Probability related to Normal Distribution

1. Jan 21, 2013

songoku

1. The problem statement, all variables and given/known data
Variable A has mean 55 and variance 9, variable B has mean 65 and variance 25. If A and B are normally distributed, find P (B > A)

2. Relevant equations
z = (x - μ) / σ

3. The attempt at a solution
Can this be solved? What is the meaning of P (B > A)? The probability of random variable B is bigger than random variable A?

Thanks

2. Jan 22, 2013

hikaru1221

I suppose the problem assumes A and B are statistically independent. Then the formal way to do this is:
$P(B>A) = \int_{b=-\infty}^{+\infty} \int_{a=-\infty}^{b}f_A(a)f_B(b)dadb$
This is a very nasty integral.

3. Jan 22, 2013

Ray Vickson

Yes, P(B > A) means exactly what you said. It can be solved quite easily if A and B are independent, but I cannot give more details now: first you have to make a serious attempt to solve it yourself.

4. Jan 22, 2013

songoku

Sorry I don't get the idea how to do the integration and why can it be like that

I don't have idea how to start. The question I saw until now always contained number such as P (B > 23). Please give me clue how to start solving this problem. What is the first step I need to take? I can't find the value of z because I don't know the value of the random variable.

Thanks

5. Jan 22, 2013

Ray Vickson

Look at the random variable Y = B - A.

6. Jan 22, 2013

songoku

OK maybe I get it

Let: Y = B - A

E(Y) = E(B) - E(A) = 10
σ(Y) = σ(B) + σ(A) = 34

P(B > A) = P(B - A > 0) = P(Y > 0) = P(Z > -1.715) = 0.9568

Do I get it right?

Thanks

7. Jan 22, 2013

Ray Vickson

It looks right.

8. Jan 23, 2013

songoku

ok thanks

Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook