What you need to know about MT4 strategy? - MQL4 and MetaTrader 4

mandor 2006.09.29 22:42 #11

Renat писал (а):
And you do the research and publish the discrepancies - it would be of interest to everyone. First and foremost, it would be of interest to us. If you do it as an article for the Articles section, we will pay for it (so far we have paid $1,720 for a number of articles).

Well ... I do not aspire to the dubious fame of a boffin-writer in the "Grail on Forex" genre. I am a practicing programmer, trader and investor. MT4 terminal is very good, MQL4 language is not bad, MT4 strategy tester is ugly. Until you understand it, you will be making endless mistakes. So far the only correct way to evaluate the strategy is by eye on the charts first, and then in real time on the demo. The MT4 tester gives only illusions. All that stuff with optimization and visualization is useless until there is a normal sequence of output data. Apparently developers see the real market model too primitively.

© Herurg

From theory to practice A pattern. Strategy Testing EA, Maximal

Renat Fatkhullin 2006.09.29 22:54 #12

mandor:

Renat:
And you do the research and publish the discrepancies - it would be of interest to everyone. First and foremost, for us. If you do it as an article for the Articles section, we will pay for it (at the moment we have paid $1,720 for a number of articles).

Well ... I do not aspire to the dubious fame of a boffin-writer in the "Grail on forex" genre. I am a practicing programmer, trader and investor. MT4 terminal is very good, MQL4 language is not bad, MT4 strategy tester is ugly. Until you understand it, you will be making endless mistakes. So far the only correct way to evaluate the strategy is by eye on the charts first, and then in real time on the demo. The MT4 tester gives only illusions. All that stuff with optimization and visualization is useless, until there is a normal sequence of output data. Apparently developers see the real market model too primitively.

© Herurg

Do bother to _prove_ your words right here with all the proofs. Otherwise, the excessive amount of unsubstantiated expressions "glory to the bozos writer", "MT4 is ugly", "MT4 tester gives only illusions", "nothing" and "Apparently, the developers have too primitive a model" will result in certain actions on our part.

This is not the first time you have allowed directly insulting and unsubstantiated statements. I didn't say that for nothing:

Trying to play with words in a mode of "if I said something wrong, you just misunderstood me".

you don't need excuses - prove with all the details and screenshots what you said or make a public apology.

"The 'perfect' trading system New MetaTrader 5 platform The perfect filter

mandor 2006.09.29 23:02 #13

Renat писал (а):

Do bother to _prove_ your words right here with all your calculations. Otherwise, the excessive amount of unsubstantiated expressions of "bozo glory", "MT4 is ugly", "MT4 tester gives only illusions", "empty" and "Apparently the developers see the model too primitively" will result in certain actions on our part.

There are links in post 4. They prove a significant difference in the set of ticks for different periods. Concerning the translation of stats calculated for periods M15-Daily, I'll provide later. I have a dark night here, it's almost morning. What have I done to make MQ so unpleasant that I'm considered my personal enemy?

© Herburg

Impulse Metatester Agent on Windows Machine learning in trading:

Renat Fatkhullin 2006.09.29 23:08 #14

There are no references in post 4. And most importantly - it's not about references, it's about absolutely accurate evidence and full details operated by honest programmers (not forum fluffers). Any reader of this forum should easily understand, reproduce and accept this evidence.

Do bother precisely tomorrow to provide a number of (1) screenshots, absolutely (2) detailed data and (3) full EA code that will show you are correct. Otherwise you will be permanently banned from the site for making silly and unsubstantiated accusations.

Why does MQL5 delete Standard misconceptions in trying Should I report an

Nikolay Zhilin 2006.09.30 04:30 #15

Renat, please don't put pressure on a colleague :) He is right somewhere. So we need to find out where exactly. Before the championship I spent quite a lot of time testing, including a real-time demo. There are some unpleasant things I don't understand. My program is not ticking, it's rather crude. It seems that it should not react to small things. But when explicit conditions are not met, I don't get it. In addition, very often there is an opening outside the bar. This is nonsense to me because I thought all ticks should be accounted for in the bar. But no. So, let's try to find the reason. The sooner we find and fix it, the sooner your product will be even better. You should be more interested in this than anyone else. If you want, I will try to restore a couple of fragments with ambiguities as soon as I have time.
On the other hand I agree that working purely on ticks and minutes is a road to nowhere. I don't know, maybe I don't understand something yet, but I don't want to try it. And on TF greater than M1 it is very difficult to make MTS working equally on all TFs. I had such a misconception before. An example is using parallel channels where we must inevitably manage the channel period when changing TFs and volatility measurement parameters. I think there are a lot of such examples.

Questions from Beginners MQL5 Experts: SwingCyborg Help, I don't know

Dmitry Fedoseev 2006.09.30 05:55 #16

1159315200 = 27.09.2006 0:00

M1 Time	H4 Time	M1 Close	H4 Close
1159315200	1159315200	1.2688	1.2689
1159315215	1159315215	1.2688	1.2688
1159315259	1159315259	1.2689	1.2689
1159315260	1159315275	1.269	1.269
1159315275	1159315335	1.269	1.2688
1159315319	1159315379	1.2688	1.2689
1159315320	1159315380	1.2688	1.2687
1159315335	1159315392	1.2688	1.2687
1159315379	1159315404	1.2689	1.2686
1159315380	1159315439	1.2687	1.2687
1159315392	1159315499	1.2687	1.2688
1159315404	1159315560	1.2686	1.2687
1159315439	1159315566	1.2687	1.2688
1159315499	1159315578	1.2688	1.2687
1159315560	1159315584	1.2687	1.2688
1159315566	1159315590	1.2688	1.2687
1159315578	1159315596	1.2687	1.2688
1159315584	1159315619	1.2688	1.2687
1159315590	1159315679	1.2687	1.2688
1159315596	1159315695	1.2688	1.2687
1159315619	1159315739	1.2687	1.2686
1159315679	1159315740	1.2688	1.2685
1159315680	1159315755	1.2687	1.2685
1159315695	1159315799	1.2687	1.2684
1159315739	1159315800	1.2686	1.2686
1159315740	1159315859	1.2685	1.2685
1159315755	1159315979	1.2685	1.2686
1159315799	1159315995	1.2684	1.2685
1159315800	1159316039	1.2686	1.2686
1159315859	1159316400	1.2685	1.2685
1159315919	1159316459	1.2686	1.2686
1159315979	1159316579	1.2686	1.2687
1159315980	1159316580	1.2685	1.2686
1159315995	1159316639	1.2685	1.2687
1159316039	1159316775	1.2686	1.2688
1159316400	1159316790	1.2685	1.269
1159316459	1159316819	1.2686	1.2689
1159316579	1159316835	1.2687	1.269
1159316580	1159316879	1.2686	1.2689
1159316639	1159316892	1.2687	1.269
1159316760	1159316904	1.2688	1.2689
1159316775	1159316939	1.2688	1.269
1159316790	1159316940	1.269	1.2689
1159316819	1159316952	1.2689	1.269
1159316820	1159316964	1.269	1.2687
1159316835	1159316999	1.269	1.2689
1159316879	1159317059	1.2689	1.2688
1159316880	1159317060	1.269	1.2687
1159316892	1159317074	1.269	1.2685
1159316904	1159317088	1.2689	1.2686
1159316939	1159317095	1.269	1.2687
1159316940	1159317119	1.2689	1.2686
1159316952	1159317140	1.269	1.2687
1159316964	1159317179	1.2687	1.2684
1159316999	1159317180	1.2689	1.2683
1159317059	1159317204	1.2688	1.2685
1159317060	1159317212	1.2687	1.2684
1159317074	1159317220	1.2685	1.2683
1159317088	1159317239	1.2686	1.2684
1159317095	1159317240	1.2687	1.2683
1159317119	1159317270	1.2686	1.2684
1159317120	1159317282	1.2687	1.2683
1159317140	1159317299	1.2687	1.2684
1159317179	1159317300	1.2684	1.2682
1159317180	1159317316	1.2683	1.2684
1159317204	1159317324	1.2685	1.2683
1159317212	1159317340	1.2684	1.2682
1159317220	1159317359	1.2683	1.2683
1159317239	1159317360	1.2684	1.2682
1159317240	1159317368	1.2683	1.2683
1159317270	1159317376	1.2684	1.2682
1159317282	1159317419	1.2683	1.2681
1159317299	1159317599	1.2684	1.2683
1159317300	1159317615	1.2682	1.2684
1159317316	1159317659	1.2684	1.2683
1159317324	1159317720	1.2683	1.2681
1159317340	1159317779	1.2682	1.2682
1159317359	1159317780	1.2683	1.2681
1159317360	1159317839	1.2682	1.2682
1159317368	1159317959	1.2683	1.2681
1159317376	1159318319	1.2682	1.2683
1159317419	1159318395	1.2681	1.2682
1159317599	1159318439	1.2683	1.2681
1159317600	1159318499	1.2684	1.2682
1159317615	1159318500	1.2684	1.2681
1159317659	1159318559	1.2683	1.2682
1159317720	1159318560	1.2681	1.2683
1159317779	1159318580	1.2682	1.2684
1159317780	1159318619	1.2681	1.2683
1159317839	1159318979	1.2682	1.2684
1159317959	1159319115	1.2681	1.2685
1159318319	1159319130	1.2683	1.2683
1159318380	1159319159	1.2682	1.2685
1159318395	1159319160	1.2682	1.2684
1159318439	1159319180	1.2681	1.2685
1159318499	1159319219	1.2682	1.2684
1159318500	1159319230	1.2681	1.2683
1159318559	1159319279	1.2682	1.2684
1159318560	1159319415	1.2683	1.2683
1159318580	1159319459	1.2684	1.2684
1159318619	1159319519	1.2683	1.2683
1159318979	1159319639	1.2684	1.2684
1159319100	1159319819	1.2685	1.2683
1159319115	1159319939	1.2685	1.2682
1159319130	1159320119	1.2683	1.2681
1159319159	1159320132	1.2685	1.268
1159319160	1159320179	1.2684	1.2681
1159319180	1159320479	1.2685	1.2682
1159319219	1159320480	1.2684	1.2683
1159319220	1159320539	1.2683	1.2682
1159319230	1159320719	1.2683	1.2683
1159319279	1159320839	1.2684	1.2682
1159319400	1159320840	1.2683	1.2683
1159319415	1159320860	1.2683	1.2684
1159319459	1159320899	1.2684	1.2683
1159319519	1159321019	1.2683	1.2684
1159319639	1159321020	1.2684	1.2685
1159319819	1159321079	1.2683	1.2686
1159319939	1159321139	1.2682	1.2685
1159320119	1159321379	1.2681	1.2686
1159320120	1159321560	1.268	1.2687

Dmitry Fedoseev 2006.09.30 05:58 #17

Here is the expert on which this data is derived:

//+------------------------------------------------------------------+
//|                                                          111.mq4 |
//|                                                                  |
//|                                                                  |
//+------------------------------------------------------------------+
#property copyright ""
#property link      ""
int h;
//+------------------------------------------------------------------+
//| expert initialization function                                   |
//+------------------------------------------------------------------+
int init()
  {
//----
   h=FileOpen(Symbol()+Period()+".txt",FILE_CSV|FILE_WRITE);
 
//----
   return(0);
  }
//+------------------------------------------------------------------+
//| expert deinitialization function                                 |
//+------------------------------------------------------------------+
int deinit()
  {
//----
   FileClose(h);
//----
   return(0);
  }
//+------------------------------------------------------------------+
//| expert start function                                            |
//+------------------------------------------------------------------+
int start()
  {
//----
 
  if(CurTime()>=StrToTime("2006.09.27 0:0")){  
     FileWrite(h,CurTime(),Close[0]);
  }
//----
   return(0);
  }
//+------------------------------------------------------------------+

Renat Fatkhullin 2006.09.30 06:04 #18

rebus:
There are some unpleasant misunderstandings. My programme is not a ticking programme, it's quite crude. It seems like it should not react to small things. But when explicit conditions are not met, I don't understand it. In addition, very often there is an opening outside the bar. This is nonsense to me because I thought all ticks should be accounted for in the bar. But no. I.e. let's try to find the reason.

Please publish the full code and screenshots of charts with openings outside the bar. Very interesting to see.

Renat Fatkhullin 2006.09.30 06:39 #19

Integer, thanks for the code and research. Only there is no clear conclusion (interpretation) in your post, which allows unsophisticated users to draw wrong conclusions based on 1-2 pips different values. The average person will draw a perfectly simple conclusion "Well yes - that's right! You see - two numbers in the line and they differ by 1-2 pips! That's right - it's a mistake!", isn't it?

You forgot to mention that you did a time scale overlap, which is not absolutely correct. The tester is trying to more or less evenly cover the time interval being tested and the exact time is absolutely irrelevant to it. It makes no difference if we set the time to 10:00:01 or 10:00:03, especially in the H4 timeframe. The main thing is to simulate the price movement as accurately as possible, not its time-per-second characteristics. And you are taking the exact second-by-second time frame as the basis and drawing the wrong conclusions.

It's by directly comparing bars per second that you get the small difference of 1-2 pips (and even more so when comparing M1 and H4). Which is absolutely normal and a fairly accurate calculation. I'll say it again - you should write loaded EAs, not react to noise!

I also ran this Expert Advisor and got almost the same results. All the checkpoints (start to end) minutes (M1 and H4) are exactly the same. And the charts are exactly the same.

How to code? Questions from Beginners MQL4 What makes an unsteady

Renat Fatkhullin 2006.09.30 06:45 #20

rebus:
Renat, please do not put pressure on a colleague :)

He will become a colleague when he can prove his words in detail and publicly.

Advice on not using the MetaTrader 4 Strategy Tester - page 2