Correct calculation of currency indices. - page 15

 
faa1947:

And can you clarify what this dollar index means at the moment ?

For it is hard to prove that there is no black cat in a dark room, especially when it is not there :))

And please don't make flooder accusations of others being flooders.

 
BoraBo:

And can you clarify what this dollar index means at the moment ?

For it is hard to prove that there is no black cat in a dark room, especially when it is not there :))

And please don't make fluderal accusations of others of fludging.

I do not intend to explain anything to anyone. Read my posts. They are always with explanations and evidence. The same post above in the thread. I had hopes that there was some sense behind your words "about nothing", not justified.


Thank you for your attention to my posts.

 
faa1947:

I had hopes that there was some sense behind your words "about nothing", but that was not the case.


I can live up to your hopes, if you don't mind.

Let's imagine the following picture. At a certain period of time EURUSD rose by 5%. EURGBP, EURCHF, EURJPY and all other EUR pairs also rose by 5%. All pairs, which do not include EUR, remained in place. One would think, what has the dollar got to do with it? Wouldn't it be more logical to assume that EUR index has risen by 5%, while all the others have not changed? But the dollar index formula used all over the world somehow makes the whole world believe that the dollar index has changed and quite significantly. That is why I also think it is about nothing.

The only thing that can justify its existence is the possibility of trading this index. But that trade smells bad as the subject of the trade itself is unclear to me, it seems as if it does not exist at all, such financial instruments belong in a casino.

 
AlexeyFX:


I can live up to your expectations, if you don't mind.

Let's imagine the following picture. At some point in time the EURUSD went up by 5%. EURGBP, EURCHF,EURJPY and all other EUR pairs also rose by 5%. All pairs, which do not include EUR, remained in place. One would think, what has the dollar got to do with it? Wouldn't it be more logical to assume that EUR index has risen by 5%, while all the others have not changed? But the dollar index formula used all over the world somehow makes the whole world believe that the dollar index has changed and quite significantly. That is why I also think it is about nothing.

The only thing that can justify its existence is the possibility of trading this index. But that trade smells bad as the subject of the trade itself is unclear to me, it seems as if it doesn't exist at all, such financial instruments belong in a casino.

I cannot agree.

Let's take an ordinary TS. It works on EURUSD. There are a lot of such TSs. We try to extract and use some information from lag values of this pair, but all information is within the pair.

Plenty of traders who talk about multicurrency is an attempt to get some additional information outside the EURUSD pair.

The dollar index is an attempt to attract some kind of information that is not in the eurodollar pair. In contrast to the stock market in forex we have very limited additional information to build a more complete model.

We have practice. In my thread I showed that the model on the index is about 1.5 times more profitable than on the Eurodollar, which confirms the above reasoning.

Now about the calculation.

Above I gave the divergence between the traded index and the calculated index. Which is better? I do not know. Purely practical considerations prevail here. On H1 there are essentially no quotes for three hours each. But it does not matter. Let's take the computable index. This is a multi-currency index. And the selection of proportions between the currencies is not spontaneous, but somehow reasonable. Can we make a model that involves each of the pairs directly and select the ISC coefficient? I doubt it. The question of choosing regressors is rather tricky, and here there is a market opinion - that's something.

 
faa1947:

I cannot agree.

Let's take an ordinary TS. It works on EURUSD. There are a lot of such TS. We try to extract and use some information from lag values of this pair, but all information is within the pair.

Plenty of traders who talk about multicurrency is an attempt to get some additional information outside the EURUSD pair.

The dollar index is an attempt to attract some kind of information that is not in the eurodollar pair. In contrast to the stock market in forex we have very limited additional information to build a more complete model.

We have practice. In my thread I showed that the model on the index is about 1.5 times more profitable than just on the Eurodollar, which confirms the above reasoning.

Now for the calculation.

Above I cited the divergence between the traded index and the calculated index. Which is better? I do not know. Purely practical considerations prevail here. On H1 there are essentially no quotes for three hours each. But it does not matter. Let's take the computable index. This is a multi-currency index. And the selection of proportions between the currencies is not spontaneous, but somehow reasonable. Can we make a model that involves each of the pairs directly and select the ISC coefficient? I doubt it. The question of choosing regressors is quite a tricky one, and here there is a market opinion - that's something.


Even from this perspective, using a common index formula does not seem sensible. It is clear that additional information from other pairs should increase profitability, and it is probably not worth the trouble to prove the obvious. But why should exactly THIS formula be used? It takes into account only 6 currencies and the validity of this accounting is questionable for me, although indices can be further used in different ways, someone may do so... But you can calculate as many as 20 currencies or as many as 200. Or do you think that currencies which are not included in the known formula do not contain useful information?
 
faa1947:

You take the ratios in your calculations based on the main US foreign trade turnover as of March 1973.

Hence we have the relative price of the dollar as if the structure of US foreign trade turnover had not changed since 1973.

But since those bearded years, the USA has moved more than half of its production to Asia, because labour is much cheaper there. China has emerged and become the world's second economy. And its share of US trade is ~11%.

That's why I think this index is about nothing.

And you just calculated, based on more than dubious assumptions, your index and compared it with the index the DC provides you with.

The result is a stable difference of ~7%, and if it helps you in trading, it is absolutely indifferent which one to use, because the relative movement of both indices is absolutely identical.

However, both of these indices are not correct, because it is absolutely unclear, that in the current economic situation reflects the US foreign trade turnover for March 1973.

PS

And although you don't want to explain, you only calculated your index, but I would like to hear, what does this index show in your understanding?

 
BoraBo:

This formula shows :

The calculation of the dollar index (USDX) by a basket of six currencies is not accidental - it coincides with the data used by the US Federal Reserve in calculating the trade-weighted dollar index by the currencies of those countries which form the main US foreign trade turnover. Most of US international trade is with the Eurozone (57.6%), followed by Japan (13.6%), the UK (11.9%), Canada (9.1%), Sweden (4.2%) and Switzerland (3.6%).

And it was derived a hell of a long time ago. Since then, the US foreign trade ratio has changed considerably (e.g. China stormed in like a typhoon). So this formula, at the moment, is about nothing.

Coefficients can be taken from SDR, it is calculated quite often.

https://forum.mql4.com/ru/40230

https://forum.mql4.com/ru/39417

 
AlexeyFX:

Even from this perspective, using a generally accepted index formula does not seem sensible. It is clear that additional information from other pairs should increase profitability, and it probably wasn't worth the trouble to prove the obvious. But why should exactly THIS formula be used? It takes into account only 6 currencies and the validity of this accounting is questionable for me, although indices can be further used in different ways, someone may do so... But you can calculate as many as 20 currencies or as many as 200. Or, do you think that currencies that are not included in the known formula, do not contain useful information?

Contains.

But the usefulness of the index is that it is traded as such. Like an index on a stock exchange. The Dow is made up of 30 issuers, which is a very representative index. Representativeness is determined not only by the formula, but also by the number of people who take note of the value.

 
BoraBo:

In your calculations you take the coefficients based on the main US foreign trade turnover as of March 1973.

The euro came into existence in 2000, not 1973.

And although you don't want to explain, you only calculated your index,

Above I gave not only the formula, but also the result of the calculations in the hope of hearing at least someone's opinion.

 
faa1947:

Contains.

But the usefulness of an index is that it trades as such. Like an index on a stock exchange. The Dow has 30 issuers and is a very representative index. Representativeness is determined not only by the formula, but also by the number of people who take note of the value.


Well, if you need to take this index and these people into account (you should check this need first), you can take it into account in the same way as you would take currency pairs into account.