Absolute courses - page 37
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Everything is fine. In the evening I will show the graphs with the following results (on the computer where I have only the outdated semi-manual algorithm so I want to wait at home):
for file 1: will show E, D, Y correlated with each other with a coefficient of about 0.97, with EDx, EYx, DYx representing E/D, E/Y, D/Y respectively correlated corr(ED, EDx) = 0.9999+ and similarly EY, EYx and DE, DYx
for file 2: will show E, D, Y correlated with each other with a coefficient of about 0.998, with EDx, EYx, DYx representing E/D, E/Y, D/Y respectively correlated corr(ED, EDx) = 0.9999+ and similarly EY, EYx and DE, DYx.
These numbers are provisional as the algorithm is normally prescribed on another computer. But the conclusion from them will be:
file 1: real.
file 2: HSPF.
Why? Because the PRNG signal is not internally connected in any way and E, D, Y can correlate with each other as 0.998 even with my semi-manual algorithm now. So in the limit it can go up to 1. But on the real data the limit is further away from 1. I think it may go up to 0.99, but no more. We will have to see in the evening. Or maybe it'll stay like 0.97. And this difference is the thin effect that determines the forms of the RELATIONShips ED, EY, DY. In the case of real quotes there is a limit (sum corr(E,D)+corr(E,Y)+corr(D,Y) cannot reach 3. But in the case of HSPF it can. For there are no internal hidden connections. So it goes like this. I'll explain it in more detail tonight.
contrariwise)) First real (M15 EURUSD and USDJPY from 10/02/2009 (00:00)). The transformation is simple (divide by the first term of the series i.e. by the value at 00:00)
second one is gpc
The real ones:
why so early again?
Everything is fine. In the evening I will show the graphs with the following results (on the computer where I have only the outdated semi-manual algorithm so I want to wait at home):
for file 1: will show E, D, Y correlated with each other with a coefficient of about 0.97, with EDx, EYx, DYx representing E/D, E/Y, D/Y respectively correlated corr(ED, EDx) = 0.9999+ and similarly EY, EYx and DE, DYx
for file 2: E, D, Y will be shown to correlate with about 0.998, with EDx, EYx, DYx, representing E/D, E/Y, D/Y respectively, corr(ED, EDx) = 0.9999+ and similarly EY, EYx and DE, DYx.
These numbers are provisional as the algorithm is normally prescribed on another computer. But the conclusion from them will be:
file 1: real.
file 2: HSPF.
Why? Because the PRNG signal is not internally connected in any way and E, D, Y can correlate with each other as 0.998 even with my semi-manual algorithm now. So in the limit it can go up to 1. But on the real data the limit is further away from 1. I think it may go up to 0.99, but no more. We will have to see in the evening. Or maybe it'll stay like 0.97. And this difference is the thin effect that determines the forms of the RELATIONShips ED, EY, DY. In the case of real quotes there is a limit (sum corr(E,D)+corr(E,Y)+corr(D,Y) cannot reach 3. But in the case of HSPF it can. For there are no internal hidden connections. So it goes like this. I'll explain it in more detail tonight.
You're blowing it so early again.
don't worry. i'll get home, do a proper analysis, post pictures here, and it seems everything will be ok.
if in doubt, i would be happy to guess one real one among dozens of hspc. get your files ready.
Why are you burning again so early?
Well, the author has conducted an analysis and announced his conclusions - why wait?)
Well, you are making up explanations as you go along)) Making speculative conclusions about internal incoherence, etc. and how it affects the correlation of series obtained by who knows what algorithm. All in all, it's clear :)
Of course I'm making it up as I go along. I explain what I see. How else could I? This is normal research. I think I've already figured out the reason for my initial error. If I post data in the evening that will repeat what I have now said and you have quoted, then, I repeat:I will gladly guess one real one among a dozen hspc. prepare the files. Evening MSC, 9-12 o'clock.
don't worry. i get home, do a proper analysis, post pictures here, and it seems everything will be ok.
if in doubt, i would be happy to guess one real one among dozens of gspc's. prepare the files.
well it's understandable that you'll be able to guess at some point))
(It's understandable that you'll guess at some point))
Once again: I show the pictures, make a conclusion. Then I guess 1 out of 10 on the first try. Do the files by evening. And better 288 bars, for in short sequences hidden patterns are harder to detect. With a sequence length of 2-3 bars it's impossible at all, it's clear.
Once again: I show the pictures, make a conclusion. Then I guess 1 out of 10 on the first try. Do the files by tonight.
I'm done - you want to do it yourself. I've got it all figured out
I can do it myself. The topic is interesting, about differentiating between quotes and gpsh. My algorithm probably can do it, although it was not planned. I will post the general material.