{
  "slug": "risk-set",
  "term": "Risk set",
  "aliases": [
    "At-risk set",
    "Number at risk"
  ],
  "definition": "A risk set contains the observed units eligible to experience the specified event immediately before a given analysis time.",
  "note": "In a first-failure analysis, its size n_i is the denominator for events d_i at that time, not the fleet's starting size, final headcount or number of telemetry rows. Earlier failures and withdrawals reduce subsequent membership. [km-1958]\n\nWith the usual failure-before-censoring convention for tied recorded times, a device censored at an event time still contributes to that event's denominator. This convention does not recover the order hidden by coarse timestamps. [km-1958]\n\nDelayed entry requires accounting for when observation begins on the chosen clock. Document entry and exit endpoint conventions; for counting-process records, R survival uses intervals open on the left and closed on the right. Do not count unobserved pre-entry time as exposure. [survfit-options]\n\nRisk-set membership is specific to the event definition and model. Repeated observations from one drive are not new independent drives. A correct denominator alone does not establish non-informative censoring or eliminate dependence between devices. [survival-methods]",
  "related": [
    "greenwood-variance",
    "kaplan-meier-estimator",
    "log-rank-test"
  ],
  "citations": [
    {
      "id": "km-1958",
      "title": "Kaplan and Meier (1958): Nonparametric Estimation from Incomplete Observations, sections 1.4 and 2.2, equation (2f)",
      "url": "https://web.stanford.edu/~lutian/coursepdf/KMpaper.pdf",
      "accessed": "2026-11-19"
    },
    {
      "id": "survfit-options",
      "title": "R survival: Surv, counting-process entry and exit endpoints",
      "url": "https://stat.ethz.ch/R-manual/R-devel/library/survival/html/Surv.html",
      "accessed": "2026-11-19"
    },
    {
      "id": "survival-methods",
      "title": "R survival package: Methods, sections 2.2.1-2.2.2, confidence intervals and robust variance",
      "url": "https://stat.ethz.ch/R-manual/R-patched/library/survival/doc/methods.pdf",
      "accessed": "2026-11-19"
    }
  ],
  "status": "shipping",
  "advantages": [],
  "limitations": [],
  "supersedes": [],
  "supersededBy": [],
  "firstSeen": null,
  "lastSeen": null,
  "updated": "2026-11-19"
}
