How to minimise index correlation - page 2
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
If you have a lag-free smoothing method, that means you know the future price values, it cannot be otherwise. This means that you already have the grail in your hands, and you do not need any indices or indicators.
That's the thing, some people argue that multi-analysis is secondary. they claim that their anticipation is derived from one pair, and multi-analysis only serves for more favourable entry conditions for one instrument or another.
But it seems to me that this very anticipation is the consequence of multianalysis, i.e. the anticipation is not obtained for pairs separately, the anticipation can be obtained only from multicurrency.
The inertia proper appears not in the price itself, but in the filter coefficients, but this inertia cannot be obtained for each individual pair, only in multicurrency analysis.
Why the curve? The point is not in the index calculation, but in the question of how to make the orthogonality (zero correlation) of currency indices at each successive calculation.
There are many things in the database. You have a redundant system: for 8 currencies you can find a brokerage company with quotes for 27 currency pairs. This is a defect of the method. Hence the "veers" and "correlations" that don't exist. You invent them, count them and then get rid of them (minimise them).
The task is much easier if you realise that not only do you not know the exact values of the crosses, but also the time to which these unknown values are known to you. Therefore, any exact solution methods will only increase the errors. You need generalized methods for solving this redundant system (there is no exact solution). You can find many variants in the Code Base. And the "geometric mean" is not taken there from scratch.
Then if the answer to the question "What for?" is "the indexes are essentially needed to select better tools". Probably you won't get it.
But indexes allow you to form a basket of instruments for trading, which is called "synthetics".
There are many things in the base. You have a redundant system: for 8 currencies you can find a DC with quotes for 27 currency pairs. This is a flaw in the method. Hence the "veers" and "correlations" that don't exist. You invent them, count them and then get rid of them (minimise them).
The task is much easier if you realise that not only do you not know the exact values of the crosses, but also the time to which these unknown values are known to you. Therefore, any exact solution methods will only increase the errors. You need generalized methods for solving this redundant system (there is no exact solution). You can find many variants in the Code Base. And the "geometric mean" is not taken there from scratch.
Then if the answer to the question "What for?" is " then the indexes are essentially needed to select better tools". Probably you won't get it.
But indexes allow you to form a basket of instruments for trading, which here is called "synthetics".
what the hell is a redundant system? why would more tools in the calculation be a defect? how is it that there are no fans and correlations?
the standard formula from here https://www.mql5.com/ru/forum/114579/page18 (first post) can't it be applied to a basket of tools?
The problem is that the wands need to be replaced with something else.
what the hell kind of redundant system is this? why is more tools in the calculation a defect? how is it that there are no fans and correlations?
the standard formula from here https://www.mql5.com/ru/forum/114579/page18 (first post) can't it be applied to a fan of tool masks?
the problem is that the wands need to be replaced with something else.
That's what I mean, the volatilities are different, therefore the amplitude characteristics are different, that's why we compare, say, these 2 sinusoids (example)
not their absolute values, but the coefficients under the functions. i.e. phase changes.
MetaDriver has clearly, mathematically, prescribed the construction of its own index, but it seems to me that the scale of prices should also be taken into account. Otherwise, for example, JPY movements are completely lost against the background of heavyweights.
There is a "magic word". rationing.
That's the thing, some argue that multi-analysis is secondary.
Maybe the multicurrency (index) is more stationary?
is that a statement, or what? I don't think the index is more stationary.
There's a 'magic word' - rationing.
is that a statement or what? I don't think the index is stationary.
Null Hypothesis: EURUSD has a unit root
Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=34)
t-Statistic Prob.*
Augmented Dickey-Fuller test statistic -1.464364 0.5519
Test critical values: 1% level -3.431160
5% level -2.861783
10% level -2.566941
.
.
.
Null Hypothesis: DX (dollar index) has a unit root
Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=14)
t-Statistic Prob.*
Augmented Dickey-Fuller test statistic -2.228522 0.1968
Test critical values: 1% level -3.457865
5% level -2.873543
10% level -2.573242
The probability that the series is non-stationary is highlighted. I.e. the pair Eurodollar is "more" non-stationary than the index, though the hypothesis that both are non-stationary cannot be rejected.
The index is taken in geometric weights. If we work with the weights, we might get a stationary index.