takes so long to release these "services".

 

MT5, even in mono currency mode, is able to run an TS which MT4 cannot, as MT5 has a timer. To put it very crudely, MT4 is a for loop by ticks, MT5 is a for loop by milliseconds.

But all these facts don't negate the previously listed.

 
Alexander Laur:

I did not suggest you to run a multi-currency Expert Advisor for nothing.

Forum on trading, automated trading systems and strategy testing

MT4-Tester VS MT5-Tester

fxsaber, 2017.05.10 17:40

The testers being compared are the transducer of the original price series into an equity series. I.e. two functions are compared
double MT4Tester( const double &Prices[], double &Equity[], void* TradeSystem );
double MT5Tester( const double &Prices[], double &Equity[], void* TradeSystem );

As the branch showed, both functions return the same result(with the same input values). So it was decided to compare these functions in terms of speed performance. I also noticed that the first argument (const double &Prices[]) is rigidly given in MT5Tester and there is no such limitation in another similar function.

Forum on trading, automated trading systems and strategy testing

MT4-Tester VS MT5-Tester

fxsaber, 2017.05.10 21:56

MT5 even in monocurrency mode is able to run a TS that MT4 cannot, as MT5 has a timer. To put it very crudely, MT4 is a for loop by ticks, MT5 is a for loop by milliseconds.

But all these facts don't negate the previously listed.

Forum on trading, automated trading systems and strategy testing

MT4-Tester VS MT5-Tester

fxsaber, 2017.05.10 19:11


There is also a third argument (void* TradeSystem). The set of all the TS that are suitable for MT4Tester is a subset of all the TS that are suitable for MT5Tester. This is the advantage of the second function. So the advantages and disadvantages are stated objectively. All results are reproducible, that is actually the highlight of this comparison

1...34567891011
New comment