Thoughts on some of the absurdity of multi-currency analysis. - page 38

New comment
 
 
 
Has anyone built a synthetic volume (at least a tick volume) for the currency index itself?)
 
Romanlo:
Has anyone built a synthetic volume (at least a tick volume) for the currency index itself?)
Not very realistic.
 
Romanlo:

I'm just going to tell a little more to everyone who is interested in where to "dig", because I think I'll soon leave this forum because it's not necessary :)
There are millions of strategies on this forum, they search the market through certain "weighting coefficients". Please excuse me for not being precise, because I'm not fluent in all terms. If we consider these strategies in the physical sense (sorry again, I love physics that explains all processes, not mathematics that just gives formalized representation) - it turns out that these strategies reduce to creation of some "virtual scales", these scales have 8 levers and 8 currencies are on the scales, so there are many strategies that search for pairs that tilt these scales in different directions and as a result - another EA that works on correlated pairs in different directions. I can agree that on some time intervals it all works and is very effective, but it does not always work. When it stops working there is some noise saying that brokerage companies have changed quotes again, etc. The essence of my research is the following - the weighting coefficient of the currency is not so important, and what is important is statistical movement of this currency - in the physical sense it looks like this Draw 8 dots on a piece of paper and sign the currencies, take the statistical indicator MA with a large period and draw the movement of these dots depending on this indicator - up/down (I can say right away I do not draw anything on a piece of paper - it is just visible by eye), and do not forget that currencies can be swapped in columns to show similar sequences of movement, and after some time you should see how these currencies move relative to each other, types of movement I have described above. Based on this movement you can draw conclusions about the movement between the currency pairs, then you can see when the correction is going on and when the trend starts, you can see when the pullback is going on. That is probably the whole TS. Probably you will not find more on the Internet :).

and this branch https://forum.mql4.com/ru/9154 , But with such decompositions one should not forget about normalisation of spectrum and phase alignment, i.e. take into account integral equality of functions formed by two neighbouring filters.

I am a bit busy at the moment, I will continue later.


"very efficient, but it doesn't always work" a wind turbine is not always blowing the right wind force, the ACS determines the ability to generate the necessary power (analogy).

"Thoughts about some absurdity of multicurrency analysis."I do not agree with"some absurdity".

Like the daily chart

and a smaller hourly chart.


if you look carefully, you may see five asterisks )) (A bit of combinatorics).

 
Сosty, I posted it in another thread, but the flooding is so fast that it may go down faster than you can see it. I'll move it here.

Why do you always make your screenshots in the form of mush? What's the use if you still need to arrange them in ascending order of price distance from the machine with each count, all pairs with the same currency as in your screenshots. The goal is to identify the most dispersed and preferably simultaneously reversed currencies.

PS: I don't get the five stars.
 
costy_: If you look closely, you can also see five stars )) (a bit of combinatorics).
Tried combinatorics, didn't work out for shit. Maybe you can do it.
 

Romanlo:

Why do you always make your screenshots in the form of mush? What's the use if you still need to arrangethem in ascending order of price distance from the machinewith each count , all pairs with the same currency as in your screenshots. The goal is to determine the most dispersed and preferably simultaneously reversed currencies.

I've decided not to use something that was done long ago, double-checked and thrown in the trash, a mess of 8 major and 28 derivatives, and 28 flips is rather a tray with sandwiches))

size from mashka to what price to take, maybe a median, in what dimension to count, maybe in gold, and what will it yield ))

Mathemat:
Tried combinatorics, didn't work for shit. Maybe you can do it.

but to try to count, the yen, 7 out of 7 down, the eu, 7 out of 7 up, the maximum change and direction is there, well a minute can be watched for entry))

you can do it another way, only the yen and the british go down...

 
costy_: but to try to count, the yen, 7 out of 7 down, the eu, 7 out of 7 up, the maximum change and direction is there, well a minute can be watched for entry))

could be different, only the yen and the british are going down...

It's not that simple. When implementing the algorithm there is a need for new hidden parameters. Well, for example, what is "down" / "up". Statistics are needed here to find, say, a minimum threshold above which it can be considered a move.

It's just a word of mouth.

 
Mathemat:

It is not that simple. When implementing the algorithm, there is a need for new hidden parameters. Well, for example, what is "down"/"up". The statistics are needed to find, say, a minimum threshold above which it can be considered a move.

This is only in words.


I agree, there are many methods, choose the most understandable (your own) and use it,

For instance, the GBP and the Yen are both going down, the chart shows increase of the GBP/Yen and decline of the others, so it turns out that the Yen is moving more than the GBP.

We use this analogy to look for the strongest growth, and then make a decision (this is from another operetta).

 
costy_: For example british and yen are falling, graph british/yen shows growth of british, for others it turns out that there is more movement in yen than in british.

We are looking for the strongest growers, and then make a decision (this is from another operetta).

Exactly, I'm saying that in words everything is simple, but in concrete implementation different nuances appear, which make my head spin.

My implementation weighs >20K code with dozens of functions (I prefer short functions, no more than 30-40 lines).

I think I will go back to single currency now, but at a new level.

1...31323334353637383940
New comment