MT4 doesn't have long to live

 

These are the pies. See https://www.mql5.com/ru/forum/6337/page4


Renat:

We have already achieved a great speed boost in the development version (it is for internal development) in the Open Price mode and solved the issue of speed difference between MT4 and MT5 testers. The speed in Open Price became comparable to MT4 by abandoning the "fair" generation of intermediate ticks, which were used to accurately control the market environment.

We also did a lot of work on completely rewriting the bar state caches, which gave a significant speedup in multiple passes during optimisation.

In the next 2 builds we will release a new faster tester.

 
What's good about the open-price test, can you tell?)
 
OnGoing:
And what is good in the open price test, can you tell?)

Well at least there is no contact chatter, i.e. indices and oscillators on every tick. Many strategies become more profitable, or at least less unprofitable, if they are converted to open prices, because the number of false signals is dramatically reduced.

But for pipsies this mode is not good enough.

 

Renat:

We have already in the development version (it is for internal development) achieved a huge increase in speed of testing in the Open Price mode and solved the problem of speed difference between MT4 and MT5 testers. The speed in Open Price is now comparable to MT4

Once it was promised that the speed in MT5 was several times faster than in MT4. But it turned out that we were only able to achieve comparable speed... Where is the super speed of MT5?
 
Reshetov:

Well at least there is no contact chatter, i.e. indices and oscillators on every tick. Many strategies become more profitable or at least less unprofitable if translated to opening prices, as the number of false signals is dramatically reduced.

But this mode is not very good for pips.

)) Look at what a stone flower for example turns out on the hourly bar opener.

When we tried to run it on tick history from a Dukas, the grail was immediately deflated)

 
sever31:
The strategy tester is a detriment to the trader.

Yes, if you put it that way, everything else is harmful too. For example, life is a dangerous venereal disease as it is sexually transmitted and always ends in death.

At least MT5 tester lacks some features that led to self-deception of novice traders in MT4 tester, for example:

1. MT5 has forward tests. It means that if on MT4 you cannot tell from the backtests whether it was a forward or a naked fit, on MT5 there is a separator already on the chart (though Photoshop rules and a vertical line can be added in any graphic editor).

2. The inability to increase the balance at the expense of losses by means of closing only profitable counter transactions.

And so on and so forth.

 

Now for another picture. To quote from my own post in the village thread.

Look at how the tester runs through the open bars. I.e. candlestick tails are not considered at all! And they are the ones that are used to take the stops, and the shorter they are the more accurate.

-----------

Opening prices

Some traders do not want to depend on peculiarities of intra-bar modelling and write EAs that trade on formed bars. The fact that the current price bar is fully formed is visible when the next one appears. Thatis what the "Open prices"modelling mode is for such Expert Advisors.


In this mode first bar opening is modelled (Open = High = Low = Close, Volume = 1) that allows expert to identify the end of the previous price bar formation. It is on this bar that testing of the Expert Advisor is started. At the next step the Expert Advisor receives the current fully formed bar,but it is not used for testing!

 
OnGoing:

)) Look at what a stone flower for example turns out by the opener of the hour bars.

When we tried to run it on tick history from Dukas, the grail was immediately deflated)


Only for EAs with explicit bar opening control

Hence the question: have you tested an Expert Advisor (on open and ticks) that has an explicit control of bar opening?

 
zxc:

Once upon a time they promised speed in MT5 many times faster than in MT4. But it turned out that only comparability has been achieved so far... Where is the super speed of MT5?

Outside the tester, applications on mql5 run at least an order of magnitude faster than on mql4.

 
zxc:


Only for Expert Advisors with explicit control of bar opening

Hence the question: have you tested an EA (by opener and by ticks) which has explicit control over bar opening?

Yes, when to open the positions makes absolutely no difference. The tester passes dazedly to the next open price and does not care at all about how the price moved inside the candle or outside the two open prices.

 
Reshetov:

Outside the tester, mql5 applications run at least an order of magnitude faster than mql4.


This is good of course. But outside the tester few people care about the speed, even on weak PCs. It is precisely the speed that is important in the tester.
Reason: