Consider X > Y => X-delta = Y , where delta >0 .

From that you obtain X =Y + delta OR Y= X-delta depending on which quantity is of interest to be the dependent or independent variable.

I am not sure I understand. If X and Y are random variables then you could have X a uniform random variable on [0,1] and Y an independent random variable uniformly distributed on [3, 5]. Then Y>X but Y and X remain independent.

Dependency of two variables can only determine logically or by using a planned experimental design with having control on both X and Y variables.

I am agree with Ehsan Khedive.

If X and Y are random variables, where the values taken by X and Y belongs to R, restricting the condition that max (X)

Ok I see what you mean now. Thanks for clarifying.

In general your claim is false : Y

Respected Yvik Swan,

Actually in my case there are two mappings in two separate metric spaces, for data taken by the random variable X forms a mapping and Y also forms a mapping on two different metric spaces. We can not consider these two variables X and Y as we used in regression analysis.

Dhruba, if we may say that Y

F(x,y)=P(X\leq x, Y\leq y)=P(X\leq x, Y\leq y, X

Respected Emilio, You mean to say that these two distributions are dependent.

Yes, Dhruba, I think they are mutually dependent from our ordering point of view. emilio