Miracles with the tester. - page 6
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Rather the wonders of real execution. In the tester, the Expert Advisor and the trade server (test) are one-to-one. In the tester there is only one EA (no interfering neighbours) and all executions are synchronous.
In the real life your Expert Advisor is on the same computer with 4 dozens of other Expert Advisors, which also work like yours. In real life, the trade server is working not only with your EA, but also with thousands of others. In real life, the execution is asynchronous.
Have you run a test of your EA in mode "random execution delay"?
In the "random delayed execution" mode the result of market entry is almost the same, i.e. only two buy trades. My profit is about 4000 less.
I'll look at the indicators now ( .
Do you only close positions on stops or are there any other forced closing mechanisms?
The tester did show all the deals that were really in the championship, but for some reason only today?
Well, if the tester is so perfect in the five, there might be a chance of success )).
The tester has not been testing for the last 24 hours.
Renat, why does the tester in MT4 test the last day and not in MT5? Are there such big differences in the essence of ticking?
By the way, yes, we sorted out the optimization - cloud, desynchronization, etc. But all the same there are not clearness, I think it is possible to limit the end point (to specify the exact time). But never mind, not very important.
But a single test with specific parameters on my computer.....
To compare the actual trade on a real / demo and in the tester.
?
Renat, why does the tester in MT4 test the last day, and not in MT5? Are there such big differences in the essence of ticking?
In MT4, the tester is "built into" the terminal and directly uses the history. The optimizer can easily "freeze" the history at the moment when the optimization starts.
In MT5, the tester is a separate application that synchronises the history with the terminal every time. The last day's data is constantly changing, so:
1. Constant pumping of the last day's minute container.
2. Different history data on different agents, depending on when an agent is connected
In MT4, the tester is "embedded" into the terminal and directly uses the history. The optimizer can easily "freeze" the history at the moment when optimization starts.
In MT5, the tester is a separate application that synchronises the history with the terminal every time. The last day's data is constantly changing, so:
1. Constant pumping of the last day's minute container.
2. Different history data on different agents depending on when the agent is connected
And if we limit optimization to a daily batch, but allow testing (single) of the last day? It is a very popular feature, especially when the market quickly (relatively) changes. The point is that it is important to verify whether the parameter implementation selected in the optimizer is really able to successfully trade in the current time.
In MT4, the tester is "embedded" into the terminal and directly uses the history. The optimizer can easily "freeze" the history at the moment when optimization starts.
In MT5, the tester is a separate application that synchronises the history with the terminal every time. The last day's data is constantly changing, so:
1. Constant pumping of the last day's minute container.
2. Different history data on different agents depending on when the agent is connected
this is ok. paging, data are all sources of ticks.
But the question is why don't we allow to check the tests on the ticks of the current day? Why is there no such possibility?
but the question is - why not let the tests be checked on current day ticks? Why is there no such possibility?
Because there is a problem with the time shift and the sync point.
To avoid having problems with data in cludes (where there are huge distributed caches) and remote agents, we specifically limit it to the previous day. A day (daily block) is the minimum portion of synchronization in our network and in a distributed network it is economically unprofitable (completely contraindicated) to deal with the constantly changing part of the minimum block.
That is, it is a technical condition and cannot be changed.