Professionals, let's create an EA together! - page 7

 
Pulat Rikhsibaev:

Please make the chart show the data of earned pips, max drawdown, etc.

Note
 
Evgeniy Gutorov:

Looking at how you build a zig-zag gives the impression that it is built by eye...

If you look at August 15 there was formed a shoulder ZZ exactly the same shoulder was formed on August 20 ... i.e. unclear system of building ZZ in the first place, it turns out it is necessary to consider the course of a longer shoulder ZZ ... ideal construction of ZZ can be built on randomly generated BP, in the market are three or four stages that can lead to ambiguous interpretation when building a mathematical model of behavior ...

In the beginning of this branch I demonstrated a case of ideal behaviour of Fibonacci levels... In reality there is 9% dilution, or if you use yen you can see unidirectional or shortened second leverage movements...For DE there are also movements with 23% spread which are quite often shown by observations...

Using a Zig-Zag is just one way to implement this Fibo method in an automated form. Yes, in some charts the lines were drawn by eye. I do not exclude the possibility of differences between the indicator readings and traders' opinions.

By the way, we are going to implement option one to build Fibonacci lines for now and as soon as it works out, we will switch to option two where lines are built a bit differently taking into account the strength of the initial movement. At the second option (you can find video projections on the website of the method's author) we could enter at reversal of EURUSD, if you pay attention, you will see that the price did not go back to 50% level - where many participants were waiting with pending sell orders.

 

Interesting strategy, thanks. I have slightly corrected the attached EA so that requests to the zigzag indicator were cached in the array, it will work faster.

I also added parameter TFTRADE to make the first zigzag build not by chart timeframe but by this parameter to optimize it. Depth, Deviation and Backstep have been made separately for two different zigzags, also for possible optimization.

Files:
Fibo.mq4  16 kb
Fibo.ex4  32 kb
 
Dr.Trader:

Interesting strategy, thanks. I have slightly corrected the attached EA so that requests to the zigzag indicator were cached in the array, it will work faster.

I also added parameter TFTRADE to make the first zigzag build not by chart timeframe but by this parameter to optimize it. Depth, Deviation and Backstep have been made separately for two different zigzags, also for possible optimization.

Thank you, I will report on my tests.
 
Izzatilla Ikramov:

Using a Zig-Zag is just one way to implement this Fibo method in an automated form. Yes, in some charts the lines were plotted by eye. I do not exclude the possibility of differences between the indicator readings and the traders' opinions.

By the way, we are going to implement option one to build Fibonacci lines for now and as soon as it works out, we will switch to option two where lines are built a bit differently taking into account the strength of the initial movement. At the second option (you can find video projections on the website of the method's author) we could enter at reversal of EURUSD, if you pay attention, you will see that the price did not go back to 50% level - where many participants were waiting with pending sell orders.

You could try this Fibo variant https://www.mql5.com/ru/forum/67132
Многоразрядные уровни Фибоначчи
Многоразрядные уровни Фибоначчи
  • www.mql5.com
Многоразрядные уровни Фибоначчи. - - Категория: общее обсуждение
 
Yousufkhodja Sultonov:
You can try this Fibo option https://www.mql5.com/ru/forum/67132

I wouldn't want to complicate life at least until all the possibilities of the current strategy have been exhausted.

Try asking the programmers here who are implementing the EA - it might include it as an option, if there is a specific ToR.

 
Yousufkhodja Sultonov:
You can try this variant of Fibo https://www.mql5.com/ru/forum/67132

I understand that you are suggesting to stretch the Fibonacci grid on the same chart segment as before (points "0" and "100" will be in the same places in both cases), the only difference will be intermediate lines, which will relate to each other in different proportions than usual?

Then in principle, you don't need to change anything in the Expert Advisor. Just convert the value of your multi-digit Fibonacci to two-digit, and change values of parameters FiboOpenDeal, TakeProfit1, TakeProfit2, StopLoss.

For example you have a value of 25 there in the table in the first post, instead of it use 38.19 in your Expert Advisor according to the second column. But imho this is not the best way, you want the chart to move by the description you thought up, while in fact your description should move to the chart. In general, you should do the opposite - optimize your EA on real data and look at the values you get for these parameters.

 
Dr.Trader:

I understand that you are suggesting to stretch the Fibonacci grid on the same chart segment as before (points "0" and "100" will be in the same places in both cases), the only difference will be intermediate lines, which will relate to each other in different proportions than usual?

Then in principle, you don't need to change anything in the Expert Advisor. Just convert the value of your multi-digit Fibonacci to two-digit, and change values of parameters FiboOpenDeal, TakeProfit1, TakeProfit2, StopLoss.

For example you have a value of 25 there in the table in the first post, instead of it use 38.19 in your Expert Advisor according to the second column. But imho this is not the best way, you want the chart to move by the description you thought up, while in fact your description should move to the chart. In general, you should do the opposite - optimise the EA on real data, and look at the values obtained for these parameters.

You're right, I want to test the hypothesis that the more we take the number of influencing values of a series (n with total N) on the last value of the series - the better, i.e. in the limit n=N. Standard Fibo has n=2 for any N.
 
Yousufkhodja Sultonov:
You are right, I want to test the hypothesis that the more we take the number of influencing values of the series (n for total N) to the last value of the series - the better, i.e., in the limit n=N. The standard Fibo has n=2 for any N.

You can't take forex by science or statistics. It's time to draw conclusions.
 

Fibonacci's not going to take it either to be honest :) But it seems that it may use zigzags.

I have tested my Expert Advisor on eurusd and it seems that it does not have universal settings and in the long term it either loses profit or balance goes up and down. However, it is good for optimization and keeps profit for some time. The optimization period should be ten times larger than the period of further trading. For example, one can optimize on two years and then trade for two months.

Also, I do not think it makes sense to close orders halfway, it's better to close the entire order at once.

Fibonacci levels are not universal, they must be changed according to optimization. Rebound points should also be changed, as well as stop loss, take profit and so on. Only the rule itself is universal: "If the price rebounds at X level of the previous zigzag, it will go up with some chance, and down with some chance. In order to catch these chances, a statistically profitable stop and take is set.