We just finished scoring 30 Rock. 138 episodes, seven seasons, 9,283 jokes scored using the same v2 methodology that's been applied to the other six shows (9 craft dimensions, 3-run consensus, Bayesian shrinkage on per-character ratings, the format coefficient deprecated as of April).
The result: 30 Rock is now #1 on the Humor Index at 84.3, with a 2.3 point gap to Arrested Development at #2. That's the biggest #1-to-#2 gap we've ever measured. Bigger than the gap from AD to Parks. Bigger than the gap from Parks to The Office.
Here's the leaderboard now:
| # | Show | Humor Index | Episodes | |---:|---|---:|---:| | 1 | 30 Rock | 84.3 | 138 | | 2 | Arrested Development | 82.0 | 84 | | 3 | Parks and Recreation | 78.8 | 124 | | 4 | The Office | 78.6 | 186 | | 5 | Seinfeld | 77.8 | 172 | | 6 | Friends | 77.5 | 235 | | 7 | Schitt's Creek | 77.3 | 80 |
(All shows shifted slightly downward on May 15 when we recalibrated the display scale so the all-time top episode caps at 100 instead of clamping. Rank order is unchanged. More on the recalibration here.)
Before getting into what the data says, the usual caveat: items 3 through 7 are statistically a wash. The 95% confidence intervals overlap heavily. The order of "is The Office actually funnier than Seinfeld" cannot be answered with this data — those numbers are within scoring noise. The two real claims this leaderboard supports are: 30 Rock is meaningfully above AD, and AD is meaningfully above everyone else. Everything below #2 is an elite cluster.
What 30 Rock is doing differently
The story of why 30 Rock won is a story about one statistic: 62.7% of every joke in 30 Rock is character comedy. That's the highest percentage of any show in the dataset. The next-closest is Arrested Development at 54.4%. After that, the rest of the field drops into the 30s and 40s.
What "character comedy" means in our taxonomy: a joke whose punchline depends on you knowing who the character is. Jokes that don't translate without that. A line that works because Jack Donaghy said it, and would land flat from anyone else.
This is the structural opposite of, say, Seinfeld, which leans much more heavily on situation + observational humor (jokes that would work delivered by a stranger). It's also why 30 Rock's repeat-watch value is so high — once you know the characters, you can keep finding new layers of the same lines, where for situation-driven comedies the joke is essentially fixed in its first viewing.
Two other 30 Rock distinctives in the data:
JPM of 2.86 — top of the dataset alongside AD. 30 Rock's per-episode joke density is roughly 50% higher than Seinfeld's and double Friends'. Every minute of 30 Rock contains more comedic content than every minute of any single-camera show we've measured.
Absurdist at 24.5% — also the highest in the dataset. The cutaways. The fantasy sequences. The Tracy Jordan everything. About one in four jokes on 30 Rock is structurally non-realistic.
Put those two together — character-driven AND high JPM AND high-absurdist — and you have a comedy formula nobody else in the dataset has executed. The closest analog is Arrested Development, which is character-driven and absurdist but has lower JPM. 30 Rock is denser.
The top 10 episodes
| Rank | Title | Score | Season | |---:|---|---:|---:| | 1 | Game Over | 94.8 | S7 | | 2 | Cooter | 94.1 | S2 | | 3 | Verna | 93.8 | S4 | | 4 | Nothing Left to Lose | 93.0 | S6 | | 5 | Alexis Goodlooking and the Case of the Missing Whisky | 92.8 | S6 | | 6 | Reaganing | 92.4 | S5 | | 7 | The Rural Juror | 92.3 | S1 | | 8 | Mrs. Donaghy | 91.6 | S5 | | 9 | ¡Qué Sorpresa! | 91.5 | S5 | | 10 | Jack Gets in the Game | 91.3 | S2 |
A note on the top of the scale: the display formula was recalibrated on May 15, 2026 so that the all-time top episode in our dataset — Office Dinner Party at raw 8.34 — anchors near 100. Under the previous calibration, 30 Rock had seven episodes scoring above 100 on the old display scale (along with one Office and one Parks episode), which made the scale lose its meaning. The recalibration preserved all rank order and just compressed the slope. (Full writeup of the recalibration.)
The peak isn't concentrated in one era either:
- Season 4 is the peak season by aggregate (89.9 display), but only by a hair over Season 6 (89.5)
- Seasons 2, 5, and 7 are all 87+
- Even Season 1, the weakest, is at 85.5 — which would rank above the show-level average for The Office, Seinfeld, Friends, and Schitt's Creek
The variance across seasons is the lowest in the dataset. 30 Rock didn't have a "the show found itself in season 2" arc the way most sitcoms do. It came out swinging and stayed there.
The character data
Top per-joke craft for 30 Rock characters, minimum 50 jokes:
| # | Character | Per-joke craft | Total jokes | |---:|---|---:|---:| | 1 | Dr. Spaceman | 7.68 | 68 | | 2 | Kenneth | 7.49 | 678 | | 3 | Jack Donaghy | 7.39 | 1,913 | | 4 | Tracy Jordan | 7.36 | 1,221 | | 5 | Jenna Maroney | 7.31 | 819 | | 6 | Pete Hornberger | 7.22 | 172 | | 7 | Liz Lemon | 7.20 | 2,024 |
Two genuinely surprising things here.
Kenneth's per-joke craft is higher than Jack's. If you'd asked any 30 Rock fan to rank the craft of the cast, almost nobody would put Kenneth in front of Jack. But the data is unambiguous: every line written for Kenneth does more structural work per beat — the character's worldview is so specifically constructed that essentially every line reinforces it. Jack's lines are individually elite but more variable; Kenneth's lines are uniformly elite within his narrower band.
Liz Lemon has the LOWEST per-joke craft of the main cast — and the most jokes by a wide margin (2,024). She's a comedy workhorse: she takes the most volume, and the rubric reads the average of that volume as lower-craft because she's the audience surrogate. Her job is to react. Reactions, by the rubric's design, score lower on character integration than self-driven punchlines from characters with strong points of view. This is by design — we don't want the reaction-shot framework to drive the leaderboard. If we did, Liz would top this list and everyone we just listed above her would drop.
Worth restating: this is per-joke quality. By total comedic output (Liz's 2,024 jokes at 7.20 craft + her impact + her quotability), she'd run away with the WAR leaderboard against most of her cast. The number above isn't measuring "who's the funniest" — it's measuring "who does the most work per beat." Different questions, different answers.
How this changes the cross-show character leaderboard
We're going to update the cross-show WAR leaderboard next week with 30 Rock characters added. Expected impact:
- Jack Donaghy is going to land in the top 5 cross-show by WAR. 1,913 jokes at 7.39 craft is elite roster-filling.
- Liz Lemon will probably be top 10 by total WAR despite the per-joke craft number — volume + ubiquity matter.
- Tracy Jordan will likely beat several main characters from other shows on Impact specifically — his per-joke Impact is 7.3+, which is at the elite tier.
- Kenneth will be a craft-leaderboard outlier — he'll rank surprisingly high given how few people think of him as a "top tier" character.
The methodology caveat we always include
The single-run scorer noise floor is ~5 points per episode (ICC = 0.28). Three-run consensus tightens this dramatically — show-level CIs at the aggregate are <2 points wide. But this matters for how you should read the leaderboard:
- "30 Rock is meaningfully above AD" — supportable; 2.3 point gap with 95% CI separation
- "AD is meaningfully above Parks/Office/etc" — supportable; ~5 point gap to the next cluster
- "Within the elite cluster, Parks (80.55) is funnier than Office (80.22)" — NOT supportable; that's a 0.33 point difference inside the noise floor
This is the part we keep saying because rankings get screencapped and become claims. The right way to read this leaderboard:
That's the honest version.
What's next
30 Rock makes seven scored shows. The next four on the backlog: Brooklyn Nine-Nine, It's Always Sunny, The Big Bang Theory, Two and a Half Men. The drip-release plan from May has B99 next at the front of the queue.
Two analyses falling out of this dataset are also coming:
- An updated cross-show WAR leaderboard with 30 Rock added — likely shifts the top 5 significantly
- The "characters per show on the craft top-20" stat — 30 Rock will have several names; how that shapes up cross-show is the most interesting question this data raises
If you want the underlying numbers, every per-episode score and per-joke detail is at thehumorindex.com/shows/30-rock.
---
30 Rock is the seventh show fully scored on the Humor Index. The methodology is documented at [our methodology page](/methodology), and the show-level confidence intervals are explained at [our Bayesian credible intervals post](/blog/bayesian-credible-intervals).