Discussion of article "Filtering Signals Based on Statistical Data of Price Correlation" - page 2

[Deleted]  
Urain:

Another positive feature of filtering is that at 4-fold reduction of trades the profit drops by only 25%

It is not a secret that every entry into the market is a risk. And the winning strategy in coin is not to play at all.

And the fact that you can make much fewer trades with almost the same profit is good, very good.


If you trade on several symbols, but with sufficiently well filtered signals, you can reduce the number of trades on each individual symbol to get a profit of 25% more.

Or as an option, you can get the initial profit, but reduce the drawdown.

And this is definitely GOOD.

 
Suppose markets behave in a chaotic manner, like a coin. <br/ translate="no">

Therefore, when a new bar appears, the price has an equal opportunity to go up or down, and the previous bars do not affect the current one in the slightest way. Idyll! Create a trading system, set Take Profit higher than Stop Loss (i.e. bring the expectation matrix into the positive zone), and you are done. It's simply breathtaking.

On the quote: I didn't understand anything at all!!! What was that??? Poetry? :) And where is the maths?

And second: It doesn't matter the probability of closing direction, if more probable small profits will have less probable fat losers. Your statistics rather depends on the risk management details omitted from it ;)

Too few examples to analyse in your article. You could say that the example you gave is a fitting to the story.

 

Let's assume that markets behave like a coin, in a chaotic way.

Therefore, when a new bar appears, the price has an equal opportunity to go up or down, and the previous bars do not affect the current one in the slightest way. Create a trading system, set Take Profit higher than Stop Loss (i.e. bring the expectation matrix into the positive zone), and you are done. It's breathtaking.

Author, have you ever thought that a moose can trigger even in every trade, despite the fact that the number of upward closes will be equal to the number of downward closes. So, even if it is known that there will be as many upward closes as downward closes, the strategy you described is not suitable.

 

Hi Тарачков,

thank you for your fine article,

correct me if i am wrong....

you extracted data from the past and run the filtered expert in the same time?(both in 2010)

if this is the case with all due respect i think there is no point in this filtering. This kind of filtering proves nothing because obviously makes the results better....

I am not a fan of totally random walk market but i think you should have used one period(for example 2005) to extract the data for filtering and run your filtered expert for the next year(2006) and continue till the last year then compare it with the original expert to see if there is any correlations between the past price behavior and its future trends.

 

It´s really a interesting articel, very good.

Thank you for this writing.

 
Thank you for your article, I'll test and try to improve your idea on some of my EA's.
Thank you again !!!