Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I have two normally distributed random variables, X and Y.

X has mean 66 and standard deviation 6.

Y has mean 77 and standard deviation 7.

The correlation between the random variables is given as 0.8.

The task is to find the probability $$P(X > 0.7Y)$$


This is my attempt:

Create the random variable $$D = 0.7Y - X$$

and calculate $$P(D < 0)$$

The variable D should also be normally distributed with these properties:

$$\mu_D = 0.7 \times 77 - 66 = -12.1$$ $$\sigma^2_D = 0.7^2 \times 49 + 36 + 2 \times 0.7 \times -1 \times 0.8 \times 7 \times 6 = 12.97$$

Then, $$P(D < 0) = F_Z(\frac{0 + 12.11}{\sqrt{12.97}}) \approx F_Z(3.36)$$

But, this is not the answer I'm supposed to get (which is instead $$F_Z(1.17)$$ according to the textbook). Where am I going wrong?

share|improve this question
    
I see... but I don't see why. The formula I have to find the variance of D does not use squares. It's just $$a^2\sigma^2 + b^2\sigma^2 + 2abCorr(X, Y)\sigma\sigma$$ (I'm being sloppy with the subscripts here, but yeah). –  vicvicvic Jan 15 '12 at 16:26
    
Actually, you are right using that formula (my answer was incorrect, I forgot to compute the new covariance....). –  David Mitra Jan 15 '12 at 16:40
    
Well, I'm starting to feel stupid here. We've been taught to consider $$D = aX + bY$$ which in this case should be $$D = -X + 0.7Y$$, but then I still get $$2ab\rho\sigma\sigma = 2 \cdot 0.7 \cdot -1 \cdot 0.8 \cdot 6 \cdot 7$$ which is negative? –  vicvicvic Jan 15 '12 at 16:43
add comment

2 Answers

up vote 2 down vote accepted

You did not go wrong anywhere, your book did. Whomsoever worked out the answer in the book missed the negative sign on the $2ab\rho\sigma_X\sigma_Y$ and calculated $$\text{var}(0.7Y - X) = 60.01 + 47.04 = 107.05$$ instead of $$\text{var}(0.7Y - X) = 60.01 - 47.04 = 12.97.$$ Thus, the book answer is $\displaystyle F_Z\left(\frac{12.11}{\sqrt{107.05}}\right) = F_Z(1.17)$ instead of $\displaystyle F_Z\left(\frac{12.11}{\sqrt{12.97}}\right) = F_Z(3.36)$.

share|improve this answer
add comment

I can't see where you went wrong. It seems right to me: $$\eqalign{ \text{Var}(.7Y-X)&=\text{Var}(.7Y)+\text{Var}(-X)+2\text{Cov}( -X, .7Y)\cr &=(.7)^2\text{Var}( Y)+\text{Var}( X)+2(-1)(0.7)\text{Cov}( X, Y)\cr &=(.7)^2\text{Var}( Y)+\text{Var}( X)+2(-1)(0.7) [ \text{Corr}( X, Y)\sigma_X\sigma_Y]\cr &=(.7)^2 \cdot 7^2+6^2+2(-1)(0.7)(0.8)\cdot 6\cdot7\cr &=12.97. } $$

share|improve this answer
    
Hm. The formula I have is $$D = aX + bY$$ and I used a = -1, b = 0.7. Then (from my book), $$\sigma^2_D = a^2\sigma^2_X + b^2\sigma^2_Y + 2ab\rho\sigma_X\sigma_Y$$ –  vicvicvic Jan 15 '12 at 16:35
    
I'm not discounting the possibility that my book is wrong; it has an awful lot of errors... If you stand by your latest answer? –  vicvicvic Jan 15 '12 at 18:42
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.