The Geometric Distribution

Picture flipping a coin, rolling a die, or inspecting parts off a line, over and over, waiting for one particular outcome. The geometric distribution answers: what's the chance that outcome first shows up exactly on trial k? Because every trial before the successful one has to be a failure, and failures are less and less likely to stack up the longer you wait, the probabilities only shrink as k grows. That's why the distribution slides downhill: k=1 is always the most likely single value, and every later value is smaller than the one before it.

The mechanism is just the multiplication rule for independent events, applied trial by trial. If success has probability p on each trial, then failure has probability (1-p). To have the first success land on trial k, you need failure on trials 1 through k-1, then success on trial k. Multiply those probabilities together: P(X=k) = (1-p)^k-1 times p. Nothing fancier than that. The trials have to be independent and p has to stay constant, which is why this only applies to repeated identical chance processes, not sampling without replacement from a small population.

This distribution sits right next to the binomial in the unit because both come from the same kind of repeated, independent trials with constant success probability p. The binomial fixes the number of trials and counts successes; the geometric fixes the goal as one success and lets the number of trials vary. Later, when you meet expected value for random variables, the geometric distribution gives a clean formula, mean equals 1/p, that reappears whenever you're estimating how long you'll wait for something rare.