This Exam P sample reference tests Bayes' Theorem. The three-year failure probabilities are 1-exp(-3/2) for type A and 1-exp(-3/4) for type B. Bayes' theorem with prior weights 0.10 and 0.90 gives posterior type-A probability 0.140595, so choice B is correct.
These notes identify the calculation error associated with each wrong letter when that error is reproducible.
AThis is the joint probability of type A and early failure, about 0.0777. It is the Bayes numerator before division by the total early-failure probability.
CThis is the probability that a type-B item survives three years, exp(-3/4). It reverses the observed event and omits Bayes' theorem.
DThis is close to the overall early-failure probability, the denominator that normalizes the two weighted likelihoods. It is not a type posterior.
EThis is the likelihood of early failure for type A, 1-exp(-3/2), without accounting for the small prior share of that type.
Original practice · fully worked
Original variant: one module failure by inspection
A research rack contains four independent modules whose lifetimes are exponential with mean 5 years. At an inspection two years after installation, calculate the probability that exactly one module has failed.
A 0.2019
B 0.3297
C 0.3972
D 0.7981
E 1.3187
Variant answer in brief
Each module fails by year two with probability 1-exp(-0.4). The number failed is binomial with four trials, giving exactly-one probability 4(1-exp(-0.4))exp(-1.2)=0.397191 and choice C.
Setup
Setup
Compute the failure and survival probabilities for one module at the inspection time.
p=Pr(T≤2)=1−e−2/5=1−e−0.4
1−p=e−0.4
Model
Model
Independence makes the number of failed modules at year two binomial.
K∼Binomial(4,1−e−0.4)
Compute
Compute
Choose which one module failed and require the other three to survive.
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.