Econometrics: one step ahead forecast - page 27
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
We summarise the past 24 hours and make a forecast for the current day, 16 November.
Wrong prediction again yesterday.
We summarise the past 24 hours and make our forecast for the current day, 16 November.
Wrong forecast again yesterday.
Yes ((( Forecast. So without going into details of the system I want to continue, because the topic is interesting.
If i look at the strategy, i can guess that trading goes not through 1000 pairs, but only those where the spread is the smallest. I think that we will gain 7 pairs at most. Therefore, I would build a portfolio only of these pairs and somehow estimate the volumes of buying and selling. Perhaps it should be an indicator. Looking at its work, maybe one step forward ))))
If it goes that way, maybe we could try it.
Yep ((( Predictor. I don't want to go into details of how the system works, because the topic is interesting.
If we compare Forex with stock market, we can assume that trading goes not for 1000 pairs, but only for those where the spread - commission is the smallest. I think that we will gain 7 pairs at most. Therefore, I would compile a portfolio only of these pairs and somehow estimate the volumes of buying and selling. Perhaps it should be an indicator. Looking at its work, maybe one step forward ))))
If it goes that way, maybe we could try it.
There is a dollar index calculated on 6 pairs. You may try to make a prediction of EURUSD by the index or in plain terms: EURUSD by pairs included into the index. Anyway, the pairs included in the index were selected by thinking. This does not seem interesting to me. If you have any thoughts, I will try to implement it.
Can I have more details?
- index name
- source
- portfolio composition
I myself have chosen for my portfolio so far:
AUDUSD,EURCHF, EURGBP, EURJPY, USDCAD, USDCHF, USDJPY, EURUSD
I will try to put all volumes on one indicator )))
That's not what I'm saying. You have grasped at autocorrelation checking with a dead hand to remove the dependence in the residuals. And I'm saying that's not enough, because Pearson autocorrelation only explains linear dependencies, not all of them. In the branch about feature selection alexeymosc already gave an example when dependences calculated by information theory (not only linear, but all in a row!) were extremely high even at very large lags. The vast majority of participants in that thread, including its author, agreed that it was all about volatility. (In this model, by the way, there are no notions of trend/flat, although they can be drawn there as well).
I still don't see enough evidence to confidently say that it is the ox that is to blame. Perhaps on the days, yes, but I have said several times that the mutual information on the days is much less than on the watches or 4H. Almost nobody was interested in it, and all results were still only posted for days. So the conclusions were accordingly, i.e. incomplete.
Mathemat, faa1947, hello!
The conclusions are not complete, I agree. And I myself have noted with interest the different amount of mutual information on different timeframes. It's just that I already saw at intuition level that the result for H1 would be about the same as for the daily bars. And that would upset us even more. )
Alexei, your methodology, on forecasting, is good, just very good, but we applied it in the wrong place. My opinion is this.
And the linear dependence studies discussed in this thread, I think, can lead to a modest 12% return per year - at best, as is the case with the big players in the market, who all have excellent Harvard and Yale degrees in econometrics. It's up to the owner, basically.
Can you elaborate on that?
Wiki quote:
USDX is an index that shows the US Dollar against a basket of six major currencies: the Euro (EUR), Yen (JPY ), Pound Sterling (GBP), Canadian Dollar (CAD), Swedish Krona (SEK) and Swiss Franc (CHF ).
The index is calculated as a geometric weighted average of these currencies using the formula:
U S D X = 50.14348112 * U S D E U R0.576 * U S D J P Y0.136 * U S D G B P0.119 * U S D C A D0.091 * U S D E K0.042 * U S D C H F0.036,
where the power coefficients correspond to the weights of the currencies in the basket:
The dollar index (as a futures contract) is traded around the clock on the ICE: https://www.theice.com/productguide/ProductDetails.shtml?specId=194.
The first factor in the formula brings the index value to 100 on the start date - March 1973, when the major currencies began to quote freely against each other.
In the terminal it is DX
Mathemat, faa1947, hello!
And the studies of linear dependencies discussed in this thread, I think, can lead to a modest return of 12% per annum - at best, as is the case with the big players in the market, who all have excellent Harvard and Yale degrees in econometrics. It's the owner's business, in general.And studies of linear dependencies, which are discussed in this thread
Suggest non-linear ones
Can result in a modest return of 12% per annum - at best, as is the case with the big players in the market
It's the portfolio managers who are trying to outperform the index. Econometrics has almost nothing to do with them. It can be used to calculate risks for portfolios, but no more than that.
A quote from the Wiki
The index is calculated as a geometric weighted average of these currencies using the formula:
U S D X = 50.14348112 * U S D E U R0.576 * U S D J P Y0.136 * U S D G B P0.119 * U S D C A D0.091 * U S D E K0.042 * U S D C H F0.036,
What is the average price taken as the basis? I do not see DX in the terminal.
The formula is not quite clear to me as the reasoning on the weight of currencies has emerged. How true is this percentage inclusion in a basket of currencies?
Let us assume that the Forex market is a virtual one, so let us eliminate the weights and assume that the weights will be provided by the volumes of buying and selling. Let's assume that the price is dictated in such a way that the more profit the better. Probably, we can say from this that it is impossible to forecast the price in terms of historical data for a long period because the market situation is unstable due to unpredictable behavior of traders. For example, in 1 hour they buy by 1000 and sell by 100. Then the trader should change the quote in the direction of falling of the rate. The trader's actions in this case are predicted by closing losing trades, in any case))) Then he will surely earn, but those who know the behaviour and predict the price from the quote will also earn, won't they?
Very much like supply and demand? But it is a market, even if it is virtual.
it is possible to make mathematically accurate and verified forecasts even 1,000 (a thousand) steps ahead...
In terms of the quotation movement function, they (the forecasts) will fit perfectly into the theory.
With the development of high technology forex managers are struggling with the introduction of news.
The news will confuse any neuronet or any high-tech algorithm.
And there is nothing you can do about it.