Pair trading and multicurrency arbitrage. The showdown. - page 109

New comment
 
Maxim Dmitrievsky Gann nets and Andrews forks again :)

It is a strange assumption that real correlations on the financial markets were analysed by digital processing of price charts))))) At most statistical dependence. And the causal things on the financial market are definitely not TA and COC))))

 
Maxim Dmitrievsky Gann nets and Andrews forks again :)

This is a layman's opinion. For such a loud statement you need to analyse about a million tickers on all stock exchanges of the world.
Cointegration is successfully used on the stock market and possibly on other markets. Yes, it doesn't work on forex.

Theconcept of cointegration was proposed by economists, not by COCs.

 
Alexander Sevastyanov #:

This is the opinion of a dilettante. For such a loud statement one should analyse about a million tickers on all stock exchanges of the world.
Cointegration is successfully used on the stock market and possibly on other markets. Yes, it doesn't work on forex.

Theconcept of cointegration was proposed by economists, not by COCs.

Is it successfully used by you?
 
Aleksey Nikolayev #:
But this is a much more advanced torba than their usual one - searching for cycles in prices by means of Fourier. It will be 100 years older than cointegration.
You can't argue with that, it's quite beautifully done :)
 
Maxim Dmitrievsky #:
Have you used it successfully?

Used by those I started with about 18 years ago.
For a number of reasons, I went a different route.

 
Alexander Sevastyanov #:

Used by people I started with about 18 years ago.
For a number of reasons I went the other way.

Already the references have gone infotsygan, yeah, and also some famous youtuber some famous youtuber uses :)
 

Gentlemen, can you have a childish debate about"whose sensei is cooler" in another thread ?

 
Maxim Kuznetsov #:

Gentlemen, can you have a childish debate about"whose sensei is cooler" in another thread ?

Rena's is the coolest. )))

 
Alexander Sevastyanov #:

This is a layman's opinion. For such a loud statement one should analyse about a million tickers on all stock exchanges of the world.
Cointegration is successfully used on the stock market and possibly on other markets. Yes, it doesn't work on forex.

The concept of cointegration was proposed by economists, not by COCs.

There is a problem related to multiple hypothesis testing. If we test for cointegration with pvalue=0.05 for a million pairs, then about 50000 pairs will have "found" cointegration according to the WBC. Probably, this number will be different because of some violations of the WBC conditions, but it will be not insignificant. If we introduce test corrections like the Bonferroni correction, we may lose those pairs, where cointegration is definitely present (calendar spreads, for example).
 
Valeriy Yastremskiy #:

It is a strange assumption that real interrelationships in the financial markets were analysed by digital processing of price charts))))) At most statistical dependence. And the causal things on the financial market are definitely not TA and COC))))

Well forex is of course a market, but not exactly financial, but rather infobiz market

Stat arbitrage between indices, futures and other derivatives - we can conditionally call some of them cointegrated, but these instruments were specially created in advance, otherwise there would be no sense in them. And there the yields are small.

1...102103104105106107108109110111112113114115116...261
New comment