AMC 10 · 2005 · #14

Grade 7 probability
probability-basicconditional-probabilityparity caseworkidentify-subproblems ↑ Prerequisites: probability-basicfraction-arithmeticparity
📏 Medium solution 💡 3 insights
Problem
One dot is picked at random from all the dots on a standard die and erased. Then the die is rolled fairly. Find the probability that the top face shows an odd number of dots.

Pick an answer.

(A)
$\frac{5}{11}$
(B)
$\frac{10}{21}$
(C)
$\frac{1}{2}$
(D)
$\frac{11}{21}$
(E)
$\frac{6}{11}$
How to solve
Strategy Identify Subproblems

The experiment has two stages, and the second stage depends on the first. Splitting it into subproblems by asking only one thing about stage one -- was the damaged face odd or even? -- is enough, because that single fact determines the whole answer for stage two. The important care point is that the damaged face is chosen in proportion to how many dots it carries, so the two cases are not equally likely and must be weighted by dot counts, not by face counts.

1STEP 1

Weight the cases by dots, not faces

The dot is uniform, not the face, so each face carries its own weight.

P(odd face damaged)=(1+3+5)/21=3/7, P(even face damaged)=(2+4+6)/21=4/7
2STEP 2

Removing a dot flips one parity

Removing one dot flips exactly one face's parity.

k odd→#{odd faces}=2; k even→#{odd faces}=4
3STEP 3

Probability inside each case

So the odd-face count becomes either 2 or 4.

P(top odd ∣ odd damaged)=2/6=1/3, P(top odd ∣ even damaged)=4/6=2/3
4STEP 4

Combine the two cases

Weighting the two cases and adding gives 11/21.

3/7·1/3+4/7·2/3=1/7+8/21=3/21+8/21=11/21
5STEP 5

Check by counting all outcomes

A flat count of all outcomes confirms 11/21, choice (D).

(9 · 2+12 · 4)/(21 · 6)=66/126=11/21
Answer
11/21
The answer must sit a little above 1/2. On an untouched die the chance is exactly 1/2, and removing a dot is more likely to hit an even face (which carries more dots), which pushes the count of odd faces up. So the answer should beat 1/2 but stay well below 2/3, and 11/21≈ 0.524 fits. Choice (C) 1/2 is the trap for forgetting that the removal is biased toward even faces; choice (B) 10/21 is the same slip in the wrong direction. Denominators of 11 in choices (A) and (E) do not match a 21-dot, 6-face experiment.
💡Key takeaway

The dot is chosen, not the face, so faces with more dots are more likely to be hit; weight each case by its dot count instead of treating the six faces as equally likely.

  • Weight the cases by dots, not faces
  • Removing a dot flips one parity
  • Probability inside each case
  • Combine the two cases
  • Check by counting all outcomes