![MQL5 - Language of trade strategies built-in the MetaTrader 5 client terminal](https://c.mql5.com/i/registerlandings/logo-2.png)
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
So I'll go on to finish the thought.
The value R squared is suspiciously small. Let's ask the question, are our variables related?
It was shown above that the distribution law is far from normal - it makes no sense to use Pearson correlation.
Look at the cointegration:
we will form a cointegration equation. It has the form:
OPEN_INTEREST = C(1)*LONG_IN_OI + C(2) + C(3)*@TREND
Substituted Coefficients:
=========================
OPEN_INTEREST = 61282.4785072 *LONG_IN_OI + 144744.044992 - 211.18145894 *@TREND
The graph of the residual from the cointegration:
Most likely stationary. Let's check it just in case:
Null Hypothesis: RESID01 has a unit root
Exogenous: Constant, Linear Trend
Lag Length: 13 (Automatic - based on SIC, maxlag=18)
t-Statistic Prob.*
Augmented Dickey-Fuller test statistic -4.467506 0.0018
We can see that the probability of the residue being non-stationary is less than 2% - it means that the residue is stationary.
This suggests that arbitrage is possible.
But caution should be exercised. Let's answer the question if there is a causal relationship between the Open Interest and the Longs
Let's perform the Granger causality test:
Pairwise Granger Causality Tests
Date: 07/30/12 Time: 19:36
Sample: 1,597
Lags: 2
Null Hypothesis: Obs F-Statistic Prob.
OPEN_INTEREST by Granger not cause LONG_IN_OI 595 2.01339 0.1345
LONG_IN_OI by Granger is not the cause of OPEN_INTEREST 0.34719 0.7068
Conclusion:
Putting open interest and longs into the same equation is probably not an option. Although there is some possibility for pair trading.
In general, the relationship between all columns is simple (2 formulas for calculating through aggregate long and aggregate short position):
OI = Noncommercial Traders Long + Noncommercial Traders Spreading + Operators Long + Non-reportable Long;
OI = Noncommercial Traders Short + Noncommercial Traders Spreading + Operators Short + Non-reportable Short;
faa1947:
It has been shown above that the distribution law is far from normal - using Pearson correlation makes no sense.
It was shown above that the distribution law is far from normal...
...
The last columns are probability...Read the first post of the thread carefully:
...
I hope we never have to calculate probability and normal distribution law in this forum again
...
Read the first post of the thread carefully:
Hope is the last to die © Folk proverbWhy, pardon my immodest interest?
Read the first post of the thread carefully:
Hope is the last to die © Folk proverbReshetov, as always in his repertoire: demonstrating that you don't understand anything at all, but you could.
San Sanych ... :)
San Sanych ... :)
If he understands, it's even worse. Pulling from different places from different contexts - why? What is the purpose of the post?
C-4 posted a real system based on many variables (very rare here), suggests discussing from a different perspective than before, there is a tool for many variables, interesting because....
Can you tell me more about:
1. Hedrick-Prescott filter - as far as I understand the approximating function is this particular filter. In the picture, it appears to be a red line marked 'Trend'. It is very similar to a moving average. It takes the difference relative to it and analyses the resulting residual - the green broken line below, which is also the blue line in the chart below. It is stationary, but it also seems to be heteroskedatic (the aplitude of oscillations is different) - it is not quite clear, aren't these mutually exclusive properties?
2. about the Granger causality test. - How is it calculated, at least in general terms, and what is the meaning.