A point is chosen at random inside square . The probability that is neither the shortest nor the longest side of can be written as , where and are positive integers, , and is not divisible by the square of a prime. What is ?
- A)
- B)
- C)
- D)
- E)
Answer
A
Key insight
AP is the middle side exactly when P lies right of the bisector x = 1/2 and inside A's unit circle, or left of it and outside.
Solution
Take the square to have side , with , , , , and let with . The three sides of are , and . Since is not on line , the triangle is genuine, and "neither shortest nor longest" simply means is strictly the middle one of the three lengths. So either
Two loci do all the work. is the perpendicular bisector , with on the right and on the left. is the quarter of the unit circle centred at that lies in the square, with inside it.
So Case 1 is the part of the quarter disc with , and Case 2 is the part of the left half of the square lying outside the quarter disc.
Let be the area of the piece of the quarter disc with . The quarter disc has area and the left half of the square has area , so
and the probability sought is .
Computing . The line meets the unit circle at , a point at angle from . Split along the radius to that point:
- the right triangle with vertices , , , of area ;
- the circular sector from to , of area .
Hence , and
The area of the square is , so this is the probability; numerically it is about , which is a plausible size for the region. Reading off , , , , (and , squarefree),
The answer is .
Why this works
Ordering three lengths is the same as deciding, for each pair, which is larger, and each such comparison is a single curve in the plane: versus is a perpendicular bisector, versus is a circle centred at . Two curves cut the square into four regions, and " in the middle" is exactly two of them — one on each side of the bisector. Describing the answer as "quarter disc minus " plus "half square minus " avoids computing either region separately and makes the coefficient , hence the single and the single , appear at once. No calculus is needed: a region bounded by two radii and an arc is a sector, and everything left over is a triangle.
Alternative approach
Compute the two regions directly with the sector-plus-triangle decomposition. Case 1 is the quarter disc minus , that is . Case 2 is the thin region tucked into the corner near : the left half-square minus , i.e. . Their sum is , agreeing with . Seeing Case 2 explicitly is worth the extra minute, because its tiny area is what makes it easy to forget.
The trap
Handling only the case BP < AP < AB and missing the sliver near D where AB < AP < BP, which leaves pi/6 - sqrt3/8 and cannot even be written in the required form.
Common mistakes
- Handling only the case BP < AP < AB and missing the sliver near D where AB < AP < BP, which leaves pi/6 - sqrt3/8 and cannot even be written in the required form.
- Getting the perpendicular bisector the wrong way round and pairing "" with "inside the circle", which computes the complementary regions instead.
- Taking "neither shortest nor longest" to allow ties and worrying about the boundaries and ; these are curves of zero area and change nothing.
Techniques
Split into exhaustive cases and handle each · Cut the figure into known shapes (triangles, rectangles, sectors)