This Exam P sample reference tests Joint Distributions. The benefit is 0 only when both discrete risks are zero, 50 when their maximum is one, and 100 otherwise. Those probabilities are 1/25, 3/25, and 21/25, so the expected benefit is 90, choice B.
These notes identify the calculation error associated with each wrong letter when that error is reproducible.
AThe value 70 is obtained by applying the capped schedule to just one uniform risk level: (0+50+100+100+100)/5=70. The contract depends on the larger of two levels.
CThe value 92 results if the pair (1,1) is mistakenly assigned the 100 benefit instead of 50. That one grid cell adds 50/25=2 to the correct mean.
DTreating every positive maximum as a full 100 payment gives an expectation of 96, whose nearest listed value is 95. This discards the intermediate 50 benefit tier.
EIgnoring the cap gives 50 E[max(X,Y)]. The maximum has mean 2.8 on this grid, producing 140 rather than the capped expectation.
Original practice · fully worked
Original variant: award from the smaller rating and a mismatch fee
Two independent reviewers each assign an integer rating chosen uniformly from 0, 1, 2, and 3. An award equals 40 times the smaller rating, plus 20 additional points when the two ratings differ. Calculate the expected award.
A 15 points
B 35 points
C 40 points
D 50 points
E 85 points
Variant answer in brief
The expected smaller rating is 14/16=0.875, and the ratings differ with probability 12/16. The expected award is 40(0.875)+20(0.75)=50 points, choice D.
Setup
Setup
Let U and V be the two ratings and write the award as the sum of a minimum term and a mismatch term.
A=40min(U,V)+201{U=V}
Model
Model
Evaluate the expected minimum by its positive integer tail probabilities and count unequal ordered pairs.
E[min(U,V)]=k=1∑3Pr(U≥k,V≥k)
E[min(U,V)]=169+164+161=1614
Pr(U=V)=1612
Compute
Compute
Apply linearity of expectation to the two award components.
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.