For those who have (are) seriously engaged in co-movement analysis of financial instruments (> 2) - page 18

New comment
 

R is a very powerful and handy statistical tool.

For example, to get such a clear Scatter Plot table:

Only a few R lines are required. The code on mt4R is in the appendix.

Files:
scatterplot.mq4  3 kb
 
genro:


"When Correlation = FALSE, the practical part of optimal synthetic Recycle weights are shown. These weights reflect volumes of corresponding symbols when the synthetic is trading..." - from the description of Recycle (practical application).

I built Recycle for three Instruments and used weight coefficients as a number of lots of each Instrument to build an Equty for this synthetic:

According to the idea Equty synthetic should be in the horizontal channel, but for some reason it is not.

Explain what I'm doing wrong or maybe I don't fully understand the meaning of the indicator.

I understood that the Coefficients do NOT reflect the number of lots, but show the Capitos to be invested in the instrument...

conventionally - there is some amount of $ for transaction.... distribute them according to coefficient....

5762$ invested in EurUsd

$5756 in GbpUsd

5803$ in EurGbp

=====

it looks like this :) lots = 5762/1.30654.... 4.41 lots Eur/Usd.... etc..

 
Aleksander:

I understood that the Coefficients do NOT reflect the number of lots, but show the Capitos to be invested in the instrument...

arbitrary - there is a certain amount of $ for the transaction.... allocate them according to coefficient....

5762$ invested in EurUsd

$5756 in GbpUsd

5803$ in EurGbp

=====

it looks like this :) lots = 5762/1.30654.... 4.41 lots Eur/Usd.... etc..


That is exactly what I want to clarify.
 

I have commented negatively on the use of regression several times in the thread.

Based on this post (Mathcad file and input data attached there) I did (continuation of Mathcad file) a small comparison of regression-based synthetic versus best solution vector-based synthetic:

You can get the optimal regression vector after running Arb-O-Mat on each symbol from the FI set and selecting the best vector obtained.

Note the resulting vectors.

Optimal regression vector:

The weighting coefficients for the 2nd(GBPUSD) and the 7th(SILVER) FI are zero (the FI are not involved in creating the synthetic). Also the weighting coefficients are very different from each other.

Best vector solution:

The weight coefficients are not very different from each other. There are no zero coefficients. All FIs form synthetics with almost the same strength.

You can also see the RMS values. Accordingly, you can estimate the characteristic of the narrowness of the horizontal synthetic channel - the degree of market interconnection.

P.S. Interestingly, in one case SILVER does not participate at all in shaping the synthetic, and in the other case it is the strongest.

 
hrenfx:

You can also see the RMS values. Accordingly, you can assess the narrowness characteristic of the horizontal synthetic channel - the degree of market correlation.

1) In statistical arbitrage, I would prefer fluffy quotes.

2) One barefacedness, unfortunately. ;-)

 
  1. I would still prefer some sort of channel property preservation on the OOS.
  2. I have given my reasons above.
 

Let us imagine that the whole market is 100 FIs. Since it is the whole market, the FIs are interconnected with each other in the following way: if there is a decrease from one FI, it is redistributed to the others. This is analogous to the law of conservation of energy. That is, our market is a closed system. And 100 FIs represent a "Ring". This means that any FI can be absolutely calculated through 99 others.

Now let's imagine that we don't know any information about one of the FIs. How would you look for relationships between the remaining 99?

 

Then you want to embrace the immensity :)

In any market you will lack information :)

 
hrenfx:

Let us imagine that the whole market is 100 FIs. Since it is the whole market, the FIs are interconnected with each other in the following way: if from one FI has decreased, this has been redistributed to the others. This is analogous to the law of conservation of energy. That is, our market is a closed system. And 100 FIs represent a "Ring". This means that any FI can be absolutely calculated through 99 others.

The hypothesis of a closed market does not seem to be true at all. I mean it can be said to be true, but only in certain sections of history. More or less regularly the "energy" changes by leaps and bounds and the market forgets its previous state. It has a memory, but a short one.
 
Mathemat:
The hypothesis of a closed market does not seem to be true at all. I mean it can be said to be true, but only in certain stretches of history. More or less regularly the "energy" changes by leaps and bounds and the market forgets its previous state. It has a memory, but a short one.

This is where Dick is right. The price of an individual FI for currency pairs is calculated trivially through two currency pairs. That is, if we know the prices of two of the three FIs forming the ring, then the third FI is trivial.

Simply put, in order to compute the whole currency market: majors and crosses, only majors are necessary and sufficient, since crosses are trivial.

The law of monetary conservation is not entirely correct, because the issue of currencies is constantly changing and most often upwards due to inflation. But it can be used as a rough approximation.

1...111213141516171819202122232425...38
New comment