The question: Consider an n×n grid. The grid squares from the middle row are shaded in. What is the probability that a randomly selected rectangle contains at least one shaded square?
My (incorrect) method:
Using complementary probability, P(contains one shaded square) = 1 - P(contains no shaded square)
P(contains no shaded square) = # of rectangles with no shaded square / # of possible rectangles
In general, we can generate a rectangle by choosing any x,y such that 0 < x,y <= n.
The # of rectangles with no shaded square in the top half = $n *(n-1)/2$, we multiply by 2 by symmetry to get $n(n-1)$ total rectangles with no shaded square in both halves.
The # of total rectangles is just $n^2$, so $1-(n-1)/n = 1/n$.
What is specifically wrong with this counting method? I'm overcounting something, but I'm not sure what. I realize that in both the numerator and denominator, I am counting (for example) a 5x6 rectangle and a 6x5 rectangle separately. Is this an issue?