This Exam P sample reference tests Poisson Distribution. Conditioning removes the zero outcome and requires renormalization by 1-exp(-3). The conditional first and second moments give variance 2.6609, choice C.
These notes identify the calculation error associated with each wrong letter when that error is reproducible.
AA value near 1.63 results from over-truncating the positive Poisson tail, effectively treating counts of three or more as the conditioning threshold. The event here retains every positive count.
BThe value 1.73 understates the second conditional moment; substituting it as the variance would imply E[N squared|A]=1.73+3.1572 squared=11.697, not the required 12.6287.
DThe value 3.00 is the unconditional Poisson variance. Removing the zero outcome changes both conditional moments.
EThe value 3.16 is the conditional mean 3/(1-exp(-3)), not the conditional variance.
Original practice · fully worked
Original variant: variance after an even-code filter
A diagnostic code K takes values 1, 2, 3, and 4 with probability P(K=k)=k/10. A quality report is retained only when K is even. Given that a report is retained, calculate Var(K).
A 0.6000
B 0.8889
C 1.0000
D 3.3333
E 12.0000
Variant answer in brief
The retained values are 2 and 4 with conditional probabilities 1/3 and 2/3. Their conditional mean is 10/3 and second moment is 12, so the variance is 8/9=0.8889, choice B.
Setup
Setup
Let R be the event that the code is even and compute its probability from the two retained codes.
R={K∈{2,4}}
Pr(R)=102+104=53
Model
Model
Normalize the probabilities of the two retained values.
Pr(K=2∣R)=6/102/10=31
Pr(K=4∣R)=6/104/10=32
Compute
Compute
Compute the first two conditional moments and subtract the squared 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.