From theory to practice - page 744

[Renat Akhtyamov]

Alexander_K:

For those who have nothing better to do, post CLOSE M1 data for any year for any pair in .csv format.

Preferred file name, for example: EURGBP 2017 CLOSE M1.csv

I will check them against Variance Gamma Process and post the results here.

Thank you.

that's fine

in person

[Alexander_K]

Renat Akhtyamov:

That's fine.

in private

Thanks. But, I need data specific to CLOSE M1.

I can do it myself, but I have almost no time - work, worries... For some reason Finam only lets me download it in 2 months - then I have to 'merge' it. Long and tedious...

Maybe someone has ready-made archives?

[Yuriy Asaulenko]

Alexander_K:

For those who have nothing better to do, post CLOSE M1 data for any year for any pair in .csv format.

Preferred file name, for example: EURGBP 2017 CLOSE M1.csv

I will check them against Variance Gamma Process and post the results here.

Thank you.

Any quotes, any instrument, on any timeframe, for any year and month, in any quantity.)

And do not burden people with any bullshit.))

https://www.finam.ru/profile/forex/eur-usd/export/?market=5&em=83&code=EURUSD&apply=0&df=15&mf=10&yf=2018&from=15.11.2018&dt=15&mt=10&yt=2018&to=15.11.2018&p=7&f=EURUSD_181115_181115&e=.txt&cn=EURUSD&dtf=1&tmf=1&MSOR=1&mstime=1

[Alexander_K2]

Only on Dukas did I find what I was looking for: 5-digit Bid and Ask quotes.

Here you go, it may come in handy

[Renat Akhtyamov]

Alexander_K:

Thanks. But, I need data specific to CLOSE M1.

I can do everything myself, but I have almost no time - work, worries... For some reason Finam only lets me download it in 2 months - then I have to "glue" it together. Long and tedious...

Maybe someone has ready-made archives?

Not happy with MT?

Export to csv is a standard feature of the terminal.

[Alexander_K2]

Now it's my turn to demonstrate how to beat SB.

Take the Laplace motion SB and plot two graphs for it:

1. Laplace motion itself with dispersion channel and MA

2. The sum of increments in the sliding window with the variance channel

As can be seen, the first chart has no real advantages, at best it is 0.

And on the second one we have a convincing profit.

Conclusion:

We should convert the initial BP to the stationary form by all means, and then it is a matter of technique.

[Yuriy Asaulenko]

Alexander_K2:

I can do everything myself, but there's hardly any time - work, worries... For some reason Finam only lets you download it in two months - then you have to "glue" it together. It is long and tedious...

I`ve tried to check it. Everything is downloaded from Finam. 2 months is a load of bullshit. Or your hands are growing in the wrong place.

<TICKER>,<PER>,<DATE>,<TIME>,<OPEN>,<HIGH>,<LOW>,<CLOSE>,<VOL>
EURUSD,1,01/01/18,01:01:00,1.1996600,1.1996600,1.1996600,1.1996600,1
EURUSD,1,01/01/18,09:02:00,1.2002000,1.2002000,1.2002000,1.2002000,1
.....
EURUSD,1,15/11/18,23:59:00,1.1331900,1.1332200,1.1330000,1.1331800,815
EURUSD,1,16/11/18,00:00:00,1.1331400,1.1332700,1.1330000,1.1331800,1015
EURUSD,1,16/11/18,00:01:00,1.1332500,1.1334200,1.1331000,1.1333400,1327

It's true, for this case there's no need to glue anything, but - Long glue? No words.

Maybe there's something in the conservatory that needs tweaking?

[Alexander_K2]

In short - Novaja was right - you have to make the distribution of the increments have the classical form - Laplace, Gauss, ...

How do we do that? I don't know. Finita la comedia.

[Yuriy Asaulenko]

Alexander_K2:

In short - Novaja was right - you have to make the distribution of the increments have the classical form - Laplace, Gauss, ...

How do we do that? I don't know. Finita la comedia.

Quietly as the slate rustles, the roof goes slowly.

We have a time series, it has increments and the increments have distributions - all these are characteristics of the given BP.

We read - "it is necessary to make the distribution of increments have a classical form - Laplace, Gauss,...". which is equivalent to the following - we have to make the initial BP to be different.

We do not and will not have another one, we have only this one.))

You can, of course, adjust the initial BP by transformations to your wishes, but, after that, it will have nothing left except your wishes, and it will be simply useless for something else.

[Дмитрий]

Alexander_K2:

In short - Novaja was right - you have to make the distribution of the increments have the classical form - Laplace, Gauss, ...

How do we do that? I don't know. Finita la comedia.

Google - converting a time series to a stationary form