[Archive!] Any rookie question, so as not to clutter up the forum. Professionals, don't pass it by. Couldn't go anywhere without you - 2. - page 35
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
It slows down, of course, but it all depends on specific indicators. For simple calculations, it is quite acceptable, but it provides time savings during development. In this way you may check the idea very quickly and happily throw it in the trash. If the results are encouraging. then it is possible to reduce it to a single indicator.
Programmers in general do not trust anyone (I'm not a programmer :)) ), so when it comes to the use of indicators they are divided into dull-pointed and sharp-pointed.
Some believe that the algorithms transferred directly from the indicator to the Expert Advisor are the fastest.
Others say that the difference is not so significant as to complicate the code. And sometimes, the introduction of calculations into the Expert Advisor even slows down the testing.
There are experts who are very skillful in optimizing the speed of the code, and there are not so many of them, even among the professionals.
Read articles in the Tester and other sections, it will be interesting.
And it's more convenient for the simple country boy to keep everything in the indicator and send signals to the Expert Advisor from there. This allows to easily modify the system, change and rewrite indicators, use several indicators simultaneously, etc. It is noteworthy that one of the most experienced forum programmers is of the same opinion.
interesting excursus
The broker's response delay time negates all optimization of the code.
For pending orders - serious code optimization is not crucial (modularity - easy to develop and modify the code).
Thank you, very interesting. I've just 'made' my first EA, actually just changed a couple of lines of code in the finished one. Here's what came out:
"Shifter", no stops. Now I have decided to optimize it.
Please advise me, where can I read about good results of testing in MT4 (drawdown, mathematical expectation, profitability and other parameters given by tester or may be obtained in excel). Something not too academic-encyclopaedic, simple and practically useful.
Damn, I think I've hit the net... :)
Thank you, very interesting. I've just 'made' my first EA, actually just changed a couple of lines of code in the finished one. Here's what came out:
"Shifter", no stops. Now I have decided to optimize it.
Please advise me, where can I read about good results of testing in MT4 (drawdown, mathematical expectation, profitability and other parameters given by tester or may be obtained in excel). Something not too academically encyclopaedic, simple and practically useful.
Damn, I think I've hit the net... :)
The best result will be achieved during testing in real-time mode. Other times, in the tester, an EA is good, but in real time, it loses everything without paying attention to any expected payoffs. The easiest option is to use a demo account. But even better, open a cent account with mimimal lot = 0.01 and having 10 quid put there, throw the EA on the real account, because testing on the demo account differs from the real one. Well, 10 quid with this lot will give you at least the equivalent of the EA on the account with a deposit of 10 000 dollars. Even with errors, but this is a real-time trading on a real account. If you don't mind shelling out 10 quid, go for it.
...Shit, I think I've hit the nets... :)
Do a search on the magic words you named and you will be horrified at the amount of material and opinions expressed in it. I still haven't formed my own opinion.
The only thing I'm firmly convinced of is: optimize on one timeframe, then run single tests on another one (OOS or forward, as they're called) with better data. Of interest are EAs which show similar results on new (forward) timeframes. If they don't plummet dramatically, it's already something, you can fiddle around. If they keep zero or show a stable trend, then this is something worth to improve.
If the behavior of the Expert Advisor on the forward section is radically different from the optimized one - in the recycle bin without any regrets, cheating and tweaking.
All this applies to codes, which do not use questionable technologies, playing on quotes peculiarities of broker (pips, divergence catchers.
And, well, Vladimir reminded me - practice is the only criterion of truth. Any promising result in the tester is tested at least on demo. If the results are significantly different - to the bin. Whatever the reasons, it is planned to trade on the real, not in the tester.
Real-time testing will give the best results.
It's a long time, I don't have intraday... I've had about 100 trades in a year - this way you'll be testing real time until your retirement :)
Long, I don't have intraday... I've had about 100 trades in a year - this way you'll be testing them in real time until your retirement :)
I don't even consider strategies with such a frequency of trades, because it would take a lifetime to test them. You'll find out it's unprofitable on your deathbed and ruin the solemnity of the moment.
Pray, the process is irreversible now... :))
Search through the magic words you mentioned and you'll be astonished at the number of materials and opinions expressed in them. I still haven't formed my own opinion.
The only thing I'm firmly convinced of is: optimize on one timeframe, then run single tests on another one (OOS or forward, as they're called) with better data. Of interest are EAs which show similar results on new (forward) timeframes. If they don't plummet dramatically, it's already something, you can fiddle around. If they keep zero or show a stable trend, then this is something worth to improve.
If the EA's behavior on the forward area is drastically different from the one on the optimized area - no regrets, cheat and tweak in the basket.
Thanks very much for the opinion!!!
I was just thinking myself that a certain abnormal section on one period could ruin, sorry, devalue the test results - say, the tseta was dangling in swings of three to seven figures during the year, but there was one short section where it flew twenty figures without stopping - the value of such optimization for future trading is probably very low, because averaging will play its part...
Yes, there are a lot of materials on this subject, I know. I will look into them.
Thank you once again.
I don't even consider strategies with such a frequency of trades, because it's not enough life to check them. You'll find out it's unprofitable on your deathbed and ruin the solemnity of the moment.
:)
I think it's still cheaper than wading through a swamp of fluctuations and noise on minute charts. A trade once every 1-5 days is not that long.
Good day! I can't figure out function 'BarsSince' - function is not defined C:\FXstart\Documents/experts/other.mq4 (72, 96)
I did not understand how to write it, it stays in black letters in MetaTrader editor and so barssince, I googled it on the question (how to memorize a bar), turns out the function memorizes the bar under certain conditions, gives a value for how many bars passed from one to the other, let's say, intersection of one mean, mentioned on a dozen forums, and the help can not find it!
I am stupid, it is a mystery. I am still thinking on my own.
Good day! I can't figure out the function 'BarsSince' - function is not defined C:\FXstart\Documents\experts/other.mq4 (72, 96)
I do not fully understand how to write it, it is in the editor of Metatradera remains in black letters and so barssince, I found it on google on the question (how to remember the bar), it turns out that the function memorizes the bar under certain conditions, gives a value for how many bars passed from sabitiyi, let's say the intersection of one midway another, mentioned a dozen forums, and in Help find it ne magu!
I am stupid, it is a mystery. I am still thinking on my own.
What's the function?