What's new in MetaTrader 4 and MQL4 - big changes on the way - page 63

 
borilunad:
The only difference is that in the tester it is impossible to run even on real quotes, only on past ones! And on demo and real, you can't check on past ones to compare with the result in the tester! Invent a time machine, that's when you can compare! In the meantime...
Compare is not even a problem, you run the test in real-time, the next day on the same area in the tester, compare the results.
 

Renat, why am I rooting for ticks? You wrote an excellent article about tick generation in the tester (opened our eyes, so to speak), and the ticks from MQ really fit well into the scheme described in the article, but the ticks of other dealers do not, and that's the problem.

It should be tested on the data with which you will work. At least there should be a test mode (even if not deep) which exactly repeats the result of realtime. So to speak the final run in the tester TC before putting it on realtime.

ZZY Let it be a month (yes I understand that over time demands will grow, but for the start is normal), but it will be an accurate test reject all huskies (including tester graals).

ZZZY Shortly while the sticking point in the political decision, it will be, will be given opportunities, and with opportunities will grow and demands, and users will be more than 2.5 people per week.

 
serferrer:

https://www.mql5.com/ru/forum/1031/page18#comment_520781

743*35 = 26 gigabytes for 35 symbols, actually it will be even less, because EURUSD is the most liquid instrument (it has the most ticks), these are real ticks from April of 2007 and some symbols appeared later than 2007.

+ If MetaTrader 5 is a terminal for exchanges, why is it unable to perform accurate tick testing?

MetaDriver:
It's a bad start. If the platform supports testing on real ticks, demand will jump. Right now, not all users can use ticks for testing.


It's OK. In the terminal settings, create a variable (like MaxTickHistory), the default setting is 0. Not everyone needs ticks, here's an opinion for example:

borilunad:
У меня информация из моего опыта! Когда строю ТС, исходя из котировок на ТФ от М5 до Н1 всё в порядке! Но когда пробую только прибавить данные с М1, то всё летит к чёрту! Представляю, что дадут тики! Лишний шум, не больше! Делайте ТС толстокожей, что всё ей было до лампочки!

He does not need ticks. But I need them and I will adjust settings to needed depth.

For my trading I need ticks to the depth of a couple of days, for testing - a couple of months + days, but for research of market dynamics I would download 6-8 years for 20-30 symbols. Capacities of disk memory allow (two 3 terabytes + 1 terabyte hard drives).

So let's think. If declaration of intent to introduce a long-awaited innovation will appear - people will be sympathetic to all the problems that arise (for example, download speed limitation). There may even be useful suggestions for implementation. For example, there is an idea to distribute traffic among users, if user terminals are united in a controlled "peer to peer" network which sends "fellow traders" missing pieces of tick history. In this case, the broker will actually provide the network with only new ticks + initial filling of the history base for a small percentage of users.

All in all it is solvable. In today's Internet conditions the problem should not horrify so much the developers as, say, 6-7 years ago. Speeds are great, the amount of disk memory is available to both brokers and "tick fans".





C-4:

You must remember that ticks are remarkably compressible. Standard zip archivers compress the tick history of one active RTS day from 8 mb to 350 kBytes. And yet it is a plain unoptimized text file. With your own highly specialised archiver, the result can still be improved. So in the worst case (about 200,000 ticks per day) it is 100mByte per year (350kByte * 270 days). Liquid tools are 30 - 40, which gives us 3-4 GBytes per year or only 30-40 GBytes in 10 years - not much by today's standards.

Urain:

The key word here is "optional".

No need to break what you have, add options.

First, custom ticks instead of those generated in the tester,

furthermore, save ticks to history,

with growth of request, broadcasting of tick history from the dealing server, etc.

ZS and each item is optional, you can even send statistics on how the options are used to keep you informed.




++++ PCA. Already in this thread it is clear that the demand is far from 2.5 people. Moreover, in terms of developing the platform, this is not a minus, but a big plus. It is important for professionals, even if there are not many of them, but they will lead thousands of ordinary users. After all, the opinion of authoritative participants guides the rest of the majority. And, for example, why isn't MT4 quoted in the exchange trading environment? - Because of the lack of a transparent tick history - like a toy for suckers, in which everything is not transparent (I think the MT5, in that environment, is promoted with difficulty - the previously formed opinion affects it).

I don't understand the company's policy - to make requests in spite of real requests, instead of following the trend. Who will be ill served by it? But why such an important (and simple) request - so many years! - hasn't been solved by Metacquotes? It is not clear to me. Do you want to wait until other companies create third-party crutches with the same functionality, and then ruin their service accusing them of stealing a market share from you?

There's a lot I don't understand...

 
MetaDriver:
It is a bad start. If the platform will support testing on real ticks - demand will jump. Not all users can use ticks for testing now.
...

So let's think about it. If a declaration of intent to implement a long-awaited innovation comes along, people will be sympathetic to any problems that may arise (e.g. download speed limits). There may even be useful suggestions for implementation. For example, there is an idea to distribute traffic among users, if user terminals are united in a controlled "peer to peer" network which sends "fellow traders" missing pieces of tick history. In this case, the broker will actually provide the network only with new ticks + initial filling of the history base for a small percentage of users.

All in all, everything can be solved. In present Internet conditions the task should not horrify developers as much as, say, 6-7 years ago. Speeds are great, disk storage capacities are quite affordable for both brokers and "tick fans".

Ah yes how come I didn't read such a valuable post :)

Torrent technology perfected long time ago, so there will be no problem with distribution of tick history, the more so since MT is not a torrent, there sitters and peers will be a few orders of magnitude higher.

 
They've got their tics all over the thread, it's hard to find a post on the subject. Who needs them at all, these tics, except a few maniacs. And they don't need them either. :)
 
Eufemio:
They got sick of their ticks, they pissed off the whole branch, you have to strain your eyes to find a post on the subject. Who needs them anyway, these tics, except a few maniacs. And they don't need them either. :)

Full moon walks on the planet :)

I'm not sure what to do with them, they're all down to H1 and I turn on the terminal once a week, what's the point of the terminal, you can get quotes by mail, e-mail is fast, you can use it to give orders" :)

ZZZY and orders can also be sent to the post pigeons :)))), and what's all, go once a quarter, bought from dealing with a dozen or two pigeons, need to send orders, papierka rolled a tube, a foot tied to his will, he still in the dealing to the female in 3 days will fly, that order and my deliveries. But the system is indestructible, no government will not interfere with the Internet or electricity. I am sick and tired of modern technology, my head is spinning :))))

 
Urain:

Full moon walks the planet :)

...


If you're not too far away from dealing and you're out of pigeons, you can light a fire. The smoke can be made red, which means a signal to sell, and blue, respectively, a signal to buy. It is very convenient. )))
 
Urain:

A full moon walks the planet :)

It all depends on the depth of the dive, ...

Well, no matter how deep you dive - and whether for mechanical or manual trading timeframe less than 5 minutes is white noise (or pink, for ladies). In this case the timeframe is a white noise (or pink, for ladies).
 
Eufemio:
No matter how deep you go - timeframe less than 5 minutes for mechanical or manual trading is white noise (or pink, for ladies). In this case, we may use a lot of different instruments, and the price of these instruments will be different.

Have you checked?

There's red noise. It is predicted normally.

 
Eufemio:
But even if you dive deeper, the timeframe less than 5 minutes for mechanical or manual trading is white noise (or pink, for ladies). If we look at the timeframe, we will see, that it is a white noise (or pink for ladies).

And I have raised a couple of grand on this white noise during a week, I only wish I had done it on a demo.

If I've tried to open positions I have often had to catch bugs in codes, I had to open some small or big losses.