To be honest: this is exactly the kind of mod I had in mind, but I didn’t have enough time to implement it yet:

Looking at your code, and letting G’s reply aside for now, I would suggest some minor modifications of your code. Your main routine is as follows:

int start() {
double delta=2.0/(1.0+VZOPeriod);
double TV=0.0,VP=0.0;
for(int…[Read more]

I think what you’re marking as ‘inefficiencies’ or ‘false signals’ is the typical behaviour of an indicator that’s meant to be very reactive to fast changing short term conditions: works as designed

Taken as the only indicator defining your entries it will certainly empty your account in no time. But show me a single indicator this rule does…[Read more]

Ok, hope that this is working now. At least I did not manage my new Clear Method or ‘Peak & Trough’ indicator to display anything unstable.

Main difference as compared to my previous approaches: I allowed for a certain amount of redundant calculation instead of shifting historic values around in static arrays. Makes life easier from time to time…[Read more]

@smallcat:

This is not really true. You can calculate values for lower timeframes than shown on chart without any problem. You just can’t visualize those numbers in the way we are used to when looking at MT4 indicators.

Most simple case: the same timeframe on chart and on indicator (red volume histogram on my pic). We’re used to display a…[Read more]

Agree:

The one @Lowphat has posted is giving the same signals as my current WIP version. So as long as we’re using iCustom to test why not take that one?

It should be possible to derive a function for direct EA use from only one of our approaches, no matter which one. The logic of the algorithm demands to keep track of current and previous…[Read more]

Hey, nice exercise!

But I think it would be better for practical use to have oscillator and histogram in two separate indies … easier to scale independently.

And there’s a bug in the higher timeframe section: your loop is

for (i = limit; i >= 0; i--) {

and a few lines after that you have

if (!Interpolate || y == iBarShift(NULL,…[Read more]

Thank you again @Anti for the efforts you’re expending to get us started in R:

Take your time, will wait patiently! During the holiday season, most of us will have several activities outside the forum and outside MT4!

Best,

s.

No, you’re not:

I’ve read so many good posts of yours, I’m sure you’re not!

s.

Hi Kiads,

Could you possibly provide a link to mladen’s original (mq4 instead of ex4)?

Thanks,

s.

Thanks for pointing me there:

Interesting discussion, but different topic. They’re mainly talking about changing timeframes or pairs. When changing timeframes, I would expect a static var to be re-initialized. That’s absolutely ok, and it’s not the point in our discussion. There must be a bug in my code. I would assume it’s too simple for…[Read more]

Thanks for testing:

I would really be surprised if this was repainting (have a look at the code). I have a test running at M1, no repainted bars so far.

Re. strategy tester: I wonder whether strategy tester can cope with static variables static CLEAR_SAVE _save[];. Maybe this causes the problems. Any ideas anybody?

s.

EDIT: now I see the…[Read more]

Thanks for testing:

I would really be surprised if this was repainting (have a look at the code). I have a test running at M1, no repainted bars so far.

Re. strategy tester: I wonder whether strategy tester can cope with static variables static CLEAR_SAVE _save[];. Maybe this causes the problems. Any ideas anybody?

s.

Kiads already answered, and there’s not much to add.

Re. your code: it would be nice if MT4 indicators worked that way! If you want to have different colours on screen, you got to declare one buffer per colour. From the point of view of a software architect, this is suboptimal, trying to use friendly words. A total waste of memory…[Read more]

Thank you, @LittleCaro:

I’ve been informed by PM about discrepancies regarding my indicator on chart vs. inside backtester, called by iCustom(). Seems to be perfect on chart, but problematic in backtester.

Everybody interested, please check this and report back.

s.

That’s correct:

You’re pointing out the same detail I discussed with G. by PM: 3-bit vs. 4-bit logic. If you refer to my code:

enum CLEAR_STATE {
CLEAR_NONE,            // undefined condition
CLEAR_BULL,            // bullish condition
CLEAR_BEAR            // bearish condition
};

you will see my ‘clear’ preference for 3-bit, just…[Read more]

Yep:

I stepped in the same (at least similar) trap in my first attempt to code this! My latest version is absolutely simple, with no parameter at all.

s.

I’m afraid there’s some misunderstanding here. As far as I got it, V. does NOT talk about replacing all iCustom indicator calls by functions calls to mqh libs. It’s all about proper code management. Please trust him.

@VlanFx: please trust yourself and your expertise and do what you know to be right!

Re. iCustom() calls pro’s and con’s (we had…[Read more]

Absolutely correct! In the end we will have several mqh-libs, each serving a specific purpose. These will be glued together by the main program. Code maintenance and enhancement will be much easier that way!

When counting code lines, the largest project I was involved was several million lines for an investment bank. We…[Read more]

My simplified Clear Method Indicator, as requested. My clear method mqh lib is included in the mq4 file for convenience.

Buffer 0: output = 1 for bullish, else 0
Buffer 1: output = 1 for bearish, else 0

No input parameters, so iCustom() should be rather simple:
double iCustom(
string symbol, // symbol
int…[Read more]