EURUSD - Trends, Forecasts and Implications (Part 1) - page 50

New comment
 
EvgeTrofi >> :


And if you add a competent technical analysis and not just open deals at random, you can get some stability and high profitability:

This is a 10 year old GBPUSD M30 test

Please help me to find the parameters: The goal is to reduce the number and size of drawdowns in maximal life time on the tester for the new version of TENGRI_Trofi 1.4.13.

Attached here are some of my examples (test.rar file) and the new version of TENGRI_Trofi...

I have a very good working filter (personally invented by me and my fellow countryman leonid553), which we called a channel filter. Just for "near-martingale" systems.

The idea is that the very first deal of the cycle is opened not randomly but only in the direction of the trend when it touches or breaks the channel.

This channel is drawn by the Expert Advisor itself.

It goes like this:

In the external parameters we insert:

extern string  _____  = "=== Channel-Фильтр ==="; 
extern bool     DrawChannel=true;
extern int      Cbar=8;
extern int      Min_ch=15;

Next, define the "drawing" part (function) itself, which we place at the very end of the EA code (outside the START function)
void DrawChannel (double HCh, double LCh, double MidCh) {
    //--- Рисуем верхнюю границу канала
   if (ObjectFind("HCh_"+Time[0]) < 0) {
      if (ObjectFind("HCh_"+Time[1]) < 0) double PrevHChY= HCh;
      else  PrevHChY=ObjectGet("HCh_"+Time[1], OBJPROP_PRICE2);
   ObjectCreate("HCh_"+Time[0],OBJ_TREND,0,Time[1], PrevHChY,Time[0], HCh); 
   ObjectSet("HCh_"+Time[0], OBJPROP_RAY,false);
   ObjectSet("HCh_"+Time[0], OBJPROP_COLOR,Yellow);  
   ObjectSet("HCh_"+Time[0], OBJPROP_WIDTH,2);     
   }
   else ObjectSet("HCh_"+Time[0], OBJPROP_PRICE2, HCh);
   
   //--- Рисуем нижнюю границу канала
   if (ObjectFind("LCh_"+Time[0]) < 0) {
      if (ObjectFind("LCh_"+Time[1]) < 0) double PrevLChY= LCh;
      else  PrevLChY=ObjectGet("LCh_"+Time[1], OBJPROP_PRICE2);
   ObjectCreate("LCh_"+Time[0],OBJ_TREND,0,Time[1], PrevLChY,Time[0], LCh); 
   ObjectSet("LCh_"+Time[0], OBJPROP_RAY,false);
   ObjectSet("LCh_"+Time[0], OBJPROP_COLOR,Yellow);  
   ObjectSet("LCh_"+Time[0], OBJPROP_WIDTH,2);    
   }
   else ObjectSet("LCh_"+Time[0], OBJPROP_PRICE2, LCh);   
   
   //--- Рисуем среднюю границу канала
   if (ObjectFind("MidCh_"+Time[0]) < 0) {
      if (ObjectFind("MidCh_"+Time[1]) < 0) double PrevMidChY= MidCh;
      else  PrevMidChY=ObjectGet("MidCh_"+Time[1], OBJPROP_PRICE2);
   ObjectCreate("MidCh_"+Time[0],OBJ_TREND,0,Time[1], PrevMidChY,Time[0], MidCh); 
   ObjectSet("MidCh_"+Time[0], OBJPROP_RAY,false);
   ObjectSet("MidCh_"+Time[0], OBJPROP_COLOR,Blue);  
   ObjectSet("MidCh_"+Time[0], OBJPROP_WIDTH,1);       
   }
   else ObjectSet("MidCh_"+Time[0], OBJPROP_PRICE2, MidCh);  
}



In addition, in the inite and deinite also add :
int init()
{
//----------------------------------------------------------
ObjectCreate("lowline",OBJ_HLINE,0,0,0,0,0);
ObjectCreate("highline",OBJ_HLINE,0,0,0,0,0); 
ObjectCreate("avline",OBJ_HLINE,0,0,0,0,0); 
ObjectSet("lowline", OBJPROP_BACK,1); 
ObjectSet("highline", OBJPROP_BACK,1);
ObjectSet("avline", OBJPROP_BACK,1);  
}
//-----------------------------------------------------------
int deinit()
{
ObjectDelete("lowline");
ObjectDelete("highline"); 
ObjectDelete("avline"); 
if (!IsTesting()) ObjectsDeleteAll(0,OBJ_TREND);  
}

Next, set the START function to :
int start()
{
       
//--------------------------------------------
Hbid= iHigh(NULL,0,iHighest(NULL,0,MODE_HIGH, Cbar,0));
Lbid= iLow (NULL,0,iLowest (NULL,0,MODE_LOW, Cbar,0));  
if ( DrawChannel) DrawChannel( Hbid, Lbid, sredn);
ObjectMove("lowline",0,Time[ Cbar], Lbid); 
ObjectMove("highline",0,Time[ Cbar], Hbid); 
ObjectMove("avline",0,Time[ Cbar], sredn);

Now we have everything we need to set the conditions for the inputs.

For the initial, - the very first entries with a min lot.

if ( High[1]>= Hbid && ( Hbid- Lbid)> Min_ch*Point ){// если предполагается
// движение вверх - // открываем сделку бай
.... .... ....
if ( Low[1]<= Lbid  && ( Hbid- Lbid)> Min_ch*Point) {//если предполагается
 //движение вниз// то открываем сделку селл

Application of the described filter gives approximately 15-35 percent additional profit. Especially for trend movements when the Expert Advisor gives entries only for the trend.

For example :

It can be clearly seen that with such movements, the Expert Advisor will give signals only in the NELL direction.

BRN oil. M5



Parameters

extern int Cbar=12;
extern int_ch=28;
can be optimized.

 
forex-k писал(а) >>

"grail toys"

I'm preparing a third version.

8) I'm shocked, where is the drawdown? or from, but a small profit drawdown is not much? what is the optimal parameters to choose? there is a distance between orders and the size of the profit

 
strangerr писал(а) >>

What kind of miracle is this?

I'm shocked by the results, I gave up on this system when I realized that I can not sit long in the box, only on the move it makes money

 
firemast >> :

I'm shocked with the results, I gave up on the system when I realised I couldn't sit in the box for long, it only makes money on the action.

I've been messing around with it, too.

I'm looking for the right entry points and distances between orders.

 
firemast >> :

I'm shocked at the results, I gave up on this system when I realized that I can not sit long in the box, only on the move it makes money


Yuri, why even look for grails? It is clear that no profitable mechanical trading system exists, because price movements on markets are the fear and excitement of the crowd. It is a much more complex system, and a computer will not be able to calculate it in 50 years. And you are talking about robots assembled on their knees, which can hardly add 2+2.
 
I used to do programming and it taught me one simple thing - a computer was invented by man according to a very simple principle - 0 and 1, YES and NO, but the thing is that man also operates with
a word like UNKNOWN and that's something the computer can't calculate. But when it will be able to calculate the term, we will have an artificial mind. Then no one will trade. I admit that there are good - semi-mechanical systems, but the success of these systems depends on the carriage driver, who presses the pedals. If it's Warren Buffett, then the system MAY go in the plus, but if it's you and me, then it's like flying an aeroplane with no idea how to do it.
 
Krotu >> :


It's clear to everyone that there is no single profitable mechanical trading system and......

My real (Broko). Schedule from 17 Aug. to seg. (i.e. one month)

95 PROC. Deals realized by Scalper Advisor on small timeframes.


 
rid >> :

My real. Schedule from 17 Aug. to seg. (i.e. one month)

95 PROFITS. The Expert Advisor realizes the deal.



Reid - I'm talking about a robot, not an EA. Two questions - do you work on real or demo? Does the advisor decide on entry, exit and lot size or yourself?
 
Krotu >> :
But the point is that a person still operates with such a word as UNKNOWN and this is just the computer can't calculate.

Quite capable.... UNKNOWN is when there is neither YES nor NO, then UNKNOWN.

But quite agree with you 100% as there are many more interesting words a person knows e.g. MAYBE.

In general, it's not even a question of words, the computer is not able to feel...

 
RomanS >> :

Quite capable.... UNKNOWN is when there is neither YES nor NO, then UNKNOWN.

But I quite agree with you 100%, because many more interesting words a man knows e.g. CAN BREAK.

Actually, it's not even about words, a computer can't feel...


You're right on the money.)
1...434445464748495051525354555657...2305
New comment