Thread: Nonparametric Association Measures in Systematic Trading

Interestingly enough, here is a good example of Spearman's rho in trend detection. The newspaper is somewhat technical. So a normalnigeriaforextradingmember may bypass most mathematical elements to get an idea of the suggested methods.

My view: the suggested tests may function as one input in a more complex detection method. They aren't sufficient on their own.
https://www.nigeriaforextrading.com/...1560627830.pdf

Originally Posted by ;

Interestingly enough, this really is a good illuion of Spearman's rho in vogue discovery. The paper is somewhat specialized. So a regularnigeriaforextradingmember may bypass most mathematical parts to get an notion of the suggested methods. My view: the suggested evaluations may serve as one input in a more complex detection system. They aren't enough on their own. picture

This eduional paper is full of formulas that may or may not have any relevance to trading.

Complexity in trading doesn't work; simplicity will work. I've spent heaps of time in study in the past; such as artificial intelligence; and nothing seems to be solid for long-term trading.

I've got two bookshelves at home; one for matters that works; and one for intellectual master*****n things.

Take my advice; and for for KISS

Originally Posted by ;

quote This academic paper is filled with formulations that may or may not have any relevance to trading. Complexity in trading does not work. I have spent heaps of time in research previously; including artificial intelligence; and nothing appears to be strong for long term trading. I have two bookshelves in the home; one for things that works; and you for intellectual master*****n stuff. Take my advice; and for for KISS

Alphadude, If you're so easily terrified by formulas, then I told you to bypass the mathematical components and focus on the idea of the evaluations. You sound lazy enough not be willing to read strict text and only chat, chat, chat on this forum. Any method may be invisibly in formulations (even if it's unnecessary). So by your logic, almost any method is wrong... I use higher math to generate profits every week, by conducting non-trivial classifiion algorithms.

Originally Posted by ;

quote Alphadude, if you're so easily scared by formulas, I told you to bypass the mathematical parts and revolve around the idea of the tests. You sound lazy enough not be pleased to read rigorous text and just chat, chat, chat on this forum. Any method may be invisibly in formulations (even if it's unnecessary). So by your logic, almost any procedure is incorrect... I use higher math to generate profits weekly, by conducting non-trivial classifiion algorithms.

My poor. I will take a look at time this weekend.
Did a fast research in youtube; and got to know that the formulas bit better:




I will see a possible use for these correlations; and the paper above.

Thank you, Alphadude.

Staassis, have you attempted to invent a spread series between correlated pairs?

Basically you exchange the spread the Exact Same way you exchange the cross pair (i.e. exchange EUR/USD and USD/JPY, instead of EUR/JPY)

Alphadude, I think trading the spread between A and B is a particular case of pairs trading (when I have understood you correctly). In greater generality, pairs trading allows one to select best weights for A and B. I am also not certain that A and B should be chosen based on significance only. Some time ago I did pairs trading at which the option of pairs and conclusion of weights was based on cointegration. This was in a asset class but not FX. I'd have encouraging results but, regrettably, for institutional reasons I never ended up making millions in my egy. In my opinion, that egy worked because there were fundamental reasons for its resources to co-move. Time series analysis alone is inadequate, unless one transactions very big portfolios for a brief time period.... Having said that,

_______ Kendall's tau or Spearman's rho cointegration established trading fundamental reasons for its resources to show similar dynamics

seems interesting. Please be aware that all those ingredients are present in the paper by Landgraf et al.. . I am cautiously curious about copula-based trading but haven't backtested anything like that however. Purely intuitively, the option of pairs has to be more significant than determining between copula based implementation and cointegration established execution.

Great sharing.

Originally Posted by ;

AlphadudeI think trading the spread between B and A is a particular case of pairs trading (when I've understood you correctly). In larger generality, pairs trading enables one to select optimal weights for A and B. I'm also not certain that A and B should be chosen based on significance only. Some time ago I did pairs trading where the choice of pairs and determination of weights was based on cointegration. This was in a macro asset egory but not FX. I'd have encouraging results but, regrettably, for institutional reasons I never ended up making millions...

another means to exchange the spread (aka pairs trading); would be to make stationary timeseries. An illuion would be to detrend the price collection.

Another problem that seems is mean float of the stationary timeseries.

I think that it is possible to make a portfolio of FX instruments with varying weights; to make a stationary timeseries; and trade which portfolio.

Originally Posted by ;

quote Another method to trade the spread (aka pairs trading); is to make stationary timeseries. A good illuion is to detrend the price series. The next problem that appears is mean drift of this stationary timeseries. I believe it is likely to make a portfolio of FX instruments with varying weights; to make a stationary timeseries; and trade that portfolio.

Alphadude, I believe there's possible in this direction. There are at least two means of using stationary combinations of financial indiors:

1) The combination is tradable and contrasts to the price of some portfolio, perhaps with always rebalanced weights,
2) the combination isn't tradable, e.g. RSI(strength 1) - RSI(strength 2) (just a silly example).

Different posts and books have concentrated on case 1. Nonetheless, in the FX world case 2 may have greater prospects. The static mix may function as a signal creating instrument for trading closely associated assets. Backtesting several ideas in that way is really on my list.