This Exam P sample reference tests Poisson Distribution. The expected number of events beyond the first two is E[(N-2)+]=-1+3exp(-1)=0.10364. Multiplying by 1000 gives about 104, choice B.
These notes identify the calculation error associated with each wrong letter when that error is reproducible.
AThe value 80 is 1000P(N≥3), rounded from 80.30. It pays only one unit whenever the threshold is crossed and ignores additional units for counts four and above.
CThe value 368 is 1000E[(N-1)+]=1000exp(-1). It applies a one-event retention instead of the stated two-event retention.
DThe value 512 implies an expected paid count of 0.512, which exceeds E[(N-1)+]=0.368 even though a second retained event can only reduce payment. It adds rather than removes retained-event exposure.
EThe value 632 is 1000P(N≥1)=1000(1-exp(-1)). It pays once for any positive count and ignores the two-event retention.
Original practice · fully worked
Original variant: variance of a threshold-triggered payment
The number N of system alerts in a week is Poisson with mean 2. A maintenance contract pays 100 credits if at least two alerts occur and pays nothing otherwise. Calculate the variance of the weekly payment, in squared credits.
A 0.594
B 59.399
C 100.000
D 2411.651
E 5939.942
Variant answer in brief
A payment occurs with probability 1-3 exp(−2)=0.593994. The payment is 100 times a Bernoulli indicator, so its variance is 100 squared times p(1-p)=2411.651, choice D.
Setup
Setup
Represent the payment by an indicator for crossing the alert threshold.
N∼Poisson(2)
Y=1001{N≥2}
Model
Model
Evaluate the Bernoulli success probability for the payment indicator.
p=Pr(N≥2)=1−Pr(N=0)−Pr(N=1)
p=1−3e−2=0.5939942
Compute
Compute
Scale the Bernoulli variance by the square of the payment amount.
Var(Y)=1002p(1−p)
Var(Y)=10000(0.5939942)(0.4060058)=2411.6510
Answer
Answer
The weekly payment variance is approximately 2411.651 squared 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.