St · Statistics & Probability

What Is the Probability the Square’s Center Lies Inside the Circle?

Published By the 9to5Science editorial team

At a glance

  1. 1. Interpret the circle

    Assume the two independently chosen points are the endpoints of the circle’s diameter.

  2. 2. Use the diameter test

    The center lies inside that circle exactly when the vectors to the two points have a nonpositive dot product.

  3. 3. Fix one point

    For any fixed first point, the square is split into equal areas by the line through its center perpendicular to that point.

  4. 4. State the probability

    The second point lands on the required half of the square with probability one-half.

Advertisement

Assume the two points are chosen independently and uniformly from the square, and that the circle has the segment joining them as its diameter. Under that standard interpretation, the probability that the square’s center lies inside the circle is 1/21/2.

The wording matters: two distinct points do not determine a unique circle unless an additional condition is given. If the points are meant only to lie somewhere on the circle, there are infinitely many possible circles. The calculation below applies when the points are the diameter’s endpoints.

The geometric test

Let OO be the square’s center, and let AA and BB be the two random points. Translate coordinates so that O=(0,0)O=(0,0). The circle with diameter ABAB has center (A+B)/2(A+B)/2 and radius ∣A−B∣/2|A-B|/2. Thus OO is inside or on the circle precisely when

∣A+B2∣2≤∣A−B∣24.\left|\frac{A+B}{2}\right|^2 \leq \frac{|A-B|^2}{4}.

Multiplying by 44 and expanding both squared lengths gives

∣A+B∣2≤∣A−B∣2⟺A⋅B≤0.|A+B|^2 \leq |A-B|^2 \quad\Longleftrightarrow\quad A\cdot B \leq 0.

So the center is inside the diameter circle exactly when the vectors from the square’s center to the two points meet at a right or obtuse angle.

Why the probability is one-half

Fix the first point AA. If AA is not the center, the condition A⋅B≤0A\cdot B\leq 0 says that BB must lie on the side of the line A⋅B=0A\cdot B=0 where the dot product is negative. This line passes through the square’s center and is perpendicular to the vector AA.

A square is centrally symmetric: if a point BB is in the square, then so is its opposite point −B-B. The map B↦−BB\mapsto -B pairs points with A⋅B<0A\cdot B<0 with points having A⋅(−B)>0A\cdot(-B)>0. Therefore, the line divides the square into two regions of equal area. Conditional on any fixed noncentral AA, a uniformly chosen BB has probability 1/21/2 of being in the required half.

The first point lands exactly at the square’s center with probability zero. Likewise, the second point landing on the dividing line has probability zero, so including the circle’s boundary does not change the result. Averaging over the possible first points gives the same probability: 1/21/2.

A concrete check

Take a square centered at the origin and fix A=(1,0)A=(1,0). The dot-product condition becomes

A⋅B=(1,0)⋅(bx,by)=bx≤0.A\cdot B=(1,0)\cdot (b_x,b_y)=b_x\leq 0.

The second point must be in the left half of the square, which has half its area. If instead A=(1,1)A=(1,1), the condition is bx+by≤0b_x+b_y\leq 0. The dividing line is diagonal, but central symmetry still makes the two regions equal in area. The direction of the first point changes the dividing line, not the probability.

Common mistakes

  • Assuming the answer depends on the square’s side length. It does not: scaling the square does not change the condition that the dot product is nonpositive.
  • Treating the points as independent directions with uniform angles. A point chosen uniformly in a square does not have a uniformly distributed angle. The half-area argument avoids needing the angle distribution.
  • Using the test for a different circle. The dot-product test applies to the circle whose diameter is ABAB. If the points merely lie on an unspecified circle, the event is not defined until the circle is specified.
  • Counting the boundary as a substantial extra case. The center lies exactly on the circle when A⋅B=0A\cdot B=0. This requires the second point to lie on a line, an area-zero set, so it adds no probability.

The same half-probability result holds for any centrally symmetric region centered at OO, not just a square, provided the two points are independently and uniformly chosen from that region and define a circle with their connecting segment as diameter.

Spotted an error? Send a correction. Read how we write and check articles in our editorial policy.