MT4 List or Array? | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Comprehensive list of results list of MT4 EAs and indicators 0 replies

Two Dimensional Array vs One Dimensional Array 13 replies

MT4 - Indicator Array Problem. 0 replies

MT4 array funkiness... wtf? 4 replies

Calculate Stochastic Formula on a array (non indicator buffer array) 2 replies

Attachments: MT4 List or Array?
Exit Attachments

MT4 List or Array?

  • #21
  • Edited 1:41am Jun 12, 2014 1:13am | Edited 1:41am
Quoting MaxDoom
{quote} you are still going about it the hard way imo test the outcomes just once and generate a unique number using binary, then simply use a case statement for the different outcomes. in this way you only test the scenarios once and do not have to do loads of IFs with condition checking. simplest,quickest, most efficient way of doing this
I understand it may be a caveman way of programming. But, "loads of ifs with condition checking" is what I know I can start implementing immediately. I know it's redundant and prone to errors of repetition and errors of omission.

It's not that I disagree with you. I hear what you are saying. I simply would not know where to begin unless I saw live implemented code to work with and copy, paste, customize and test.

I'm not asking you to post proprietary code or to code it for me or others here. I wouldn't ask you do that unless you wanted to, and more importantly unless I could immediately understand it.

Even if you posted it like rockit has done, I probably would not understand more than 2 or 3 lines of the code.

I understand a lot of the *basic* syntax of standard MT4 commands and logic and can understand suggestions in MQL and/or can look them up on online. However, the syntax of more advanced coding like nested loops and arrays or binaries (as you suggest) is not immediately in my mental grasp.

It's like this. I know how to drive a car. But, could I drive an F1 at top speed on a live course with other cars flying around me? Probably not effectively.

I appreciate the F1 type programming suggestions. The seeds have been planted. Right now I've realized even more that rock solid logic, immediate implementation and testing is more important to me than efficient or advanced coding methods.

Should it become necessary that other coding methods are absolutely necessary to implement what I need or want to implement, then it's time to acquire more advanced programming skills or partner up with someone that can. Right now it's testing what ideas I have, as quickly as possible and then learning more effective and more efficient coding later.

Thank you.
  • #22
  • Jun 12, 2014 2:20am Jun 12, 2014 2:20am
Quoting Pipologist
{quote} I understand it may be a caveman way of programming. But, "loads of ifs with condition checking" is what I know I can start implementing immediately. I know it's redundant and prone to errors of repetition and errors of omission. It's not that I disagree with you. I hear what you are saying. I simply would not know where to begin unless I saw live implemented code to work with and copy, paste, customize and test. I'm not asking you to post proprietary code or to code it for me or others here. I wouldn't ask you do that unless you wanted...
what ever works for you I guess....

but here is the basic principle:

You have 4 conditions to check, lets call then ABCD and each can be true or false.

Create an integer called STATUS and set to 0.

Check A and if TRUE add 1 to STATUS
Check B and if TRUE add 2 to STATUS
Check C and if TRUE add 4 to STATUS
Check D and if TRUE add 8 to STATUS

The value of STATUS now represents the complete set of answers from all 4 conditions.

If STATUS = 12 then you know that A&B = false C&D = TRUE and so on.

now you can use a Switch operator to carry out commands based upon the results as follows:

Inserted Code
int Status = 0;
if TestA  Status += 1;
if TestB  Status += 2;
if TestC  Status += 4;
if TestD  Status += 8;
 
switch(STATUS)
{
   case 1:
      Print("Trigger = A");
      break;
   case 2:
      Print("Trigger = B");
      break;
   case 3:
      Print("Trigger = AB");
      break;
   case 4:
      Print("Trigger = C");
      break;
   case 5:
      Print("Trigger = AC");
      break;
   default:
      Print("Trigger = none");
      break;
}

you can have as many case statements as required.
you can replace the Print commands with any number of lines of code that you wish, you might just set a boolean to True for a buy for instance and then after the Switch statement say IF DOBUY = TRUE .....

The switch is simply a tidy way doing all the IFs
The binary conversion is a simple way of not repeating all the status checks

The big advantage of the switch is that once it has processed the right case statement the whole stitch code is finished processing ie: it will not process any further case statements after it finds the match which is not true for a long list of IF statements
  • #23
  • Jun 12, 2014 3:56am Jun 12, 2014 3:56am
Quoting MaxDoom
{quote} what ever works for you I guess.... but here is the basic principle: ......The switch is simply a tidy way doing all the IFs The binary conversion is a simple way of not repeating all the status checks The big advantage of the switch is that once it has processed the right case statement the whole stitch code is finished processing ie: it will not process any further case statements after it finds the match which is not true for a long list of IF
statements
Thanks. That is POWERFUL. With the STATUS...

So this method is much faster and of course less redundant and less prone to error and once laid out, easier to code as long as one knows what they are doing.

So I suppose with 30 simple IF statements the difference in processing speed of the code is negligible.

However, from your experience at what point is there a measurable difference in speed of code, in other words how long of a list would there be for the status checks/switching code to be noticeably faster? An IF statement list of 100 conditions and checking, or 500 or more? My intuition tells me with a list of 100 or more somewhat complex if statements and checking would definitely benefit from the method you provided.
  • #24
  • Jun 12, 2014 6:37am Jun 12, 2014 6:37am
Quoting Pipologist
{quote} Thanks. That is POWERFUL. With the STATUS... So this method is much faster and of course less redundant and less prone to error and once laid out, easier to code as long as one knows what they are doing. So I suppose with 30 simple IF statements the difference in processing speed of the code is negligible. However, from your experience at what point is there a measurable difference in speed of code, in other words how long of a list would there be for the status checks/switching code to be noticeably faster? An IF statement list of 100 conditions...
it's easier and more tidy to code, making maintenance easier as well

as for speed benefit it is more relevant depending on how close to market you are trading, the nearer you are the sooner you want to process. I have not measured timings but the maintenance issue alone is worth using this technique.
  • #25
  • Last Post: Jun 14, 2014 6:43pm Jun 14, 2014 6:43pm
Regarding the speed... Just realized via experience, a longer list of if statements and condition checking takes much longer to run backtests. Running a multi parameter and multi EA, EA backtest now, with 29 buy conditions, 29 sell conditions, and it takes forever...my genetic algorithm pass is queued for 364 hours. Must learn and implement tidy coding if I want to expand my EAs further.
0 traders viewing now