Bill Williams and his strategies... - page 30

New comment
 
ZZZEROXXX:

Oooh, that's nice, even very nice. The only thing I would like to know is how many orders were opened in one direction at the most, maybe to limit it somehow.

"Senior trend filter on D1" - sounds mysterious, but seems to be effective.


I completely forgot - there are 10 orders in market simultaneously in one direction. All at 0.1 lot.
 

I am pasting up the optimization results: working TFs, stop levels and trawl types from 06.2002 till 06.2010, then up to the present time - forward, to the right of the red line.

First, on H4, trawl by shadows of 42 bars with an indentation of 35 real points, PB=15.9; second, on D1, trawl by shadows of 70 bars with an indentation of 15 real points, PB=12.5 .

The received cuts, I think, are interesting (especially it pleases, that entrances on five measurements of B.Williams... :-)))) and worthy of attention...

Anyway, there is definitely something in organizing market entries with pending orders according to five measurements by B.Williams... Let's move on...

Excluding premature exits (beyond the red line) of the alligator, we don't have sharp drops in the chart of balance change (as in the previous screenshots (pages(-s) above) but smooth spots where the market is slightly down.

areas with insignificant minus, but when catching a trend... All can be seen on the chart - it's not for nothing that the system is trending medium term.

 
Roman.:

In any case, there is definitely something in the way of market entry with pending orders according to B.Williams' five-dimensional method... Let's keep digging...


I suspect Williams is not to blame for such a curve ))), probably, it may be just about any system with fillings. It would be interesting to see the first of the three charts, which is without optimization, without the "senior filter". And also the question of how to calculate finrez. If the maximum 10 lots are involved, the profit and drawdown should also be divided by 10. And then we get a modest profit and negligible drawdown - an investor's dream. Or am I wrong?
 
ZZZEROXXX:

I suspect Williams is by and large not to blame for such a curve )))), probably any system with refills could have been in his place. It would be interesting to see the first of the three charts, which is without optimization, without the "senior filter". And also the question of how to calculate finrez. If the maximum 10 lots are involved, the profit and drawdown should also be divided by 10. And then we get a modest profit and negligible drawdown - an investor's dream. Or am I wrong?


What is that for five measurements with no returns? :-)))

When I do, I'll post these two variants you suggested as well. In the second one (without refill) we'll see how much "modest profit and paltry drawdown"... :-))) Haven't tested it myself yet - don't know.

Yes... An investor's dream - you're right... :-)))

 
ZZZEROXXX:

I suspect that Williams is not to blame for such a curve )))), probably any system with refills could have been in his place. It would be interesting to see the first of the three charts, which is without optimization, without the "senior filter". And also the question of how to calculate finrez. If the maximum 10 lots are involved, the profit and drawdown should also be divided by 10. And then we get a modest profit and negligible drawdown - an investor's dream. Or am I wrong?

I am pasting up test results of the Expert Advisor according to five B. Williams' dimensions - market entries - strictly by the system of 0.1 lot volume with a maximum of 10 orders simultaneously in the market, exits upon achievement of one of the levels - stop, take, or trailing variant.

The first of three charts (see the previous page) is a variant without optimization and filters - the same input parameters, H4 timeframe.

The second picture is similar to the second one on this page - without filters, Day, same input parameters.

The third picture is analogous to the first one on this page - H4, same input parameters.

"And another question - how to calculate the finrise. If 10 orders are used at most, we should divide profit and drawdown by 10. In this case we obtain a modest profit and negligible drawdown - this is the investor's dream".

Test picture of an owl with one order in the market - 0.1 lot - we enter the market only - "start" order (when meeting trading criteria, as recommended by B. Williams - breakdown of a fractal price), we do not use shares, working TF - day, input parameters are similar to parameters - see the screenshot above...


In all cases - on forwards - during 2010 to the present - the balance curve with an upward slope - that's the main thing.

Conclusion: B.Williams' five dimensions strategyis"working" (so far in the strategy tester)... :-))) It needs an "appropriate" approach.

In the trailer owls without any filters, with outputs - on reaching levels... Lines of code - commented out.

The structure of the owl is modular - the textbook analogue is here.

Anyone who is interested - I suggest to trust this "clean" version of the EA, by using different "filters". When they work on a different (from working) timeframe

for this purpose, the first variable in this list is selected, allowing its value to be optimized as well. 

extern int t_trend_period=7;      // в данной варианте системы не используется - это выбор ТФ - для старшего фильтра (см. строку ниже)

extern int s_signal_period=6;    // переменные периода индикаторов по пяти измерениям Б.Вильямса: 1-М1, 2-М5, 3-М15, 4-М30, 5-Н1, 6-Н4, 7-D1 - для тестирования

Include "your" filter variants and share them (both results and filters) in this branch of the forum if you get interesting results of testing and operation.

P.S. attached - EA with a set.file - working on a new bar opening (also tested using the model: "By opening prices (only for EAs with explicit bar opening control)"). Place the contents of the archive in the appropriate folders of the MT4 client terminal.

Files:
profitunityvslftpktral.zip  126 kb
 
the last picture without the "senior filter"?
 
ZZZEROXXX:
the last picture without the "senior filter"?


Yes. Look it up yourself with the same parameters - at the same time in the Criterion inclusion exclude via /* all refills and that's it...

Like this:

////-----------------------------------------------ЛОНГ-----------------------------------------------------
//  Старт - пробой фрактала ценой за пределами линии зубов Аллигатора
 if((Mas_Tip[0]==0 && Mas_Tip[1]==0 && upfractal > 0 && Bid > upfractal && Bid > Teeth)) // старт
     { 
      Print ("Функция определения ТК - покупка: старт, masstip 0  = ", Mas_Tip[0]);                                                                          
      return(10);// критерий на открытие ордера Buy при условии, что тек цена больше фрактала и линии зубов аллигатрора    
     }

/*   
if (Mas_Ord_New[0][0]<10 && Mas_Tip[0]>0)
{ 
   Print ("Функция определения ТК - покупка:  количество ордеров бай  = ", Mas_Ord_New[0][0]);  // Не более 10-ти доливок
...
...
...
...
    return(10000);    // Доливка по Линии баланса - устанавливаем байстоп в красной зоне 
   }
}
*/

//-----------------------------------------ШОРТ-------------------------------------------------------------- 
 //  Старт - пробой фрактала ценой за пределами линии зубов Аллигатора
         
 if((Mas_Tip[1]==0 &&  Mas_Tip[0]==0 && dwfractal > 0 && Bid < dwfractal && Bid < Teeth)) // старт
    {
      Print ("Функция определения ТК - продажа: старт, masstip 1  = ",Mas_Tip[1]);  
      return(20);// критерий на открытие ордера селл при условии, что тек цена ниже фрактала и линии зубов аллигатрора    
    }

/*   
if (Mas_Ord_New[0][0]<10 && Mas_Tip[1]>0)
{ 
   
   Print ("Функция определения ТК - продажа:  количество ордеров селл  = ", Mas_Ord_New[0][0]);  
...
...
...
*/
 
I.e. what's the trick in the end - is it the Williams strategy, the senior filter, or the early non-exit tactic?
 
ZZZEROXXX:
I.e., what is the main thing - the Williams strategy, the senior filter, or the early no-exit tactics?


Well, that's exactly how it all works together here...

I think it is necessary to try different variants of filtering (on higher TF) inputs according to B.Williams' five dimensions on lower TFs.

 
Roman.:


Dk that's exactly what's working in aggregate here....

I think it is necessary to try different variants of filtering (on higher TF) inputs according to B.Williams' five dimensions on lower TFs.

It is brilliant, Roman. Thank you very much for your efforts in writing EAs, everything is working fine https://www.youtube.com/watch?v=ONfnwq9n29U, I am waiting for the new versions of your filters, maybe with wipes with 240 sma and 60 ema. I would like to see the signals from Trading Chaos2 in them, after the formation of the angulation 1st sage, 2nd sage and 2 PICK





1...232425262728293031323334
New comment