Geometric probability

What just happened?

We draw a point uniformly from a square of side 1 (area = 1), with a circle of radius $r$ inscribed in it — this is the favorable region. The probability of landing in the circle does not depend on counting cases (as with a die roll), because there are infinitely many points. We compute it as the ratio of areas — and that is exactly the share of blue points among all of them.

Geometric probability:

Here the favorable area is $\pi r^2$ (circle), and the total area is $1$ (square), so $P = \pi r^2$. This extends the classical definition of probability to cases with infinitely many equally likely outcomes.

The invariant — why NOT always ½:

The probability changes together with the size of the favorable region — it equals the ratio of measures, not a constant ½. The key assumption is the uniformity of the sampling: every point of the figure is equally likely. Analogously we compute it for lengths on a segment ($\text{length}/\text{length}$) or volumes in space.

Where did this concept come from?

The first example of geometric probability was the famous Buffon’s needle problem (Georges-Louis Leclerc de Buffon, 18th century): we toss a needle onto a floor of parallel boards and ask what the chance is that the needle crosses a line between the boards. The answer leads directly to a ratio of measures (and surprisingly — to the number $\pi$). This idea also underlies Monte Carlo methods — estimating quantities through random sampling.

What does this mean in practice?

When the “event space” is continuous (a random point, a random moment in time, a random length), you cannot count cases — instead we draw regions and divide their measures. The scattered grid of points in the simulation only estimates this ratio (hits/N ≈ P), but the formula itself is exact.

Why you need this for the exam:

In geometric-probability problems the trick is always the same: draw the total region and the favorable region, compute their measures (area, length, volume) and divide one by the other. Watch out for the uniformity assumption — without it the formula does not work.

Remember: with uniform sampling, probability = measure of the favorable region / measure of the total region. Not ½ “by eye” — the exact ratio of areas.