The Process

How I ranked every episode of The Office by how funny it is:

Practically no one who quickly scans my list to find how I've egregiously under-appreciated their favorite episode will read this next sentence, but it's possibly the most important piece of information to have: One hilarious moment does not make an entire episode deserving of a top ranking.

A couple other lists like this exist elsewhere on the Internet, but none that I've seen go beyond just "best" episodes, and certainly none attempt to scientifically quantify their nebulously stated rankings. What does "best" mean? How do you compare an episode that was more about character/relationship development with one that was all about the laughs? Did the people assembling these rankings just look at a list of all episodes and rank them by memory?

That's how it seemed, and memory can be a tricky thing. There are many episodes that rank in the bottom half of this list that have some incredible individual punch lines, but this is about full episodes, not just one line from an otherwise (comparatively) okay episode.

So I watched every episode while taking detailed notes, marking every instance of humor and categorizing how funny that instance of humor is to create a "laugh score". After I began compiling this data, I presented my findings to a math PhD, who helped shape the formula by which the episodes would be ranked.

The episodes are divided up as they are shown on Netflix, which is why some two-parters like "Search Committee" are split up, while others like "Launch Party" are combined into one. Due to the sheer volume of jokes in the double episodes, these installments would be disproportionalitely boosted in the rankings. Because of this, the laugh score also takes into account episode length, giving us what basically amounts to a humor-per-minute indicator.

To test out this formula, I also watched a few random episodes of some other highly popular sitcoms. Most of them scored in the middle to lower half of The Office episodes, which suggests two important takeaways. First of all, the formula puts The Office episodes into an encouraging normal distribution bell curve, which was not broken by these other sitcom episodes. Secondly, The Office is generally funnier than most sitcoms.

Without revealing the whole proprietary formula, that process looks a little something like this:

• Every instance of humor is ranked as a 1 (mildly funny), 2 (moderately funny), or 3 (extremely funny).
• This data is formulaically combined, taking episode length into account to obtain a "laugh score".
• Immediately after watching each individual episode, the episode was given a "feeling score" based on how funny overall I felt it was.
• Each individual episode's IMDB rating was taken into account.
• "Laugh score" accounted for 50% of the final "episode score".
• "Feeling score" accounted for 30% of the final "episode score".
• IMDB rating accounted for 20% of the final "episode score".
• Ties were broken by which episode had the most "extremely funny" instances of humor.

Bottom line: The "episode score" takes into account every instance of humor in an episode, how funny those instances are, how long the episode is, the quality based on IMDB's crowdsourced rating, and how funny I generally felt the episode was immediately after watching it.