{
  "slug": "exponential-distribution",
  "term": "Exponential distribution",
  "aliases": [
    "Exponential lifetime model",
    "Negative exponential distribution"
  ],
  "definition": "An exponential distribution is a continuous probability model whose unshifted form describes nonnegative lifetimes or waiting times with constant hazard.",
  "note": "For the unshifted model with rate lambda > 0 and elapsed time t >= 0, S(t) = exp(-lambda t), F(t) = 1 - exp(-lambda t), and h(t) = lambda. Lambda has inverse-time units; lambda t is dimensionless. Rate lambda and scale 1/lambda are different parameter conventions. [nist-exponential]\n\nThe mean is 1/lambda, but the median is ln(2)/lambda. Substituting the mean into the CDF gives 1 - exp(-1), not one half. Neither summary promises an individual service life. A shifted model additionally specifies a location parameter and a different time origin. [nist-exponential-parameters]\n\nFor a >= 0 and u >= 0, the unshifted model gives P(T > a + u | T > a) = S(a + u)/S(a) = exp(-lambda u). This derived memoryless property concerns additional lifetime conditional on survival; it is not evidence that devices avoid wear. Two-parameter Weibull shape one gives this exponential case. [nist-exponential] [nist-weibull]\n\nA homogeneous Poisson process requires independent, identically distributed exponential interarrival times. Exponential individual lifetimes alone establish neither independence between devices nor a Poisson fleet-replacement stream. State the event, population and clock before transferring a rate. These are model distinctions, not a fitted storage-fleet result. [nist-hpp]\n\nFor a fixed initial cohort of N independent items with common exponential rate lambda, each counted only at its first failure and never replaced, the failure count by t is binomial with probability 1 - exp(-lambda t). Its expected count is N(1 - exp(-lambda t)), so its expected event rate is N lambda exp(-lambda t), not a constant. This is an analytic deduction from the lifetime CDF under the stated cohort assumptions, not a fitted fleet model. [nist-exponential]",
  "status": "shipping",
  "advantages": [],
  "limitations": [],
  "related": [
    "weibull-distribution",
    "survival-function",
    "hazard-rate",
    "mean-residual-life",
    "poisson-rate-model",
    "interarrival-time"
  ],
  "supersedes": [],
  "supersededBy": [],
  "firstSeen": null,
  "lastSeen": null,
  "citations": [
    {
      "id": "nist-exponential",
      "title": "NIST e-Handbook 8.1.6.1: Exponential lifetime formulas",
      "url": "https://www.itl.nist.gov/div898/handbook/apr/section1/apr161.htm",
      "accessed": "2026-11-18"
    },
    {
      "id": "nist-exponential-parameters",
      "title": "NIST e-Handbook 1.3.6.6.7: Exponential distribution, location, scale and quantiles",
      "url": "https://www.itl.nist.gov/div898/handbook/eda/section3/eda3667.htm",
      "accessed": "2026-11-18"
    },
    {
      "id": "nist-weibull",
      "title": "NIST e-Handbook 8.1.6.2: Weibull and its exponential special case",
      "url": "https://www.itl.nist.gov/div898/handbook/apr/section1/apr162.htm",
      "accessed": "2026-11-18"
    },
    {
      "id": "nist-hpp",
      "title": "NIST e-Handbook 8.1.7.1: Homogeneous Poisson Process assumptions",
      "url": "https://www.itl.nist.gov/div898/handbook/apr/section1/apr171.htm",
      "accessed": "2026-11-18"
    }
  ],
  "updated": "2026-11-18"
}
