This Exam P sample reference tests Central Limit Theorem. This is a central-limit approximation for a sum of 100 iid two-point variables. The sum has mean 100 and standard deviation 10, so the threshold is 1.5 standard deviations above the mean and the upper tail is 0.0668, giving choice B.
These notes identify the calculation error associated with each wrong letter when that error is reproducible.
AThe value 0.005 is the normal tail beyond about 2.58 standard deviations. Reaching that z-score would require an aggregate variance near 34 rather than the verified value 100.
CThe value 0.144 comes from using the raw second moment E[X²]=2 as Var(X). That gives aggregate standard deviation √(200), z=1.0607, and an upper tail near 0.144.
DThe value 0.147 compounds the same raw-second-moment error with coarse rounding of z=1.0607 down to 1.05, whose upper tail is about 0.147.
EThe value 0.440 is the offered value nearest Φ(1.5)-Φ(0)=0.4332. That computes the region between the mean and the boundary rather than the region above the boundary.
Original practice · fully worked
Original variant: expected layer payment on an aggregate
A shared reserve receives 100 independent maintenance charges. Each charge has mean 10 credits and standard deviation 2 credits, while its full distribution is otherwise unspecified. Let T be their total. At the end of the period, the reserve pays the positive part of T minus 1020. Use a central-limit approximation to calculate the expected payment.
A 0.1587 credits
B 1.6663 credits
C 3.1731 credits
D 4.8394 credits
E 10.5027 credits
Variant answer in brief
The aggregate is approximately normal with mean 1000 and standard deviation 20, placing the payment trigger at z=1. The normal positive-part formula gives 20[phi(1)-(1-Φ(1))]=1.6663 credits, so choice B.
Setup
Setup
Aggregate the means and variances of the 100 independent charges.
E[T]=100(10)=1000
SD(T)=100(2)=20
Model
Model
Approximate T by a normal variable and express the trigger in standard-normal units.
T≈N(1000,202)
z=201020−1000=1
Compute
Compute
Use the normal expected-excess identity at a standardized trigger z.
E[(T−1020)+]≈20[ϕ(1)−{1−Φ(1)}]
=20(0.2419707−0.1586553)=1.6663094
Answer
Answer
The expected layer payment is about 1.6663 credits.
The 3108-page Probability Proof Manual reorganizes 718 verified Exam P solutions by syllabus skill and adds formula proofs, error patterns, and original worked practice.