Chapter 11: The Poisson Distribution

The Poisson distribution serves as a fundamental probability model for analyzing experiments involving the random counting of discrete events occurring at a known average rate, such as radioactive decay or cosmic ray detection in physics. This distribution emerges as a limiting case of the binomial distribution when the number of trials becomes exceptionally large while the probability of individual success remains very small, making it mathematically elegant and practical for real-world experimental scenarios. The distribution is completely defined by a single parameter mu, representing the expected average count across repeated trials, and the probability of observing exactly v events is given by the Poisson function involving e to the negative mu multiplied by mu to the v divided by v factorial. A critical feature of the Poisson distribution is that both its mean and variance are equal to mu, leading directly to the square-root rule, which states that when a counting experiment yields v observed events, the uncertainty in that measurement is simply the square root of v, reported as v plus or minus the square root of v. This rule applies specifically to the actual counted numbers rather than derived quantities like rates, which require proper error propagation techniques. As the parameter mu grows large, the Poisson distribution progressively approximates a Gaussian distribution, becoming increasingly symmetric and bell-shaped, allowing experimenters to use simpler normal distribution calculations when mu is sufficiently large. The Poisson framework enables hypothesis testing by establishing whether observed counts fall within expected statistical fluctuations, with deviations greater than three standard deviations indicating potential systematic problems or false initial assumptions. In practical experimental physics, background noise from extraneous sources must be carefully isolated and subtracted from total measurements through separate background counting sessions, with uncertainties properly propagated through the subtraction process using quadrature combination of individual rate errors derived from the square-root rule applied to each counted quantity.