Econometrics: why co-integration is needed - page 20
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
Shamanism is a traditional religion on the territory of Russia. And it is the oldest.
So I would ask you not to stir things up.
I don't understand anything, please explain.
It needs two figures: kotir and profit, so it fits, especially since it is multi-currency. But it is the same TS, so its properties will be the same, if it converts non-stationary coir to stationary profit, then it is, as that is its property. or it does not. That's why we need reports from different TS. The idea is that in profitable TS the quote is cointegrated with profit and in unprofitable ones it is not. I would like to check this idea.
From the ATC 2011 website:
Open the file in Excel, put a filter on the participant and go to work. All in one package, you can't do better than that.
It's easy to put a filter by account (first column). For example, 800014. And then you see trades made on one account, i.e. by one participant.
It's easy to put a filter by account (first column). For example, 800014. And then you see trades made on one account, i.e. by one participant.
The problem was solved in the following way.
I selected two columns price and profit with 374 ones by the eurusd symbol. Calculated the balance, which is equal to the previous balance multiplied by the previous profit. I have obtained the balance accumulation in the account. It looks like this:
Graph
The unit root test indicates that the quote is not stationary
Balance graph
Unit root test indicates that the balance is not stationary.
I will also show the profit graph:
A very interesting graph. Then the unit root test for the profit:
Profit is stationary. And it is the difference for the balance.
Let's check for cointegration of price and the variables listed in the table on the full sample:
Co-integration is present.
Let's take the part of the sample, the beginning, where the TS makes a loss:
Cointegration present
Take the part of the sample where the TS makes a profit:
Co-integration present
The conclusion from the above is that TC cannot be estimated using cointegration.
You need two numbers: kotir and profit, so it fits, all the more so as a multi-currency. But this is the same TS, so its properties will be the same, if it converts non-stationary coir to stationary profit, then so it is, as that is its property. or it doesn't. That's why we need reports from different TS. The idea is that in profitable TS the quote is cointegrated with profit and in unprofitable ones it is not. I would like to check this idea.
Check it. The idea is a good one. Planned profit is only needed in pips. Everything almost has to be counted in points - entry into the market. Exit - there already in the currency of the deposit at the pair's exchange rate should be counted. Do not bother with other systems' reports. Your beginning is correct - develop your thought.
For example - quotation - stationary series, almost. Find these odds based on the current depo size. Take maximum profits mixed with minus on exit.
My system never goes into minus and constantly recalculates the parameters of the stationary series on market exit.
Good luck. All will come out. Do not believe in anything. Each pair gives about one percent a day. Don't run into couples who are dead souls. Very undesirable. Choose carefully.
The problem was solved in the following way.
Selected two columns price and profit 374 by the eurusd symbol. Calculated the balance, which is equal to the previous balance multiplied by the previous profit. I have obtained the balance accumulation in the account. It looks like this:
Price graph.
The unit root test indicates that the quote is not stationary
Balance graph
Unit root test indicates that the balance is not stationary.
I will also show the profit graph:
A very interesting graph. Then the unit root test for the profit:
Profit is stationary. And it is the difference for the balance.
Let's check for cointegration of price and the variables listed in the table on the full sample:
Co-integration is present.
Let's take the part of the sample, the beginning, where the TS makes a loss:
Cointegration present
Take the part of the sample where the TS makes a profit:
Co-integration is present
The conclusion from the above is that TC cannot be estimated using cointegration.
I read and read the books and decided to resuscitate the thread with this observation.
In fact anonymous wrote everything, I just did not understand it. I had a limited understanding of the trend, I always have in my head a picture of a straight line from the TA, which is not true at all. If the trend is considered a deterministic component of the quotient, deterministic due to the fact that it is expressed by a formula such as regression or a filter, then the picture of using cointegration is as follows.
Let's take two pairs: EURUSD - GBPUSD.
Each of these series on the specified area I(1) is integrated.
We look for the cointegration equation, which has the following form:
EURUSD = 0.828446321089*GBPUSD - 1.37009549204 + 0.000206761078317*@TREND
We get the picture
At the bottom is the difference, the residual, which is stationary, i.e. it always returns to the zero line. Now using cointegration, i.e. the bottom chart: if we have a quotient equation, by which we can select the trend direction, then:
if the trend is up, we buy GBPUSD at the points of the kotir when below zero (below the dotted line which is above)
if trending downwards, we sell EURUSD at the quotes, when it is above zero.
Pair trading. In either direction of the trend, we have an additional opportunity to enter the market and respectively exit.
I read and read the books and decided to resuscitate the thread with this observation.
A couple of remarks. Not in the sense of criticism, but as thoughts on the subject.
Using the results of the contest for experiments is not entirely justified. The matter is that some other approaches are used in contests, as opposed to "normal" trading. Therefore, the analysis results are not quite correct.
As for spread trading techniques. One of them is what you've described. This is not exactly a "pair" trade, but with the use of a synthetic indicator, it's basically the same wizard. The other method is classic, trading on the movement of the indicator towards zero. I.e. at highs (or so) we enter on both instruments (in case of basket trading, buy basket), at zero we exit. Another method is the opposite: enter at zero, exit at the highs (or at a certain distance from zero).
I have encountered on the Internet that "pair trading" has much in common with marting. In the sense of unjustified risks. Everything seems to be good, many trades are in the plus, etc., but one unlucky one can take out a deposit.
Scope is described as long term, as a hedging tool, on "physically" related or dependent assets. Or as HFT, on ticks, but there the advantage is not only in the method, but also in the technique.