What is the binomial distribution?

The binomial distribution is a discrete probability distribution that models the number of successes in a fixed number of independent trials, where each trial has only two possible outcomes—commonly termed "success" and "failure"—and the probability of success remains constant for every trial. It is defined by two parameters: *n*, the number of trials, and *p*, the probability of success on a single trial. The probability of observing exactly *k* successes is given by the formula P(X = k) = C(n, k) * p^k * (1-p)^(n-k), where C(n, k) is the binomial coefficient representing the number of ways to choose *k* successes from *n* trials. This mathematical structure makes it the fundamental model for processes ranging from quality control sampling and clinical trial analysis to simple games of chance, provided the core assumptions of independence and constant probability hold.

The mechanism of the distribution hinges on the Bernoulli process, where each trial is a Bernoulli random variable. The independence of trials means the outcome of one trial does not influence another, while the constant probability *p* ensures the process is identically distributed across trials. When these conditions are met, the binomial distribution provides a complete probabilistic description of the total count of successes. Its shape is determined by its parameters: the distribution is symmetric when *p* = 0.5, skewed right when *p* is small, and skewed left when *p* is large, with the skewness diminishing as *n* increases. Key properties include an expected value or mean of *np*, a variance of *np(1-p)*, and a standard deviation that increases with *n* but as a proportion of *n* decreases, illustrating the law of large numbers.

In practical application, the binomial distribution is used for inference and decision-making under uncertainty. For instance, in manufacturing, if a production line has a known defect rate *p*, the distribution can calculate the probability of finding more than a certain number of defective items in a shipment of *n* units, informing quality assurance protocols. In finance, it underpins the binomial options pricing model by representing possible price movements over discrete time steps. A critical analytical implication is that real-world data can violate the model's assumptions; dependence between trials or a non-constant *p* will render binomial probabilities inaccurate. Therefore, verifying the assumptions is as important as the calculation itself.

The distribution's broader significance lies in its relationship to other statistical concepts. For large *n*, it is approximated by the normal distribution, facilitating the use of confidence intervals and hypothesis tests for proportions. It is also a building block for the negative binomial and hypergeometric distributions. Its limitations define its appropriate domain: it is not suitable for modeling the number of trials until a success, which is geometric, or for sampling without replacement from a finite population, which requires the hypergeometric distribution. Thus, the binomial distribution serves as a precise tool for a specific, widely encountered class of random phenomena, and its correct application demands a clear understanding of the experimental or observational process it aims to represent.