Forget random quotes - page 54
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
Well obviously price is an independent function. Net positions of operators, speculators and small traders is a dependent variable. The net values are shown up to column 'H' (according to the Excel). Then come the calculated indicators. Correspondingly, they already depend on the net values of the operators, speculators and small traders.
Typical kind of "function":
Please specify with the file(table) name the names of the columns or their numbers. For example, like this:
table.XXXXXXX: col1 = col2, col5, col6, which means that data from col1 (may have a name) is calculated from data from col2,5,6. Can be from different tables, but then link the column to the table name
The names of the columns are given in the table headers. The columns that you are referring to are the columns (excel addressing) from 'A' to 'H' and from 'O' to 'Q'. All other columns are the values of the technical indicators calculated on the base values. I am going to briefly show you the values of the bars:
Open Interest - Open Interest
Noncommercial Long - Long positions of non-commercial traders (large speculators):
Noncommercial Short - Short positions of non-commercial traders (large speculators);
Operators Long - Long positions of commercial traders (hedgers);
Operators Short - Short positions of commercial traders (hedgers);
Nonrep Long - Long positions of Non-reporting traders (small speculators);
Nonrep Short - Short positions of non-accountable traders (small speculators).
Please pay special attention to the net positions of the group. This derivative value is calculated as the simple difference between long and short positions of a group, e.g. Net Operators = Operators Long - Operators Short, etc. It is considered to be an important ratio with predictive power.
There are at least some issues I'd like to clarify.
Obtained data trend is superimposed on the time axis, although the stationarity of these data (time dependence, seasonal factors) is not proven. Rather, there is a clear relationship with the price: price up - operators positions up, price down - operators positions down. What is the leading factor here and what is the slave and we need to find out. And we would also like to see tests on the predictive power of these data. It has been repeatedly observed that extreme values of the participants' positions reached much earlier than the market breaks, i.e. in these cases they had predictive power.
The names of the columns are given in the table headers. The columns that you are referring to are the columns (excel addressing) from 'A' to 'H' and from 'O' to 'Q'. All other columns are the values of the technical indicators calculated on the base values. I am going to briefly show you the values of the bars:
Open Interest - Open Interest
Noncommercial Long - Long positions of non-commercial traders (large speculators):
Noncommercial Short - Short positions of non-commercial traders (large speculators);
Operators Long - Long positions of commercial traders (hedgers);
Operators Short - Short positions of commercial traders (hedgers);
Nonrep Long - Long positions of Non-reporting traders (small speculators);
Nonrep Short - Short positions of non-accountable traders (small speculators).
Please pay special attention to the net positions of the group. This derivative value is calculated as the simple difference between long and short positions of a group, e.g. Net Operators = Operators Long - Operators Short, etc. It is considered to be an important ratio with predictive power.
There are at least some issues I'd like to clarify.
Obtained data trend is superimposed on the time axis, although the stationarity of these data (time dependence, seasonal factors) is not proven. Rather, there is a clear relationship with the price: price up - operators positions up, price down - operators positions down. What is the leading factor here and what is the slave and we need to find out. And we would also like to see tests on the predictive power of these data. It has been repeatedly observed that the extreme values of the participants' positions were reached much earlier than the market break, i.e. in these cases they had a predictive power.
Let's begin.
Although the task you set before me is not very clear to me.
What are we trading?
I will do some calculations and you comment.
Let's begin.
Although the task you set before me is not very clear to me.
What are we trading?
I'll do some calculations and you comment.
That would be interesting...
Here is a sample TS plan for entering the market on the kinks in market operators' net positions: "Master Class "Using CFTC reports in trading". Sorry if this is a little off-topic.
Let's begin.
Although the task you have set for me is not very clear to me.
Open Interest - Open Interest
Noncommercial Long - Long positions of non-commercial traders (large speculators):
Noncommercial Short - Short positions of non-commercial traders (large speculators) ;
Let's take the specified columns from the first table.
As not all variables have the same minimum number of observations - 597
These variables have the form.
Histogram. We can see that the distribution has little relation to the normal distribution
Autocorrelation Partial Correlation AC PAC Q-Stat Prob
.|******* .|******* 1 0.955 0.955 547.11 0.000
.|******* .|. | 2 0.917 0.053 1052.0 0.000
.|******| .|. | 3 0.886 0.074 1524.4 0.000
.|******| .|. | 4 0.859 0.040 1969.4 0.000
.|******| .|* | 5 0.839 0.076 2394.4 0.000
.|******| .|. | 6 0.815 -0.030 2796.2 0.000
.|******| .|* | 7 0.799 0.088 3182.8 0.000
.|******| .|* |* 8 0.791 0.112 3563.0 0.000
.|******| .|. | 9 0.785 0.047 3937.6 0.000
.|******| .|* |2 10 0.782 0.075 4310.4 0.000
.|******| .|. | 11 0.780 0.047 4681.3 0.000
.|******| .|* 12 0.783 0.106 5055.8 0.000
.|******| .|* 13 0.792 0.114 5439.5 0.000
.|***** | ***|. | 14 0.762 -0.389 5795.2 0.000
Probability of no correlation between observations is zero
Stationarity test after removal of bias and trend
Null Hypothesis: OPEN_INTEREST 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 -3.717413 0.0219
The probability of non-stationarity is about 2%, i.e. after removing the shift and the trend, the series is stationary
This is clearly visible after detrending with the HP filter.
At the bottom of the residual you can see that the mean does not change, and oscillates around the mean. But the magnitude of oscillation varies - this requires modelling heteroscedasticity.
If the analysis for the first variable is of interest, I can do it for the other two as well.
Open Interest - Open Interest
Noncommercial Long - Long positions of non-commercial traders (large speculators):
Noncommercial Short - Short positions of noncommercial traders (large speculators) ;
Look at the relationship of these variables. Let's ask a question: does Open Interest depend on the other two variables. if interesting, we can increase this list.
So look at the dependence of open interest on speculators.
Regression Equation
Estimation Equation:
=========================
OPEN_INTEREST = C(1)*LONG_IN_OI + C(2)*SHORT_IN_OI
Substituted Coefficients:
=========================
OPEN_INTEREST = 181072.989406*LONG_IN_OI + 215543.752303*SHORT_IN_OI
Evaluation of obtained coefficients
Dependent Variable: OPEN_INTEREST
Method: Least Squares
Date: 07/30/12 Time: 17:46
Sample: 1,597
Included observations: 597
Variable Coefficient Std. Error t-Statistic Prob.
LONG_IN_OI 181073.0 6965.536 25.99556 0.0000 Coefficient estimation error =181073.0, is 6965.536 about 4% - a remarkable result
SHORT_IN_OI 215543.8 7539.375 28.58907 0.0000
R-squared 0.276436 But weak matching of open interest to speculators!
Adjusted R-squared 0.275220 S.D. dependent var 46013.71
S.E. of regression 39173.32 Akaike info criterion 23.99272
Sum squared resid 9.13E+11 Schwarz criterion 24.00744
Log likelihood -7159.828 Hannan-Quinn criterion. 23.99845
Durbin-Watson stat 0.288368
I'll continue this a little later.
If the analysis for the first variable is of interest, I can do it for the other two as well.
Yes, of course you do. Open interest is not even that important. The most important column is Operators. In principle, noncommercial should be a mirror image (at least that's what you can tell by eye). But still there is an opinion that it is not so, and despite the external mirror "similarity" they are different charts, with different features not visible to the naked eye. Also small speculators should give their own independent series.
It is strange that the OI is stationary. In principle it should be correlated with the price, which is non-stationary. But the level of OI (especially in settlement futures) strongly depends on the time of expiration and steadily grows closer to it. Perhaps this is what stationarity is all about.
The relationship between all columns is generally straightforward (2 formulas calculate the cumulative long and cumulative 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;
I.e. open interest always has two sides: buying and selling, so if the open interest equals 1 contract, it means that two sellers entered into a deal of 1 contract. One will hold a short position and the other will hold a long position.
Everything is built around an uneven distribution between the groups. And when there is a strong imbalance it is a signal to make a trade.
Look at the relationships of these variables. Ask yourself whether open interest depends on the other two. if you are interested, you can increase this list.
So look at the dependence of open interest on speculators.
The regression equation is
OPEN_INTEREST = C(1)*LONG_IN_OI + C(2)*SHORT_IN_OI
The formula above is a tricky one, it only counts on cumulative long positions or only on cumulative short positions.