Combining multiple indicators in an EA - page 3

 
avatara писал(а) >>

50% were in each other's way!

Or what?

What's the point of trading on them if you are going to lose as much as you have traded on them?

 

We're talking about probabilities.

Make it clear.

I repeat - believe all the signals.

That's 80%.

What's wrong with that?

 

avatara, in this case you should believe only 30% of signals, and you should either forget about the remaining 70% or filter them

and work with them by another part of the system. Where did the 80% comes from? 30%.

 
Richie >>:

avatara, в данном случае надо верить только 30% сигналов, а про остальные 70% надо либо забыть, либо фильтровать их

другими индюками и уже работать с ними по другой части системы. Откуда 80% то взялось? 30 процентов.

In order to filter, these filtered indicators need to know MORE than the filtered ones. And why the hell do we need filtered ones then?

 
joo писал(а) >>

In order to filter, these filtered indicators need to know MORE than the filtered ones. And why the hell do we need filtered ones?

Several strategies can be combined in one EA. But, this is if one is not enough. Although, if each of them is bad - then increasing the

.

 
Richie >>:

avatara, в данном случае надо верить только 30% сигналов, а про остальные 70% надо либо забыть, либо фильтровать их

другими индюками и уже работать с ними по другой части системы. Откуда 80% то взялось? 30 процентов.


Richie 06.01.2010 18:07


avatara wrote >>

50% were in each other's way!

Or what?

What is the point of trading on them if you will then also lose as much as you have traded on them?

I repeat myself. So:

They lie together - 20%.

they contradict each other, but one of them says the right thing - 50%.

Together they say it right - 30%.

Hence the conclusion - 80% if you believe each - apart and together.

;)



 

My point is that it makes sense to abandon the concept of "correctness" of the signal altogether.

It is generally accepted, and I think this is what is meant here, that a signal will be considered correct if after opening on a signal to open a position the position is closed on the reverse of the indicator signal with a profit. This "discreteness" of signals, yes or no, + or -, in my opinion, is the reason for some comrades' claims that TA doesn't work. With this approach combining any of the possible combinations of indicators, we will get worse results than the performance of the indicators individually.

 

Now I understand where 80% is coming from. But, the "AND" principle implies using exactly that 30%. And 80% is

a very different principle-- "OR."

 
joo >>:

Я клоню всё к тому, что имеет смысл отказатся вообще от понятия "правильность" сигнала.

Общепринято, и здесь, я думаю, именно это имелось ввиду, что правильным будет считатся сигнал, если после открытия по сигналу на открытие позиции позиция закроется по обратному сигналу индикатора с прибылью. Эта "дискретность" сигналов, да или нет, + или -, по моему, и является причиной утверждений некоторых товарисчей, что ТА не работает. С этим подходм объединяя любые из возможных комбинаций индикаторов, мы будем получать результат хуже, чем работа индикаторов по отдельности.

This was not discussed precisely.

We are studying probabilities. The topic-starter should prove his conclusion.

Otherwise it's not clear to newcomers. ;)

 
avatara >>:

Это не обсуждалось как раз.

Мы вероятности изучаем. Свой вывод топикстартеру следует доказать.

А то новичкам не понятно. ;)

And I thought the topic meant the practical aspects of crossing indicators. OTV, I don't think, is of any help in this matter.