Discussion of article "Practical application of correlations in trading" - page 3

 
Dmitry Fedoseev:

which is all the more reason it doesn't require it.

then it will lie in some cases, or rather predict nothing.

 
Why on earth would you put some kind of distribution in here? The data may not be random at all.
 
Dmitry Fedoseev:

Correlation for calculation does not require normalisation of data.

Of course it does not, we count the average everywhere and then subtract it, so there is no need to normalise.... theoretically.

But in practice, the subtraction of the average is normalisation, but by shifting the price series along the ordinate axis (Y) we change the relationship between the neighbouring data, and this operation already carries significant distortions. If we are looking for correlation of cattle population with arable land, then of course a regular Pearson will do a great job. In general, we return again to the essence of the article - if it is for beginners, then all these points should be told, and they are not, so this conclusion is made by me.

 
Denis Kirichenko:

I stronglydisagree. There are trading systems that do not make any forecasts, but rather exploit the statistical properties of a particular financial series or a set of them.

The author of the article is right - entering the market of any trading system is always a forecast of the state of a particular financial instrument. It does not matter whether static or dynamic properties are analysed. It is still a forecast.

 
Aleksandr Masterskikh:

The author of the article is right - entering the market of any trading system is always a forecast of the state of a particular financial instrument. It does not matter whether static or dynamic properties are analysed. It is still a forecast.

Yes, everything is right, Alexander. Whatever we do on the market, the essence of these actions is to predict further price movement and earn on it.

 
Aleksandr Masterskikh:

The author of the article is right - entering the market of any trading system is always a forecast of the state of a particular financial instrument. It does not matter whether static or dynamic properties are analysed. It is still a forecast.

You have distorted the terminology. I wrote not about static or dynamic, but about statistical properties.

Roger that. But I remain in my opinion and do not impose it on anyone....

I'll try to clarify my thought.

What the author wrote about concerns trading systems for which it is important where the price will go. But there are trading systems that do not care about the price direction. The main thing for them is that the price just fluctuates.

That's why it sounds a bit loud to me:

...entering the market of any trading system is always a forecast of the state of a particular financial instrument.

However, as for the forecast, I agree with this. When opening a position, the system somehow expects that its direction has been chosen correctly. Then yes, it is a kind of forecast....
 
When installing, the error: failed with code 1, what is it?
 

The article deals with the main types of correlation, as well as the corresponding indicators and test results. Although classical correlation analysis has limitations, namely, it requires a large amount of data for accuracy (and the author studies a small sample, moreover, within the same financial instrument), the author managed to show the methodology of correlation analysis for trend detection. Although the profit factor is small (equal to 1.19), it is enough to prove that correlation thresholds (in this case within the same financial instrument) are useful parameters for trend detection.

Thanks to the author for the article - correlation analysis is not easy and now many market participants will be able to enrich their arsenal of analytical methods.

 

Thanks to the author, he explained everything on his fingers, it's a pity I can't test the Expert Advisor, it gives errors

can't open "C:\Program Files\MetaTrader 5\MQL5\Experts\Trade.mqh" include file Correlation.mq5 10 11
'CTradeBase' - declaration without type Correlation.mq5 11 1
'MarginMode' - declaration without type Correlation.mq5 35 10
'Trade' - undeclared identifier Correlation.mq5 119 7

In my opinion, this is a mathematical attempt to reveal the A plane of the three-dimensional market model according to Eric Nyman - Price - Time, it is possible to rework the code for the B plane - Volume - Time, as well as the third plane C - Price - Volume including the relation to the volume of supply and demand (will tell with SYMBOL_SESSION_BUY_ORDERS and SYMBOL_SESSION_SELL_ORDERS).
 
Kisel31:

Thanks to the author, he explained everything on his fingers, it's a pity I can't test the Expert Advisor, it gives errors

can't open "C:\Program Files\MetaTrader 5\MQL5\Experts\Trade.mqh" include file Correlation.mq5 10 11
'CTradeBase' - declaration without type Correlation.mq5 11 1
'MarginMode' - declaration without type Correlation.mq5 35 10
'Trade' - undeclared identifier Correlation.mq5 119 7

To avoid confusion, I always write in my articles:

At the end of the article is attached an archive with all the listed files sorted by folders. Therefore, for correct operation it is enough to put the MQL5 folder in the root of the terminal.

It is likely that the Trade.mqh file is not in the same folder with Correlation.mq5