neural network and inputs - page 3

 
Demi:



Bullshit.

Correlation is a pattern.

This is not nonsense, it is a typical mistake, and the proof is one and very clear - at 100% correlation there is no point in searching for a pattern, because one tool becomes exactly another tool. Why use this, the other tool, if it is the same?

In case you perceive the correlation as a tool, then there is another counter-argument - why correlate throughout, if it is enough to correlate only at certain points, where the inputs should be? And even in this case, the correlation will also take a minimum value....)))

 
alsu: It may well be that there is a correlation and zero correlation.

This is exactly what happens in the financial markets, just not everyone realises it ))))
 
LeoV:

This is not nonsense, it is a typical mistake, and the proof is one and very clear - at 100% correlation there is no point in searching for a pattern, because one tool becomes exactly another tool. Why use this, the other tool, if it is the same?

In the case of your perception of correlation as a tool, then there is a counter-argument - why correlate throughout, if correlation only at certain points where there should be inputs is sufficient? And even in this case, the correlation will also take a minimum value....)))

There is no mistake here - there is no 100% correlation in forex.

High correlation - divergence - opening a position, etc. Then the correlation coefficient went down and the TS died.

And sometimes in financial markets the correlation coefficient has changed sign

 
Demi: Over time the correlation coefficient has decreased and the euro-dollar-index correlation has increased.

All this has been known for a long time.

It wasn't about the euro-dollar-index....))))

And by the way, at that time, I checked, for the sake of interest, the correlation between euro-dollar and euro-yen was minimal, as it will be no surprise to you. You can check it all yourself - the data is all there ))))

 
Demi:
There is no mistake - there is no 100% correlation in forex.


Well fuck it, the closer you get to 100%, the bigger the error and the missing the point of searching for patterns..... How's that? ))))
 
LeoV:

It wasn't about the euro-dollar-index....))))

And by the way, at the time, I checked, for the sake of interest, the correlation between the euro-dollar and the euro-yen was minimal, as will not be surprising to you. Yes you can check it all yourself - the data is all there ))))



Show me the result.

The correlation coefficient EUR USD and EURJPY was "minimal"? How much is that? I can still understand EURUSD and JPYUSD, but this................

 
Demi: High correlation - divergence - opening a position, etc. Then the correlation coefficient went down and the TS went dead.


No, it didn't. Divergence is not a pattern on which you can make money.
 
Demi:
Show me the result.

The EURUSD and EURJPY correlation coefficient was "minimal"??? How much is that? I can still understand EURUSD and JPYUSD, but this................


In my memory, if I'm not mistaken, 0.3-0.4. So you can check it yourself - you can download the whole story and calculate.....)))
 
LeoV:

No, it isn't. Divergence is not a pattern on which you can make money.



Yeah sure it is! Pair trading to the scrapheap etc. etc.

You, of course, know better. (that's sarcasm)

 
LeoV:

In my memory, if I'm not mistaken, 0.3-0.4. So you can check it yourself - you can download the whole story and count it.....)))
Show me the result.

The EURUSD and EURJPY correlation coefficient was "minimal"??? How much is it? I can still understand EURUSD and JPYUSD, but this................
Reason: