Absolute courses - page 24

 
grell:

Normalised how?


HOW MANY TIMES. DIVIDED BY THE INITIAL BAR. WHICH is DEFINITELY for D = 1, and for E=ED in this bar (this is E0), and for Y = 1/DY in this bar (this is Y0). It's not the point at all. I can put the yen there as a unit. You don't get what I mean and you're getting hung up on nonsense. Everything is signed on the charts. In one case E is drawn, in the other the ratio E/E0. And similarly D, Y.
 
Dr.F.:


OK, I see that the missing equation, though denied, did appear. OK, that's some conversation.

But, excuse me, what makes you think that the "true values" calculated on 150 candlesticks will correlate with each other on other >9000 bars in the history? To check it, I suggest plot the missing history using the obtained ratios and see for yourself.

 
alsu:

OK, I see that the missing equation, though denied, did appear. Okay, that's some conversation.

Although, let me ask you, what makes you think that the "true values" calculated on 150 candles will correlate with each other on the remaining >9000 bars in history? To check it, I suggest plot the missing history using the obtained ratios and see for yourself.

Imagine that your chart is made up of thousands of pieces. Each piece has a certain strictly known form. A step down, for example. It's obvious to me without calculations that the entire graph would be a thousand steps down.

WHERE DID THE EQUATION APPEAR? 0_о

 
All right, I'm off for the day. We'll continue the tricks tomorrow or the day after. And then we'll move on to the bidding.
 
Dr.F.:
Imagine you have a graph made up of a thousand pieces. Each slice has a certain strictly known shape. For example, a step down. It's obvious to me without calculations that the entire graph would be a thousand steps down.


Repeat the calculations on a piece of history of 1500 and then 15000 candlesticks and see what correlation coefficients are obtained. My guess is that they will tend to zero as the sample grows.
 
So at EVERY moment all 3 currencies are moving in the same direction, but at different speeds?
 
I kind of figured it out. there is YY and we got YY from them, so we need to build YYY so that the KC between them was 1 (almost), but the additional condition is that the dollar is 144 bars ago, so the euro starts from 1.31, and the yen from 0.010815 approximately. right?
 
alsu:

Repeat the calculations on a piece of history of 1500 and then 15000 candles and see what the correlation coefficients turn out to be. My guess is that they will tend towards zero as the sample grows.

It's hard to make a more moronic assumption. Of course they can decrease slightly. Even down to 0.8. But at intervals of price changes within a day or two, the agreement of chart pieces in absolute figures will be as good as the one already shown. Do not get hung up on the value of correlations. It doesn't make any sense by itself and depends to some extent on the length of the averaging interval where you calculate them. That's not what it's about.
 
Dr.F.:

WHERE DID THE EQUATION APPEAR? 0_о


Come on now, it's not funny. Somehow you calculated the series, so you used at least arithmetic operations. And that's what formulas are made of, yes.
 
alsu:
Come on, it's not funny. Somehow you calculated the series, so you used at least some arithmetic. And that's what formulas are made of, yes.
Well, add the THIRD EQUATION to the studio and show me HOW I got the curves https://forum.mql4.com/ru/54199/page23