Problem 8.12 Easy Free

Normal Half-Plane by Symmetry

Let XX and YY be independent standard normal random variables. Find P[Y>4X]\mathbb{P}[Y > 4X].

Work it out, then compare with the answer.

SolutionShow the worked solution

Since XX and YY are independent centred normals, the linear combination Z=Y−4XZ = Y - 4X is normal with

E[Z]=0,Var⁡(Z)=Var⁡(Y)+16Var⁡(X)=17,\mathbb{E}[Z] = 0, \qquad \operatorname{Var}(Z) = \operatorname{Var}(Y) + 16\operatorname{Var}(X) = 17,

so Z∼N(0,17)Z \sim \mathcal{N}(0,17). A centred normal is symmetric about 00 and puts no mass at 00. Hence

P[Y>4X]=P[Z>0]=12.\mathbb{P}[Y > 4X] = \mathbb{P}[Z > 0] = \tfrac12 .
12\boxed{\tfrac12}

Alternative. The pair (X,Y)(X,Y) is rotationally symmetric about the origin. The event {Y>4X}\{Y > 4X\} is an open half-plane whose boundary is a line through the origin, so it has probability 12\tfrac12.

Generalisation. The same argument gives P[aX+bY>0]=12\mathbb{P}[aX + bY > 0] = \tfrac12 for any (a,b)≠(0,0)(a,b) \neq (0,0).