Interesting topic for many: what's new in MetaTrader 4 and MQL4 - big changes on the way - page 47

 
It's simple. We test at opening prices, then we fine-tune the Expert Advisor to work at opening prices. That is the whole secret. It is not correct to make a pipswitch that trades based on ticks and then blame the tester, because it is clearly stated in the documentation that the ticks are modelled. So in this case the divergence is inevitable. As for the long term, it is not necessary at all, you can test on opening prices even on minutes.
Документация по MQL5: Стандартные константы, перечисления и структуры / Константы индикаторов / Ценовые константы
Документация по MQL5: Стандартные константы, перечисления и структуры / Константы индикаторов / Ценовые константы
  • www.mql5.com
Стандартные константы, перечисления и структуры / Константы индикаторов / Ценовые константы - Документация по MQL5
 
sergeev:

It's not stupidity that sits here, but those who consider everything only under specific implementations and code.

Give algorithms, not empty words!

Code and pictures, code and pictures!

Algorithms give (programming language is not at the discretion of the coder - anything), stop using the richness of Russian language, it is not for analytics,

it's time to move on to specifics, formulas and code!

MetaDriver went over, showed everything clearly. And what has changed? Did anyone see the problem?

papaklass:

For some reason I thought there would be no example.

Why would I need someone's insight when there is a LIVE and FULL source of the idea?

Examples cut away different interpretations of ideas. Examples show the essence of the idea that the author puts into it. It is from the original source that the example is important, not its derivation.

No offence. This is just my opinion.

Is that what for? "It's not stupidity that sits here" - and continues to sit.
 
FoxRex:
It's simple. We test at opening prices, then we fine-tune the Expert Advisor to work at opening prices. That is the whole secret. It is not correct to make a pipswitch that trades based on ticks and then blame the tester, because it is clearly stated in the documentation that the ticks are modelled. So in this case the divergence is inevitable. As for the long term, it is not necessary at all, you may test on opening prices even on minutes.
https://www.mql5.com/ru/forum/146025/page49#827675
Что будет нового в MetaTrader 4 и MQL4 - большие изменения на подходе - MQL4 форум
  • www.mql5.com
Что будет нового в MetaTrader 4 и MQL4 - большие изменения на подходе - MQL4 форум
 
papaklass:

For some reason, I didn't think there would be an example.

Why should I understand someone when there is a LIVE and in FULL HEALTH of the original source of the idea?

Examples cut out different interpretations of ideas. Examples show the essence of the idea that the author puts into it. It is from the original source that the example is important, not from its derivation.

No offence. It's just my opinion.

https://www.mql5.com/ru/code/10468

https://www.mql5.com/ru/code/9234

IND_Potential - MQL4 Code Base
  • www.mql5.com
IND_Potential - MQL4 Code Base: технические индикаторы для МТ4
 
hrenfx:
This is not correct, because the moment of Low_Bid appearance does not coincide with the moment of Low_Ask appearance. But we can write another value in the Spread field: NewSpread = Low_Ask - Low_Bid. Then, indeed, we will only need changes in the algorithm for calculating the spread.
Well, then you can strain not MQ, but the quotes providers)
 
Avals:
In this case we can stimulate not MQ, but the quotes providers).

As for the old story, of course only the broker can rewrite it on his own initiative, but the algorithm for writing a new one is probably in the hands of MQ, because the MT-server writes it.

--

The greatest benefit (informativeness) from the field "Spread" in quotes will be achieved when writing in this field the difference between LowBid and LowAsk.

This is the best way to make the historical base "environmentally friendly" at the moment (with the current format of quotes).

There will be almost no statistical difference between the average spread and the one calculated using this formula but the tester will not be slippery when trading with limit orders.

Considering that all basic orders in the market are limit orders, this method of spread coding in quotes is equally correct for all order types.

 
MetaDriver:

As for the old story, of course only the broker can rewrite it on his own initiative, but the algorithm for writing a new one is probably in the hands of MQ, because the MT-server writes it.


I think it should be configurable, if they talk about the flexibility of the server part)
 
MetaDriver:

I don't think so. As for the old story, here of course only the broker can rewrite on his own initiative, but the algorithm for writing a new one is most likely in the hands of MQ, because the MT-server writes it.

Some ECN/STP venues are aware of this initiative and are directly interested in hosting it while adding MT5 trading capability.

With history, of course, it will work out. If the real-time will not allow for such an implementation, then simply introduce another informative symbol, on which a fresh and optimal for MT5-tester history will be added at a certain interval.

Of course, we will not forget about a separate real-time Ask-symbol for fighting robots - as in MT4.

P.S. Third-party developers have to tweak the platform's capabilities - give algotraders a way around its crutches.

 

Half-measures are everything. We need a proper teak tester

p.s. if it's going to be common and in the chips, that's good)

 

What is meant by a "normal teak tester"? And what is it for?

MT4 has a working tick tester as there is the ability to create a custom story. So what?

In JForex it's the same. But there is also a millisecond timestamp + volumes on bestbands. And again, so what?

The exact staff MT5 tester with an option to investigate in the cloud would be very useful for algotraders. Half-measures are useful only within the MT5 framework.

You could have created your own MQL5-tester (by ticks) using mathematical calculations for the cloud long ago. But who would really need it?