MT4 doesn't have long to live - page 5

New comment
 
OnGoing:
Have you even touched MT5 in any way? If not, you have no moral right to talk about dick)
I've been "touching" MT5 for 9 months now. Every day, from morning till night. So I have every right. You should at least look at mql5.com. ))) Personally, I'm already registered there. )))
 
OnGoing: But MT4 is like a fast and graceful gazelle compared to MT5.)

It is a gazelle for now, and only in the tester (and also - for now). In computing, MT4 is doomed because it will be single-threaded: native support for GPU computing is not expected, OpenMP is not in the plans either (however, it is not likely to be in 5 either).

And it will not have less possibilities to implement ANY strategy.

And what about ANY idea? You've obviously overstated it: MT4 arrays alone will not get you far, the structures alone are worth a lot.

I can check absolutely any idea in MT4 in half an hour,

If your ideas are that simple, then perhaps any - yours.

I have ideas that would be very difficult to implement on the quadruple without powerful computing power of a five. Well, of course, there is always a dll, but why make one over the ass when OpenCL is already supported in 5?

You have to scour through all the docs to be able to just make a basic promodify. Then you'll have to spend a couple of hours to make it happen without unnecessary swearing in your log.

These are problems of growing and mastering the language. They will pass quickly.

It's another matter - when will P5 grow to a full-fledged trading platform...

 
MetaDriver:

In fact, the increase in spread is easily simulated (in MT5). There is an option to write your own optimization criterion function, so you can recalculate the profit in it taking into account the extended spread.

The formula itself is simple: Calculated Balance = Balance-(number of trades * increment_to_spread* cost_item): when testing with a fixed lot, you can calculate in one move at the end of the run. With another MM it is a bit more complicated - we will have to make deductions during testing (and output the result at the end), but it is not difficult to solve if we want to.

--

If tester-5 speeds up to four - I'll delete four from my folders. Well, maybe I'll leave the distribution, just in case there's a million-dollar order... :) And even so - in extreme cases it is possible to download it. For very bad (:no time to download:) - let the client provide... ;-))

Wow, how many enthusiastic responses!!!)) And if in a couple of years the five will not go to the people. I will have to restore all MT4 folders again)

 
Mathemat:

It is a gazelle for now, and only in the tester (and also - for now). In computing, MT4 is doomed because it will be single-threaded: native support for GPU computing is not expected, OpenMP is not in the plans either (however, it is not likely to be in 5 either).

You've obviously missed the point about ANY possibilities: MT4 arrays alone won't get you far, the structures alone are hard to beat.

If your ideas are that simple, then perhaps any - yours.

I have ideas that would be very difficult to implement on the quadruple without powerful computing power of a five. Well, of course, there is always a dll, but why go through the hassle when OpenCL is already supported in 5?

These are problems of growth and language adoption. They will pass quickly.

It's another matter - when will P5 grow to a full-fledged trading platform...

Don't make fun of me, Alexey. It seems that MT5 adepts grasp OpenCL as the only reason (and excuse) to become a 5yo geek).

By the way, if your idea involves so many calculations that they lack a single thread, doesn't that make you wary? And if one screw accidentally falls out in combat conditions, wouldn't the whole house of cards fall down?

I always thought the simpler the system, the more reliable...

 
OnGoing:

Wow, how many enthusiastic responses!!!)) What if in a couple of years the five doesn't go to the public. I'll have to restore all MT4 folders again).

It's all right, I'll restore it if you need it. :) But I find it hard to believe...

I remember programming in F4 - I shudder. It's different to each his own, I guess.

 
OnGoing:

Don't be ridiculous, Alexey. It seems that MT5 adepts have grasped OpenCL as the only reason (and justification) to become five-root geeks).

By the way, if your idea involves so many calculations that they lack a single thread, doesn't that make you wary? And if one screw accidentally falls out in combat conditions, wouldn't the whole house of cards fall down?

I always thought the simpler the system, the more reliable...

So TC may be simple, but it is demanding on resources. And the notion of simple and complex is also different for everyone. After all, it is possible that what is very complicated and beyond your comprehension may be very simple for someone else. They take it for granted. )))
 
tol64:
So TC can be simple, but demanding on resources. And the notion of simple and complicated is also different for everyone. After all, it is possible that what is very complicated and incomprehensible to you may be very simple for someone else. They take it for granted. )))

Nah, it's not about understanding at all. It's a matter of technique. It would make sense to complicate it, that's what I mean.

 
OnGoing: Don't be ridiculous, Alexei. I have the impression that MT5 adepts grasp OpenCL as the only reason (and excuse) to become geeks of five)

Not the only one, I've said before. New language structures offer the possibility to write more intelligible code. I haven't even mentioned the cloud yet, which is just manna from heaven for fans of extensive optimizations...

By the way, if your idea involves so many computations that they lack a single thread, doesn't that alert you? And if one screw accidentally falls out in combat conditions, wouldn't the whole house of cards fall down?

On the current system - one thread is enough. But I'm also thinking of the future, when it might break.

Let one cog fall out: the system should not suffer, as it is based on diversification.

I always thought the simpler the system, the more reliable...

It is not obvious, alas. And don't make me laugh: I'm sure you're not happy with the excessive simplicity of rural methods either.

Market wizards and other gurus may tell us that this is the case, but I don't put much faith in it. A simple system is just as easily broken by any innovation, as it is usually easy to figure out.

 
OnGoing:

I'm not suggesting what you want) Tell me, what magical things have you found in MT5 that you can't find in Quartet?

1. The structures without which we couldn't make any headway in 4.

2. The calculation speed is much faster (several times faster).

The indicator buffers are more than enough.

4. a debugger.

5. Objects, with all the goodies (inheritance, encapsulation, virtualization, object arrays)

6. Message exchange with other programs.

7. More or less serious possibility to build graphical interfaces.

8. The tester is really multi-currency. (albeit slow. for now.)

9,10,11,........

Is that enough?

 
OnGoing:

Nah, it's not about understanding at all. It's a matter of technique. It would make sense to complicate things, that's what I mean.

But to make sense, you have to go a certain way. Otherwise, you'll never know if it makes sense or not. Standing is not profitable. )))
123456789101112...83
New comment