Correct calculation of currency indices. - page 5

 
voidpiligrim >>:
а как у вас корень из -1 получился? У меня только положительные числа в формулах.

It's not me, it's MathLab .
If you look at IndexesuvyMathLab.txt (from 06.02.2009 20:32) the first solutions for each Index
I've created an indicator based on it, but it's not ready yet.
I have sketched out an indicator on its motives but it is still raw and I don't have time to improve it. The idea on IndexKor is embedded in the code itself. The code is attached.

Files:
indexkor.mq4  8 kb
 

Here's my example calculation (SQL code), if you're interested:

SELECT
A1.DTM AS DTM
,POWER (1/(A1.MO*A3.MO*A8.MO*A12.MO*A13.MO),0.16666666666666667) AS CAD
,POWER ((A1.MO*A2.MO)/(A4.MO*A9.MO*A14.MO),0.16666666666666667) AS CHF
,POWER ((A3.MO*A4.MO*A5.MO*A6.MO*A7.MO),0.16666666666666667) AS EUR
,POWER ((A8.MO*A9.MO*A10.MO*A11.MO)/(A5.MO),0.16666666666666667) AS GBP
,POWER ((A12.MO)/(A2.MO*A6.MO*A10.MO*A15.MO),0.16666666666666667) AS JPY
,POWER ((A13.MO*A14.MO*A15.MO)/(A7.MO*A11.MO),0.16666666666666667) AS USD
,POWER (1/(A1.MO*A3.MO*A8.MO*A12.MO*A13.MO),0.16666666666666667)
*POWER ((A1.MO*A2.MO)/(A4.MO*A9.MO*A14.MO),0.16666666666666667)
*POWER ((A3.MO*A4.MO*A5.MO*A6.MO*A7.MO),0.16666666666666667)
*POWER ((A8.MO*A9.MO*A10.MO*A11.MO)/(A7.MO),0.16666666666666667)
*POWER ((A12.MO)/(A2.MO*A6.MO*A10.MO*A15.MO),0.16666666666666667)
*POWER ((A13.MO*A14.MO*A15.MO)/(A7.MO*A11.MO),0.16666666666666667) AS MUL
,A1.MO AS CHFCAD
,A2.MO AS CHFJPY
,A3.MO AS EURCAD
,A4.MO AS EURCHF
,A5.MO AS EURGBP
,A6.MO AS EURJPY
,A7.MO AS EURUSD
,A8.MO AS GBPCAD
,A9.MO AS GBPCHF
,A10.MO AS GBPJPY
,A11.MO AS GBPUSD
,A12.MO AS JPYCAD
,A13.MO AS USDCAD
,A14.MO AS USDCHF
,A15.MO AS USDJPY

POWER is a power increase.
 
Urain писал(а) >>

USDX = 50.14348112 *(GBPUSD, -0.133)*(EURUSD, -0.59 )*(USDCHF, 0.05 )*(USDJPY, 0.136)*(USDCAD, 0.091)

expression ( x,y ) means x to the power of y.

EURx=EURUSD/USDX

GBPx=GBPUSD/USDX

as a result, EURGBP does not always equal EURx/GBPx ???

I wonder how you calculate, for example, AUDUSD... Your formula takes quotations of 5 currency pairs as the basis...

Or for example how would the dollar index look like just for the major currencies... If you can give me the formula...

 

More currencies can be added, but then a system of more equations must be solved. Formulas will be similar but longer, because the normalization equation will be CAD*AUD*NZD*USD*GBP*EUR*JPY*CHF=1. Once again, in these formulas weights of currencies are proportional to absolute values of their crosses. To equalize the weights I take the product of moving average ratios.

It can be simpler: AUD = AUDUSD*USD, where USD is calculated using the previous formula.

 
kharko >> :

I wonder how you calculate, for example, AUDUSD... Your formula is based on quotations of 5 currency pairs...

Or, for example, what would the dollar index look like just for the major currencies... If you can give me a formula...

In the process of discussion on the forum I gave up
formula: USDX = 50.14348112 *(GBPUSD, -0.133)*(EURUSD, -0.59)*(USDCHF, 0.05)*(USDJPY, 0.136)*(USDCAD, 0.091),
And I have changed to formula offered voidpiligrim 05.02.2009 23:37 (with normalized equation). As a result I got a system laid out (Urain 06.02.2009 17:43), on the principle of which I can calculate a system for any
number of tools.

 
Urain писал(а) >>

In the course of discussion on the forum I have abandoned
formula: USDX = 50.14348112*(GBPUSD, -0.133)*(EURUSD, -0.59)*(USDCHF, 0.05)*(USDJPY, 0.136)*(USDCAD, 0.091),
And I have changed to formula offered voidpiligrim 05.02.2009 23:37 (with normalized equation). As a result I got a system laid out (Urain 06.02.2009 17:43), on the principle of which I can calculate a system for any
on which one can calculate the system on any number of instruments.

sometimes it's useful to use the search engine. 'Please help in solving a system of 3 equations with three unknown values' I think you will find it interesting

 
Prival >> :

sometimes it's useful to use the search engine. 'Asking for help in solving a system of 3 equations with three unknown values' I think you will find it interesting

I agree, it's useful. But I'll cite the imperfection of the search engine, since there were no answers to my enquiries.

By the way, one could give a reference at the beginning of a branch and discussion would pass to a branch "I ask for the help in the decision of system 3",
>>and it wouldn't be multiplied by budding (you participated there too and knew about similarity of discussed topics).
in any case it is a question from a category "if only". With respect Urain.


 

Perhaps the above formulas could be called currency index calculations... Which I doubt....

But your desire to get quotes for any currency pair by dividing currency indices is a mistake...

You have forgotten what a currency index is for... Let me remind you...

Exchange index (another name - stock index) - is a composite index of price changes of a certain group of assets(securities, commodities, derivatives) - "index basket".

In our case it is a set of currency pairs or instruments that have a common currency... For example, what prevents us from using quotations for metals in dollar index calculation...

What should the formula for calculating the dollar index look like?

VALUTA=(K1*VAL1/USD+K2*VAL2/USD+...+Kn*VALn/USD)/n

USDx = 1/VALUTA

where,

VALUTA - weighted average currency;

K1...Kn - weights coefficients;

n - number of instruments used in the calculation...

This formula is universal... We can use from 1 to nth number of instruments...

 
Urain >> :

The system has been in use for some time now, but there are not enough resources to get an accurate forecast (even though it's on the 4th stump, which is not the oldest trough),

and when the settings get too tight "we're throwing the baby out with the soapy water".

My personal experience is that the extrapolator starts to lie when USDX starts moving more than usual

i.e. USDx was relatively calm and its influence on other currencies was perceived
extrapolator sees it as noise (which can be ignored) and then USDx wakes up and "tails it".
And if you extrapolate the USDx, the lull is deciphered as a frequency beat,
and a wake-up is predicted at a moment's notice...

It would be interesting to hear more about your vision, if you don't mind

 
Urain писал(а) >>

I agree, it's helpful. But I'll cite the imperfection of the SEARCH, since there were no answers to my enquiries.

By the way, you may have provided a link at the beginning of the branch, and the discussion would have moved to the branch "I want help in solving the system of 3...",
In any case, it is a question of "How to help me solve the system of 3..." and not to multiply by kidnapping (you participated there too, and were aware of the similarity of the topics under discussion).
in any case it's a question from the category of "if only". Respectfully Urain.

It seems to me that we have helped you, or tried to help. You have now understood that the index can be calculated in different ways (that's why we asked you for the formula). When you saw it, you tried to explain its advantages and disadvantages. I was hoping that I would see something new, some interesting approach to the construction. But there is always the question of the purpose of these changes. They seem to have also touched on it (the aim), that it is better to predict the index. It may well be.

The subject of indices (clusters, triangles, etc.) periodically arises here, and somehow quickly dies out. They made a curve, but what next? They tried to predict the movement of one curve, and now there's another one with different properties. For example, it would be very interesting to me, if you have built an indicator, what to do with it afterwards? How to work with it?

Reason: