The user discusses trading strategies, emphasizing the importance of non-linear dynamic portfolios over static cointegration and highlights the limitations of traditional approaches.

Roman Shiredchenko 2026.07.10 15:12 #2911

Vitaly Muzichenko #:

I am conducting experiments using Pearson’s correlation (red) and Spearman’s correlation (blue)

Spearman’s correlation is more suitable, as it does not take into account the GEP

--

AUDNZD + EURJPY

CHFJPY + GBPCAD

--

Thank you for your research – I’ll be incorporating the spreads and settings into my trading strategy!

mvf358 2026.07.10 16:13 #2912

Vitaly Muzichenko #:
You don’t need to look for it in the majors; the dreamer is right.

It’s not about the majors. In this case, there’s no need to split the 28 pairs into majors and crosses. It’s a completely different situation here. As an example, I posted a screenshot of the ‘Portfolio Modeller’ indicator, where two synthetic pairs correlate 100 per cent, and no pair is repeated. I seriously doubt that, by following the approach you’re taking, you’ll manage to find a combination of pairs that will bring you a profit in the long run. Good luck on your difficult journey.

For major banks Considering Econometrics: bibliography [WARNING CLOSED!] Any newbie

Vitaly Muzichenko 2026.07.10 16:53 #2913

Roman Shiredchenko #:

Thank you for your research – I’ll be incorporating the spreads and trading settings into my work!

All the indicators and brief descriptions are available on my blog.

Feel free to carry out your own experiments

Roman Shiredchenko 2026.07.10 21:06 #2914

Vitaly Muzichenko #:

All the indie titles and brief descriptions are available on my blog.

Feel free to experiment

Thanks, I’ll have a look...

Aleksander 2026.07.11 01:34 #2915

How come the majors aren’t making a profit? Here you go, 1,700 per cent on the majors alone:

Files:

untitled-1_10.jpg 23 kb

Vitaly Muzichenko 2026.07.11 02:37 #2916

Aleksander #:
How can it be that majors don’t generate a profit? Here you go, 1,700 per cent on the majors:

You’ve been running around with this random image for 10 years now. Are you going to keep showing it for another 20 years?

Statistics show how people make money on majors – they can’t keep up with topping up their balances… and then it’s back to the majors again.

FOREX - Trends, forecasts Some one help me, Trend vs Flat

Roman 2026.07.11 05:44 #2917

Aleksey Nikolayev #:

Do you break the rules of the Russian language on Fridays?

Roman has further developed the theory – dynamic portfolios using the theory of stochastic process filtering.

There you go, Alexey has hit the nail on the head: it’s precisely these multidimensional, non-linear dynamic portfolios, rather than static, cointegrated ones.
Filtering theory is not the main tool, but rather a corrective measure in case the process turns out to be trending.
However, the theory of non-lagging filtering itself is abstractly embedded in the coefficient search algorithm, designed to minimise errors in the coefficients.
And it is precisely the theory of filtering that clearly illustrates the problem of shifts for different algorithms when applied to the vector of coefficients.
That is, to understand the problems. I know you all understand what I’m writing about. It’s the other people here who are studying AI, and cointegration is a new discovery for them.

Adaptive digital filters Econometrics: why co-integration is Random Flow Theory and

Roman 2026.07.11 06:27 #2918

mvf358 #:
We need to focus on developing practical skills, rather than theory, which takes us away from reality. Whenever I asked him to demonstrate his theory in practice and make a profit, he would put on a show and backtrack. As long as his theory remains nothing more than a theory, it currently has no bearing whatsoever on the subject of Sinus.

...