This Exam P sample reference tests Poisson Distribution. For a Poisson count, a zero-count probability q implies mean -ln(q). Applying this to the three groups and weighting by their population sizes gives 36.134873 thousand expected claims, which rounds to 36 and choice C.
These notes identify the calculation error associated with each wrong letter when that error is reproducible.
AThis is approximately the expected number of policyholders with exactly one claim: sum over groups of n_g lambda_g exp(-lambda_g), which is 21.39 thousand. It does not count additional claims.
BThis is the expected number of policyholders with at least one claim, 20(0.30)+45(0.10)+35(0.50)=28 thousand. A policyholder can make more than one claim.
DThis subtracts the expected claim count from the 100-thousand policy count, giving about 64. Those quantities have different meanings and are not complements.
EThis is the number of policyholders expected to have no claims, 20(0.70)+45(0.90)+35(0.50)=72 thousand.
Original practice · fully worked
Original variant: infer an array size from silent sensors
An array contains identical independent sensors. The daily alert count from each sensor is Poisson, and one sensor has no alerts on a day with probability 0.82. The expected total number of daily alerts from the array is 29.76764. Calculate the number of sensors in the array.
A 30
B 36
C 150
D 165
E 345
Variant answer in brief
The zero-alert probability gives a per-sensor mean of -ln(0.82)=0.198451. Dividing the array's expected total by that mean gives 149.999996, so the array contains 150 sensors and choice C.
Setup
Setup
Use the Poisson zero-count identity to determine one sensor's mean alert rate.
0.82=e−λ
λ=−ln(0.82)=0.1984509387…
Model
Model
For n identical sensors, linearity of expectation makes the array mean n times the individual mean.
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.