Pair trading and multicurrency arbitrage. The showdown. - page 91
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
A large window size increases the time of being in a position, and by and large it leads to a transfer into positions by time.
Here we should proceed from what we are counting. And what we are counting is profitability. This is the basis for what period we consider this profitability.
Profitability is usually considered for a day, a week, a month, a quarter, a year. Here they are all the needs.
You can also count for 8 hours, like a trading session. Therefore, everyone has his own choice here.
I don't know. I repeat that Eugene has no delays and he has already calculated zero on the screen.
In general, until I check it myself, I will not be convinced of my assumption.
But it seems there is a logic, since the coefficients are calculated at each step, which equalises the curves,
and at each moment of time all processes still give a common zero state)).
If it has ATR in the calculations, then delays and disturbances are already there and lots of them.
PS. You by the way did not discount the link you mentioned earlier :-)
but you can't fix the "curse of sliding windows" completely with no coefficients, corrections and dynamic sizes.
Let me add a thought.
I understand what curses you are talking about, these curses have a culprit. And it is statistical methods of solving the problem.
Statistical methods give a bias of estimates, from this and float curves. What scientific methods I have not tried, floats and everything.
That's why I was caught by Eugene's screen, that there are no these curses there. And there is a supposed answer to this,
the derivative of the system is always zero, thanks to the coefficients, because the formula leads to this zero state.
In other words, the error is 0.
Which link?
Let me add a thought.
I understand what curses you're talking about, these curses have a culprit. And it is statistical methods of solving the problem.
Statistical methods give a bias of estimates, and that's what makes the curves float. What scientific methods I have not tried, floats and everything.
That's why I was caught by Eugene's screen, that there are no these curses there. And there is a supposed answer to this,
system is always equal to zero, thanks to the coefficients, because the formula just leads to this zero state.
What link?
about another method ofcalculating volume in spread trading. Or was that your namesake?
about another method ofcalculating volume in spread trading. Or was that your namesake?
No, it was probably not me )))
Nah, I guess that wasn't me ))
Ah...you mean about puzzles ;-)
The solution to the puzzle about the meaninglessness of ATR: if the currencies are combined into a common base and normalised by yield, then ATR is not only constructively lagging, it is similar to the point of indistinguishability. The difference will be in the noise
and the second edge of the puzzle - the structure of ATR is initially known (when it increases/decreases). But the ATR structure can be "predicted" by comrades who have mastered Fourier, it is periodic.
Oh, you mean puzzles.)
The solution to the puzzle about the meaninglessness of ATR: if currencies are combined into a common base and normalised by yield, then ATR is not only structurally inherently lagged, it is similar to the point of indistinguishability. The difference will be in the noise
and the second edge of the puzzle - the structure of ATR is initially known (when it increases/decreases). But the ATR structure can be "predicted" by comrades who have mastered Fourier, it is periodic.
I didn't say anything about ATR either ))
I just put together what Renat wrote about.
And ATR is not needed there at all. There is nothing there, except for the formula ))
And then you can modify it at your discretion.
I didn't say anything about ATR either ))
I just pieced together what Renat wrote about.
And ATR is not needed there at all. There is nothing there, except for the formula ))
And then you can modify it at your discretion.
It's about puzzles :-)
We should also mention CME, so that no questions are asked.
Well, what he shows on the screens is nothing like what I get on the hourly chart, if we are talking about Renat's approach.
So he has his own bicycle of construction.
On this screen without the formula, I still can't get to write it.
At the moment on a simple yield on hourly charts, there is such a bifurcation.
no correct ones yet.
more or less
but it's still not right.
overmuch
I would like to share my observations, what is important to observe in automatic multicurrency trading.
1.. important point is to understand in what state is the general channel of the market (how much money is in the cash register and the direction of movement).
To build an algorithm of this understanding, you will need to take from 8 buffers of the indicator CCFp data on currencies - sum up their values, you will get the total width of the channel.
This value of channel width should be taken for the last 15 candlesticks (I have it so), take the average value which is equal to 1, relative to it, recalculate the values for all candlesticks
and then you will get different variants of entering the position. There are 4 variants and each variant has 4 sub-types of positions - i.e. 16 types in total.
2. Each variant should be described separately, Forex is like a car with four wheels on a 4-pot engine with a distributor differential.
Determine the width of the channel for the period and the average width of the channel for 15 periods ( 100 is 100% average width of the channel, 65 -75 is the minimum width of the channel according to my observations can fall to 60 - 75%,
maximum width can be 130-145)
I will write the values for the last 4 periods for example 0,1,2,3 (0-current) that you get for further calculations.
The market states are as follows:
1. gap1 66.00, 67.55, 75.10, 89.28 Channel width is less than average 100%, the channel is shrinking
2. cross1 89.28, 75.10 , 67.55, 66 . 00 Channel width is less than average 100% , channel is expanding
3. cross2 110.01, 107.55, 85.10, 79.28 Channel width is greater than average 100% , channel is expanding
4.gap2 121.00, 127.55, 131.10, 129.28 Channel width is greater than average 100% , the channel is shrinking.
It should be noted that there is a directional movement of the channel, and there are boundary values at which the direction of movement changes.
3. For each bet out of 28 pairs of 8 major currencies it is necessary to determine its share or share of the total width of the channel.
The share changes dynamically, the chart can grow, for example, and the share at the same time fall, because the total width of the channel grows faster than the share of a particular bet grows.
Determining whether the share grows or falls relative to previous periods is connected with the growth/decline of the total channel width, so this growth coefficient should figure in the formula
to determine the signal to buy or sell, or exit from the position.
4. Having determined the width (more less) and the direction of the channel movement (more less), as well as the growth or fall of the share (more less), you get two options for solving the question (sell or buy)
and accordingly 16 options for answering your question 2x2x2x2x2=16.
This is the forex differential, having sorted it out - now we need to describe 16 variants
for each bet from 28 pairs of basket of 8 currencies "USD", "EUR", "GBP", "JPY", "AUD", "CAD", "NZD", "CHF".
I would like to share my observations, ....
And how do these observations benefit you in terms of profit?