
Use a product that allows you to express a view on realized correlation... Specifics will depend on the asset/product class in question.Quote from mizhael:
Let's say you have two assets that historically have high correlation, but recently their correlation approached 0 or even became negative.
Any trading strategies?
Quote from Martinghoul:
Use a product that allows you to express a view on realized correlation... Specifics will depend on the asset/product class in question.
As to TA, there's a whole variety of TA "methods" that are fundamentally based on mean reversion.
Quote from psytrade:
simple dispersion became unprofitable in mid 2000s according to the author at Nuclear Phynance:
http://www.nuclearphynance.com/User Files/2/Dispersion - A guide for the clueless 1.1.pdf
Trading correlations profitably now would likely involve trading a particular view when relationships between risk assets are behaving out of line, rather than through vanilla equities/options

Well, the difference is simply the rigor and diligence of the analysis. A lot of TA is arbitrary mumbo-jumbo, whereas statistics is based on some sound mathematical concepts. Sorta like astrology and astronomy... This is my personal opinion.Quote from jem:
of course... for instance bollinger band bounces at say 2sd deviations seem to used by pair traders and many t/a people.
My rant was directed the guys who think statistics are not t/a. It was not a well written rant I apologize.