From theory to practice - page 144

 

Anyone interested - do your own processing of AUDCHF data (see attached file), as I did earlier for AUDCAD. Look at the results, just don't fall off your chair like I did.

Files:
AUDCHF.zip  331 kb
 
Alexander_K2:

Anyone interested - do your own processing of AUDCHF data (see attached file), as I did earlier for AUDCAD. Look at the results, just don't fall off your chair like I did.


There are no time stamps again. Is this already exponent sampling or do I have to do it?

 
СанСаныч Фоменко:

There are no time stamps again. Is this already sampling exponentially or should we do it?

It's already exponential. I will post all 20 pairs when ready.
Files:
AUDJPY.zip  362 kb
AUDUSD.zip  316 kb
CADJPY.zip  390 kb
CHFJPY.zip  414 kb
EURAUD.zip  458 kb
EURCAD.zip  464 kb
EURCHF.zip  356 kb
EURJPY.zip  456 kb
EURGBP.zip  356 kb
EURUSD.zip  395 kb
GBPAUD.zip  480 kb
GBPCAD.zip  467 kb
GBPCHF.zip  413 kb
GBPJPY.zip  473 kb
GBPUSD.zip  386 kb
USDCAD.zip  367 kb
USDCHF.zip  345 kb
USDJPY.zip  373 kb
 

The Physicists along with the Chemists will tear up the forum!
As well as their Nerds! )))

What are you guys up to? Has the forum been closed already? Because I went to the terminal today, there's still some chart hanging around. )))

 
Максим Дмитриев:

The Physicists along with the Chemists will tear up the forum!
and their Nerds! ))))

What are you guys up to? Has the forum been closed already? Because I went to the terminal today, there's still some schedule dangling there. )))

:)))

No, I can hardly do it by myself. But if a million traders will start to pump the forex by grail, then everything is possible. Well, good riddance! Let's go back to the Soviet Union, which is exactly what I want.

 

To the question of "Markovian - not Markovian.

Here's the autocorrelation.

Acf(V1_diff[V1_diff!=0], lag=100000)

We see that there is a significant correlation between ticks even above 60,000. And in the first 40,000 ticks the correlation is almost solid.

Although the distribution looks nice.


It's the usual, why the uneven scale? Why the ticks?

I don't get it. The trading idea itself is to lose when the trend corrects sideways...

 

One other thing. A very common one is cyclicality for absolute values of increments.


I've already posted it once for watchmakers - it looks very similar. How to use it is not clear

 

Guys, it's time for practice.

It's about time.

 
Alexander Ivanov:

Guys, it's time for practice.

It's time already.


it will be after the new year)

 
СанСаныч Фоменко:

One other thing. A very common one is cyclicality for absolute incremental values.


I've already posted it once for watchmakers - it looks very similar. How to use it is not clear

Chop minima/maxima (read: tails) in observation window (sample size) = 99.5% confidence level or more.
Reason: