If MetaTrader 6 comes out tomorrow - page 13

New comment
 
borilunad:
What about the daily report sent to email, means nothing, has no weight, in which the broker offers to challenge the trader within 24 hours? If you agree by default, it takes effect! Or is it a "paperwork"?

For an EA that trades (changes position) once a day or less is enough. And if the average trading time is 5-30 minutes, then there is a lot to mess around with during a day.

Cancellation of a trade causes all subsequent positions on the instrument to be shifted by the volume of the cancelled trade (only in the opposite direction).

 
Provide for all types of pending orders. You don't need 6 for that. This was needed in 3. Stop Close Only for example.
 
MetaDriver:

For an Expert Advisor that trades (changes position) once a day or less, it is enough. And if the average trading time is 5-30 minutes, then a lot can be messed up during a day.

Cancellation of one trade shifts all subsequent positions on a symbol by the volume of the cancelled trade (only in the opposite direction).

Not to work with just one position!
 
borilunad:
You can't work with just one position!

What are you talking about? Are there any alternatives? You don't need to invite me to the foursome, I was and will be there. I'm talking about netting and the MT5 order-dealing system.

When trading, it is the aggregate position that matters, the rest (trades, orders) are unimportant.

My EAs support the position recommended (the forecasted part) for a symbol by placing the orders necessary for this. And they place them correctly. If one deal is removed from the series of these transactions retroactively, the entire subsequent series will be distorted. This does not happen in real time - if any order does not trigger or only partially triggers, a new attempt is made until the difference between the recommended and actual position is zeroed out.

 
MetaDriver:

What are you talking about? Are there any alternatives? You don't need to invite me to the foursome, I have been there and will continue to be there. I am talking about netting and the MT5 order-dealing system.

When trading, it is the aggregate position that matters, the rest (trades, orders) are unimportant.

My EAs support the position recommended (the forecasted part) for a symbol by placing the orders necessary for this. And they place them correctly. If one deal is removed from the series of these transactions retroactively, the entire subsequent series will be distorted. This does not happen in real time - if an order does not trigger or only partially triggers, we try again until the difference between the recommended and actual positions is zeroed.

Thanks for the clarification! I'm looking at netting for now. Of course, there is more clarity, but there is more responsibility for each step, especially in Forex market conditions! I am not attracted to other markets. Good luck to you!

By the way, netting is more convenient even for lovers of lots! By opening the opposite position when the direction changes, the position will be reduced, and the losses and the deposit will not grow, like on the 4! Thus, not only will you not miss, but you will enter the opposite trend earlier!

 

Thought this.... maybe we should do a bit of "degradation" with the tester? :)

Or rather, to introduce mode, that turns the tester in a simple "calculator", not much different from homemade toys - all statistical information calculation is disabled (profitability, expectations and all the rest), optimizing graphing is disabled and in general everything that can brake at least something testing and optimization. Something like: "Calculating statistics on/off". The tester's task in this mode is to provide properly ticks for the requested symbols and that's all, the rest is up to the user, he will calculate everything he needs.

Of course, optimization in this mode is possible when the "Custom max" optimization criterion is chosen, all the rest are inaccessible.

After testing in this mode, the report becomes inaccessible, or as an option, to display in a report the user's statistic data, but this requires an additional type of variables, something like doubleOpt, and in the report displays these variables, in which the user will collect and add the necessary statistic information.

In my humble opinion, this mode will speed up the work of tester/optimizer many times, and possibly dozens of times.

 
MetaDriver:
What was it for?
 
joo:

In my humble opinion, this mode will speed up the tester/optimiser many times over, if not tens of times over.

There are suspicions that the main brake of MT5 tester is a model of working with data. Namely the data is copied from one buffer to another using special functions like CopyRates. Expanding data into memory is a very, very slow operation that crunches the entire performance of the system. There's no sense in omitting the MO calculation in the hope to get a multiple performance gain if each for loop transfers an array from one memory area to another.
 
TheXpert:
What was he banned for?
+1, good question.
 
Dark/night mode for the terminal borders, DOM, etc. Seems to be the only platform that still doesn't have one and the white causes burn in on my screens.
1...678910111213
New comment