What if you picked the cast?

Update, October 15: the five-player lineup and five bonus questions have been adopted for Week 4. The unsubmitted October 31 card has been replaced. The original proposal and research below remain here for reference; the launch post explains the current rules.

• Read the format-change announcement

Fantasy sports usually begins with a lineup. You choose your team, watch what the players do, and add up the points. We’re exploring whether Fantasy SNL should move more of its weekly game in that direction. Fewer questions that require keeping up with the news; more reasons to watch the cast work.

Here’s the version we’re testing: pick five cast members. Put three in Repertory slots and two in Featured slots. Repertory earns 2 points per appearance; Featured earns 4. Pick five. Feature two. Every qualifying appearance scores.

Those are your fantasy positions. They have nothing to do with SNL’s cast billing. You could put an SNL featured player in a Repertory slot and Mikey Day in a Featured slot. A small part and a lead part would earn the same appearance points.

This was published as a proposal on October 12. On October 15, we adopted the five-plus-five format for the unsubmitted Week 4 card. Earlier scores and preseason predictions keep their original rules. The discussion and historical tests below show how we arrived at the new format.

The proposed weekly card: five players, five bonus questions

The lineup would do most of the scoring. Choose five different current cast members, then feature two of them. There is no draft: different players can choose the same cast members. Your three Repertory picks would earn 2 points for each qualifying appearance, and your two Featured picks would earn 4. Lead, supporting, and small parts would all earn the same appearance points.

Your saved lineup would carry forward from week to week unless you changed it before the cutoff. You could keep all five, replace a cast member, or change which two you feature. The five weekly bonus questions would be separate choices for that episode; their answers would not roll over.

That is 15 possible bonus points. The cameo and musical-guest questions would no longer ask where the appearance happens. A celebrity cameo would still exclude the current cast, host, and billed musical guest. Update would count separately introduced guest segments, so a jointly introduced pair would count as one. These are questions about the episode you’re watching, without a news-reading assignment or an artist’s catalog to study first.

What would leave the weekly card?

• Wildcard Cast Member would become the five-person lineup.
• Cold Open Lead and the first-sketch-after-monologue, first-sketch-after-Update, and final/10-to-1 calls would leave. The lineup would reward cast appearances without asking players to predict a sketch’s form and most prominent performer several times.
• First Impression, First Recurring Sketch or Character, and Update Opening Story would leave. They demand more outside knowledge and more fine distinctions when we score them.
• The routine first-to-break question would leave. Special-event questions could return occasionally, with rules announced before a card opens.
• Musical Guest Songs would leave the scored weekly card. Music fans could still trade song predictions in the blog comments.
• Host Workload would leave. A shared host slot is a separate possibility discussed below; it is not part of the five-player, five-question proposal yet.

Your comments come before a rules change

Our original plan was to consider a change after October 31. No player comments or Week 4 submissions had arrived when we adopted the format on October 15, so we replaced that unsubmitted card ahead of the October 31 episode. Earlier cards, scores and preseason commitments are preserved. See the launch post and current standards for the adopted rules.

Tell us which parts you would keep, which you would cut, and what would make this easier or more fun to play. The public discussion is below the research. Everyone can read; a free account lets you comment. Private feedback by email is welcome too.

Does the same cast always win?

We tested all 20 regular episodes of Season 51 using its actual 17-person roster. The best lineup kept unchanged for the entire season was Ashley Padilla and Sarah Sherman in Featured, with Mikey Day, Chloe Fineman, and James Austin Johnson in Repertory. It earned 1,174 points.

That sounds like a settled starting five. But those exact five were never an optimal five-person cast for any individual episode, even allowing ties. Fifteen cast members could appear in at least one winning weekly lineup. The season’s leading five accounted for about 52% of the optimal weekly roster slots when tied performers share the available spots equally. There was considerable room for other performers.

What does changing a lineup buy you?

Perfect weekly choices add 276 points, or 23.5%, over the best unchanged lineup. But nobody gets to watch an episode before submitting a card. These are ceilings, not a promise that making changes will earn that much.

For a more practical comparison, we used the first five episodes to select a lineup, then scored only Episodes 6–20. Leaving that lineup alone earned 808. Picking each week from the previous five episodes’ appearance averages earned 838: 30 more points, or 3.7%. Using the previous three episodes earned 830; using the entire earlier season earned 822. Simply copying the previous episode’s leaders earned 810.

These are simple historical tests, not a proven prediction model. Choices used only information from earlier episodes, with alphabetical tie-breaks. Starting after three or eight episodes changed the results: the five-episode strategy’s advantage ranged from 0.6% to 6.4%. Weekly attention helped in those tests, but chasing last week’s big performance was much less useful than it might sound.

Two points and four points?

The proposal above uses 2 points in Repertory and 4 in Featured, but we’re inviting feedback on the spread. For cast points alone, 1 and 2 produces exactly the same lineup rankings as 2 and 4; it halves every cast score. Once bonus questions are included, that change also affects how much the lineup matters relative to the bonuses. We would review those values together. A 2-and-3 spread gives the Featured choices a smaller premium. Changing the spread alone won’t make different performers lead the appearance totals.

What about a host slot?

We also checked the host credits. Last season’s hosts appeared in 6–10 qualifying sketches per episode, averaging 7.25. At 1 point per host sketch, that would add 145 points over the season; at 2 points, 290. For this test, host sketches include cold opens, live sketches, and comedy pretapes, with the monologue, Update, music, introductions, and goodnights excluded.

An automatic host slot would give everyone a reliable baseline and another reason to follow the whole episode. It would also award everyone the same points each week, so it wouldn’t change the standings or reward better lineup choices. That distinction matters when deciding what we want the slot to do.

What we counted—and what still needs review

The cast test counts one credited appearance per performer per sketch or monologue, plus one Update credit for anchoring or a guest appearance. Opening credits, musical performances, and goodnights are excluded. The historical archive groups Update guest features under one segment; no repeated independent guest features by the same cast performer were identified in this season’s records. A future game needs a clear rule for that situation. Voice credits are included where the source credits the performer.

We reconciled all 1,072 cast credits across 198 performance segments against the current public SNL Archives episode pages. That verifies agreement between the records, not an independent replay audit. Host counts were extracted separately from those pages’ host-credit sections. Bowen Yang was eligible for the weekly-choice tests only through his final cast episode, Episode 9. The archive remains under editorial review, and corrections may change these calculations.

• Download every cast member’s weekly appearances (CSV)
• Download weekly lineups, scores, and host counts (CSV)
• Read the complete calculations and methodology (JSON)
• Browse Season 51 in our archive
• Source episode index: SNL Archives