AMC 10 · 2010 · #18

Grade 12 probability
geometric-probabilityvector-additionexpected-valuetrigonometric-ratios identify-subproblemsconvert-to-algebra ↑ Prerequisites: geometric-probabilityvector-addition
📏 Long solution 💡 3 insights
Problem
Three equal jumps head in independent random directions. Find the chance of landing near the start.

Pick an answer.

(A)
$\dfrac{1}{6}$
(B)
$\dfrac{1}{5}$
(C)
$\dfrac{1}{4}$
(D)
$\dfrac{1}{3}$
(E)
$\dfrac{1}{2}$
How to solve
Strategy Identify Subproblems

Three random directions at once is too much to hold. Tool #7 (Identify Subproblems) splits the trip: first glue jumps 1 and 2 into a single step vec S of random length R, then ask what the third jump must do. Tool #4 (Introduce a Variable) names the angle between the first two jumps, which is all R depends on, and Tool #13 (Convert to Algebra) turns "lands within 1 meter" into a plain inequality on a cosine. Tool #1 (Draw a Diagram) reads that inequality off the direction circle as an arc, turning it into a probability. The load-bearing move is Tool #15 (Organize Information in More Ways): switch from the three absolute directions to the pair (length of the first two jumps, direction of the third jump measured against vec S), because in those coordinates the two pieces are independent. Tool #5 (Look for a Pattern) finishes: the resulting conditional probability is a straight-line tent in the half-angle, so its average needs no work at all.

1STEP 1

Turn jumps into vectors

The jumps become three unit vectors.

P=Pr(|vec u₁+vec u₂+vec u₃| ≤ 1), |vec u₁|=|vec u₂|=|vec u₃|=1
2STEP 2

Glue the first two jumps

Gluing the first two leaves one length.

vec S=vec u₁+vec u₂, R=|vec S|∈[0,2], final displacement=vec S+vec u₃
3STEP 3

Length after two jumps

That length rides on the angle between them.

R²=|vec u₁|²+|vec u₂|²+2 vec u₁·vec u₂=2+2cosθ=4cos² θ/2 ⟹ R=2|cosθ/2|
4STEP 4

The last direction is still uniform

The last direction is still uniform and independent.

ψ=arg(vec u₃)-arg(vec S) is uniform on [0,2π) and independent of R
5STEP 5

When does the frog land home

Landing home becomes one cosine condition.

|vec S+vec u₃|²=R²+1+2Rcosψ ≤ 1⇔ 2Rcosψ ≤ -R²⇔ cosψ ≤ -R/2
6STEP 6

Read the condition as an arc

That condition reads as an arc of the circle.

Pr(cosψ ≤ -r/2)=2arccos(r/2)/2π=arccos(r/2)/π
7STEP 7

Average the tent function

Averaging a tent-shaped function gives a quarter turn.

arccos|cos t|= t,&0 ≤ t ≤ π/2 ; [2pt]π-t,&π/2 ≤ t < π ⟹ E[arccos(R/2)]=1/2·π/2=π/4
8STEP 8

Assemble the probability

The probability is 1/4, choice (C).

P=E[arccos(R/2)/π]=1/π·π/4=1/4 ⟹ (C)
Answer
1/4
Three internal checks agree. First, the conditional formula arccos(r/2)/π is correct at both extremes by inspection: 1/2 when the first two jumps cancel (r=0), and 0 when they line up (r=2). Second, the same machinery run on two jumps instead of three gives Pr(2|cosθ/2| ≤ 1)=1/3, so 1/4 < 1/3 is the right direction — an extra random jump spreads the frog out further. Third, a completely different route (below) reaches 1/4 with no integration at all. One popular shortcut deserves a warning: "ignore the first jump; the last two jumps land uniformly on a disk of radius 2, so the answer is the area ratio π/4π=1/4" gets the right number for the wrong reason. The sum of two unit vectors is not uniform on that disk — its radial density is 2/(π√(4-r²)), which piles up near r=2, whereas a uniform disk would give density proportional to r and vanish at r=0. Rerun that shortcut with a target radius of 1/2 and it predicts 1/16=0.0625, while the correct value is about 0.0481. The agreement at radius 1 is a coincidence, not a proof.
💡Key takeaway

Fold the first two jumps into one step: the chance the last jump brings the frog home is a straight ramp up and back down as that folded step shrinks, and the average of a ramp is just its midpoint, giving 1/4.

  • Turn jumps into vectors
  • Glue the first two jumps
  • Length after two jumps
  • The last direction is still uniform
  • When does the frog land home
  • Read the condition as an arc
  • Average the tent function
  • Assemble the probability