Original practice · fully worked
Original variant: predict a second roll from partial evidence
A simulation selects a fair six-sided die with probability two-thirds and a fair eight-sided die otherwise. The selected die is rolled twice. The first roll is greater than four. Given this result, calculate the probability that the second roll is also greater than four.
- A
0.333 - B
0.389 - C
0.405 - D
0.429 - E
0.500
Answer and calculation
The first result updates the die probabilities to four-sevenths for the six-sided die and three-sevenths for the eight-sided die. Averaging their chances of another result above four gives 17/42, approximately 0.405, and choice C.
Setup
Setup
Let E denote a roll greater than four. Calculate its likelihood under each possible die.
Pr(E∣D6)=62=31 Pr(E∣D8)=84=21 Model
Model
Use the first roll to update the probability assigned to each die.
Pr(E)=3231+3121=187 Pr(D6∣E)=74,Pr(D8∣E)=73 Compute
Compute
Conditional on the selected die, the two rolls are independent. Average the second-roll likelihoods over the updated die probabilities.
Pr(E2∣E1)=7431+7321 =4217=0.4047619048… Answer
Answer
The second roll exceeds four with conditional probability approximately 0.405.
0.405(C)