Correlation Coefficient Explained | Forex Factory

  • Correlation Coefficient Explained

    From newtraderu.com

    In finance and investing the correlation coefficient is used to calculate a statistical measurement of how strong movements in two markets are historically in relation to one each other. The correlation is expressed in a range of values between 1.0 and -1.0. All the readings of the formula are inside the 1.0 to -1.0 range. So any calculations that come out to more th“Diversification is the only free lunch” in investing, says the quote attributed to Nobel Prize laureate Harry Markowitz.an 1.0 or less than -1.0 are an error. A -1.0 shows a perfect inverse correlation in movement. One market goes up by the same ... (full story)

Comment Options
eztrader1
Feb 14, 2020 9:27am
Permalink
No idea about statistics, mathematics or markets.
I could vomit about those self-declared experts on any sh...

CORELATION is a measure of CO-MOVEMENT, not of CO-Strength or CO-Power. It is simply a directional measure weighted by probabilities.

Therefore a correlation of 1 (-1) DOES NOT say, that if one market goes up, the other market goes up (down) BY THE SAME AMOUNT. NOT AT ALL!!!

On goes up, the other as well (or down). THAT IS THE WHOLE INFORMATION

If you want to compare SIZES of movements, not only direction, YOU NEED BETA!!! The baby of a Markowitz disciple and Noble laureate as well.

Beta is a reformulation of the Pearson-Correlation-Coefficient, and it comes only with an ALPHA.

There are some many "helpful" websites offering statistical tools on VOLATILITY or CORRELATION. None of them has the slightest idea about these parameters. Most of them do not even disclose the used formulas at all, when you ask them, and you luckily get an answer, you know they are wrong. Some, at least, disclose their calculation process, so you can immediately realize, what crap you are looking at. Even the parameters on OANDA are wacky. They state the correct formula, but comparing their values with other websites (myfxbook,investing, etc) or my own (==correct) numbers, results always in differences.

But it is at least a site to start from. If they show some significant parameter, one start to run his own calculations. And has own calculations are usally combined with need to look for datasets, it helps to rank someone's research projects from promising and worthwhile to possible waste of time.

SO

STATISTICAL TRADER BEWARE! DO NOT TRUST ANY INFOSITE! IT MIGHT RUIN YOU!
Reply
  • Story Stats