59 of 174, better than a third, appeared in one season and never came back.
The most common thing that happens to a cast member
First, what the count is. The number of seasons attached to each person here is a sourced figure, not arithmetic on the years they arrived and left; those two disagree for most of the roster, because people are hired in November and dropped in March. So "one season" means one season credited, and it covers a full year and a handful of shows alike.
On that count, 59 of the 174 people ever cast lasted exactly one season, which is 33.9% and which the statistics page already reports. What is worth adding is that this is not merely a large group, it is the largest. The tenures run 59 people at one season and 20 at two before the numbers flatten out, so a single season is the modal outcome and is almost three times as common as the next one along. The median run is three seasons, which is the figure usually quoted and the fairer summary of a typical career; but no other single length comes close to happening as often as one.
The group has a shape. 35 of the 59 never reached the repertory company. Turned round, that is starker: of the 49 people whose only credit is featured player, 34 lasted a single season, better than two in three. The apprenticeship the history of the show describes as two years of small parts before promotion or exit is, in the data, mostly a one-year trial that most people do not pass.
What does not mark the group out is a writing credit. 17 of the 59 are credited as writers, 28.8%, against 35.7% of everyone else, and a sample of 59 cannot tell that difference from zero (z = -0.91). Whatever a writing credit predicts on this show, and it predicts the Weekend Update desk very strongly, it does not predict surviving your first year.
Nor is one season a synonym for failure. 24 of the 59 reached the repertory company, several of them immediately, and seven were later invited back to host the programme that had let them go: Billy Crystal, Robert Downey Jr., Paul Shaffer, Martin Short, Sarah Silverman, Ben Stiller and Damon Wayans. Seven of 59 is 11.9% against 27.8% of the 115 people who lasted longer, so a short run does cut your odds of the invitation to well under half. It does not remove them.
Twenty-nine of them arrived in one decade
Sort the 59 by the decade they joined and they are not spread evenly: 4 from the 1970s, 29 from the 1980s, 9 from the 1990s, 4 from the 2000s, 7 from the 2010s and 6 from the 2020s. Raw counts are the wrong figure, though, because the show has never hired at a constant rate: 50 new people in the 1980s against 22 in the 2000s. The comparable figure is the share of each intake that never came back for a second year.
That share is 26.7% for the 1970s (4 of 15), 58.0% for the 1980s (29 of 50), 24.3% for the 1990s (9 of 37), 18.2% for the 2000s (4 of 22), 24.1% for the 2010s (7 of 29) and 28.6% for the 2020s (6 of 21). Set the 2020s aside for a moment, because that decade is not over and two of its six are Saidah Belo-Osagie and Grace Reiter, hired for season 52, who have not finished a first season yet; the second of the two objections below takes it apart properly. Among the five completed decades, the 1990s and the 2010s land within a fifth of a point of each other, the 1970s and the 2000s sit a few points either side of them, and the 1980s sits more than thirty points above every one of them. Tested against every other decade combined, 29 of 50 against 30 of 124, that gap is about as solid as anything this dataset produces (z = 4.26, p below 0.0001). It is also the conservative version of the test: the comparison group still contains those two unfinished runs, and stripping out everyone whose time in the cast is not over drops the other decades to 28 of 107, which widens the gap rather than closing it.
The individual cells are a different matter and should not be read as findings. The 1970s figure rests on 15 people and the 2000s figure on 22; the eight and a half points between them is four people either way and means nothing at all. The claim worth making is the 1980s against the other 124, and that one holds.
Two ways to make that go away
The first is that the 1980s number might be a single catastrophe wearing a decade's clothing. Season six is the obvious candidate: 12 of the 15 people hired in 1980 lasted one season, the largest intake in the file. That needs separating from the other fifteen of 1980, because the two are easily run together. Fifteen stints also end in 1980, but only five belong to people who arrived that year; ten joined between 1975 and 1979, and are the original company and the 1979 intake leaving after season five rather than new hires washing out. The biggest arrival and the biggest departure in the file share a year and are largely different people. If 1980 alone is doing the work, the finding is about one management change at NBC rather than about a decade.
So take it out. Excluding everyone whose first season was 1980, the 1980s still lost 17 of 35 first-year hires, 48.6%, and the gap against the other decades survives (z = 2.79, p = 0.005). Year by year the losses run 12 of 15 in 1980, 3 of 7 in 1981, then nothing much, then 5 of 5 in 1984 and 8 of 11 in 1985. That is three separate collapses under three different regimes, which is exactly what the history page describes: the Doumanian season, Dick Ebersol's cast of established names in 1984, and the 1985 company that was almost entirely replaced after a year. Four of the six largest single-year exits in the whole file fall in 1980, 1981, 1985 and 1986, whoever it was that was leaving.
The second objection is that the 2020s figure of 28.6% is not a finished number either, and the reason is a counting artifact this site cannot avoid. Two of those six joined in 2026 and are still in the cast. They are recorded at one season because that is the season they were hired for, not because anything ended; neither has finished it, and if either of them returns for a second season they leave this group. Restricted to people who have actually finished, the 2020s intake is 4 of 8, a sample that can support no claim whatsoever.
Widening it gives a steadier answer. Of everyone who joined in 2000 or later and has since left, 15 of 55 lasted one season: 27.3%, sitting right on the 1990s and 2010s figures. The six currently counted in the 2020s can only shrink as those runs continue, and the decade's real figure will not be known until it takes in people who have not been hired yet. What it will be measured against is the 18% to 25% the show has posted in every completed decade since 1990.
There is a third thing the data cannot do, and it limits all of the above. The roster records that a run ended, never why. Someone who left after a year for a film, someone cut in April, and someone never really hired for longer look identical here, and among 59 people there is certainly a mixture. So "one season" describes an outcome and not a verdict, and the biographies are where the individual cases stop being a rate.
What the rates do support is a correction. The churn itself is real: better than a third of everyone ever hired got one year, and no reading of the 1980s makes that go away. But the version of the story in which the programme is a revolving door, in which a new cast member's odds are roughly a coin toss, is drawn almost entirely from the first six years of the 1980s. In each completed decade since 1990 the figure has sat between 18% and 25% of new hires. The one-season cast member is a real and permanent feature of the show at that rate. The rate people remember belongs to a particular stretch of the 1980s rather than to the institution.
Cast dates and tenures on this site come from Wikipedia; the counting and the reading of it above are our own. More of these are on the articles page, and corrections are welcome at info@snlroster.com.