Linear Combinations of Independent Random Variables
Variance of Independent Sums
This Exam P sample reference tests Normal Distribution. The difference of the two independent normal variables is normal with mean zero and variance 16. The required interval is therefore one quarter of a standard deviation on either side of zero, giving probability about 0.197 and choice A.
Form the prediction error as the difference of the two normal variables.
D=W−T
Model
Model
Independence makes the variances add even though the variables are subtracted.
E[D]=0
Var(D)=4+12=16,D∼N(0,16)
Compute
Compute
Standardize both endpoints of the symmetric interval.
Pr(∣W−T∣<1)=Pr(∣Z∣<41)
=2Φ(0.25)−1=0.1974126514…
Answer
Answer
The probability rounds to 0.20.
0.20(A)
Test the method on unseen problems
Probability Multivariate & Risk Models Practice Workbook
30 original multivariate random variables problems with complete solutions and wrong-choice analysis. Written for Exam P, not copied from the SOA set. One-time $19 PDF.
These notes identify the calculation error associated with each wrong letter when that error is reproducible.
BThe value 0.23 is close to using only the predictor's standard deviation and ignoring the claim variance.
CThe value 0.38 is the central probability within one half standard deviation, obtained by using only the claim standard deviation.
DThe value 0.60 is the one-sided cumulative probability at the upper standardized endpoint. It fails to subtract the probability below the lower endpoint.
EThe value 0.68 is the central probability within one full standard deviation, treating the raw threshold one as already standardized.
Original practice · fully worked
Original variant: conditional normal sensor reading
Independent normal sensor readings X and Y have means 10 and 6 and variances 4 and 12, respectively. Their combined reading X+Y is observed to equal 20. Calculate the conditional probability that X exceeds 10.
A 0.282
B 0.599
C 0.691
D 0.718
E 0.841
Answer and calculation
Given the total, X is normal with conditional mean 11 and variance 3. The threshold 10 is one divided by the square root of 3 standard deviations below that mean, giving probability approximately 0.718 and choice D.
Setup
Setup
Let S be the observed sum and calculate its moments with X.
S=X+Y,E[S]=16
Var(S)=16,Cov(X,S)=4
Model
Model
Apply the conditional normal mean and variance formulas.
E[X∣S=20]=10+164(20−16)=11
Var(X∣S)=4−1642=3
Compute
Compute
Standardize the threshold under the conditional distribution.
Pr(X>10∣S=20)=Φ(31)
=0.7181485692…
Answer
Answer
The conditional exceedance probability is approximately 0.718.