Graph TV – The Simpsons | Hacker News Reader

Graph TV – The Simpsons

graphtv.kevinformatics.com

70 pointsColinWright54 comments

This graph of The Simpsons illustrates (for me at least) that IMDB is more reflective of the viewing community's weird quirks and biases.

Honestly: is "Homer the Smithers" the best episode of The Simpsons? I doubt anyone would truly put it on their "best ever" list. It's really more "solid" than great. Most importantly: it doesn't rub anyone the wrong way.

Is "Saddlesore Galactica" one of the worst episodes ever? No. It's extremely funny and the story is structured well. It gets lots of very low votes from viewers who favor realism over humor (the episode is implausible with the horse racing then gets silly/fantasy). The episode's score reflects a community desire rather than an objective opinion.

> The episode's score reflects a community desire rather than an objective opinion.

I don't really understand your point. You think it's biased for audience members to rate an episode poorly because they didn't enjoy it and didn't think it was funny?

People that disliked something are more likely to vote on that episode to voice their dislike.

For example if the episode makes fun of christians you may have a heavy amount of christians vote the episode down even though it was a good episode. The people that thought it was good aren't as motivated to vote as the people offended.

This is also common with service providers and internet reviews. If the service is good there is no need to go online and post a review. If it's bad you are much more likely to take the time.

"objective opinion" ?

>Is "Saddlesore Galactica" one of the worst episodes ever?

No, and the graph in no way show that. From around season nine, episodes are mostly rated between 6.5-7.3, with a fair amount lying considerably lower than that and only a few outliers in the other direction.

But as a 6.5'er, "Saddlesore Galactica" is a solid "poor" episode, like many, many others. But there are a substantial amount of episodes well below that one, and they are also scoring considerably lower.

The clip shows are the truly "worst", despite having all the best punchlines. So one could argue that objectively they are the best. But that is also a poor argument.

This is a great example of how to lie with statistics, or what someone can claim by applying just college-level statistics. Those trend lines are completely worthless.

See the trend lines for Futurama: http://graphtv.kevinformatics.com/tt0149460 The difference between all the episodes make having a "trend" very doubtful. Especially season 5 is a wildly varying season where if you take one episode away the line would completely flip.

Much the same can be said for The Next Generation http://graphtv.kevinformatics.com/tt0092455 where you have basically clouds where seemingly at random a line is drawn through it.

Yes, I know there are statistical methods for determining trends, but without data on their accuracy they are pretty much worthless. And you really should use a threshold for those accuracies if you're presenting this kind of data to a very wide audience.

Forfty percent of all people know that.

https://www.youtube.com/watch?v=1j2Duy_xzEA

Love this. You can really see how some series took a while to get their feet under them.

Star Trek: TNG http://graphtv.kevinformatics.com/tt0092455

Deep Space 9: http://graphtv.kevinformatics.com/tt0106145

Enterprise: http://graphtv.kevinformatics.com/tt0244365

Or in each season

Game of Thrones: http://graphtv.kevinformatics.com/tt0944947

Or took a while to hit their stride, then left right after the peak Seinfeld: http://graphtv.kevinformatics.com/tt0098904

Had a story arc: B5 http://graphtv.kevinformatics.com/tt0105946

Or were clearly failing:

Andromeda: http://graphtv.kevinformatics.com/tt0213327

Sliders: http://graphtv.kevinformatics.com/tt0112167

Futureama: http://graphtv.kevinformatics.com/tt0149460

Or are experiencing a revival

http://graphtv.kevinformatics.com/tt0121955

I wish this could be correlated against ratings data.

And or course: Dr. Who http://graphtv.kevinformatics.com/tt0056751

edit:

Also Law & Order http://graphtv.kevinformatics.com/tt0098844

and Law & Order SVU http://graphtv.kevinformatics.com/tt0203259

and Law & Order CI http://graphtv.kevinformatics.com/tt0275140

and the Stargate series

http://graphtv.kevinformatics.com/tt0118480

http://graphtv.kevinformatics.com/tt0374455

http://graphtv.kevinformatics.com/tt1286039

and Battlestar Galactica

(1978) http://graphtv.kevinformatics.com/tt0076984

recent http://graphtv.kevinformatics.com/tt0407362

caprica http://graphtv.kevinformatics.com/tt0799862

Your Enterprise link goes to TNG again. And somehow I was sure Andromeda kept going for 7 seasons, it was just so bad I wasn't even keeping track anymore.

Fixed, thanks.

Andromeda really only had five seasons. I think it went episodic in Season 3 or 4 after blowing its load on the major story arc (putting together the federation^h^h^h^h^h^h^h^h^h^h commonwealth again). It was never a great show, but fine to put on in the background while I was doing other stuff. But then it got really really bad and I don't even think I ended up finishing it.

The Doctor Who reboot: http://graphtv.kevinformatics.com/tt0436992

Overall kinda consistent, but very scattered if you look at the episodes individually.

For what it's worth, I watched every episode over the course of the last year, so I've got a real good perspective on this.

Here's proof that IMDb is bullshit: Simpson Tide is very highly rated. That episode is one of the absolute worst.

Here's a rough overview:

S1: terrible!

S2: promise!

S3: very good!

S3-S8: the absolute best!

S9-S10: still very good—but not as good

S11: definite decline

S12: yep.

S13-17: ok wow this is pretty bad.

S18-19: a little better

S20-21: definitely better

S22-25: actually pretty good!

The series has been extremely underrated since Season 20 or so when it came out of the slump that began with Season 11. It's not the

This is great!

Here's a few interesting ones I've come across:

* The Wire (http://graphtv.kevinformatics.com/tt0306414)

Known for being a slow starting show, this is visible with the season trend lines.

* The Shield (http://graphtv.kevinformatics.com/tt0286486)

Season 4 is such a massive outlier.

* Seinfeld (http://graphtv.kevinformatics.com/tt0098904)

Held very steady until the last season.

The Wire was the first one I tried, and the most interesting.

And, of course, Breaking Bad: http://graphtv.kevinformatics.com/tt0903747

This is most impressive one I found too

Oh my, I could spend hours with this!

'Sopranos' has the same 'Season 4' problem as 'Shield':

http://graphtv.kevinformatics.com/tt0141842

The massive drop in Twin Peaks after the murderer is revealed:

http://graphtv.kevinformatics.com/tt0098936

The awful last season of 'Dexter':

http://graphtv.kevinformatics.com/tt0773262

And that's how you do a last season (Six feet under):

http://graphtv.kevinformatics.com/tt0248654

I'm amazed Dexter maintained the highs it did. Everything after Season 4 was inferior.

The 'Red Wedding' episode of Game of Thrones was the best single show rating I could find - 9.9! http://graphtv.kevinformatics.com/tt0944947

The 4th episode of True Detective was 9.8. http://graphtv.kevinformatics.com/tt2356777

Check out this outlier on Dexter's last season lol http://graphtv.kevinformatics.com/tt0773262

Breaking Bad episode Ozymandias (s05e14) has a rating of 10!

http://graphtv.kevinformatics.com/tt0903747

Interesting that for each season of Breaking Bd, episode 4 is the worst rated.

It's a great hack that uses lousy data. IMDB ratings are best perhaps with films, a lot less reliable with individual episode scores. Also, if a show goes south do viewers keep watching it and rating it? That phenomena alone sort of spoils the fun for me.

Nostalgia wins, for the Simpsons at least.

I sometimes feel this way about TV in general (early stuff is better).

Example (circa 1967-1968):

http://www.dailymotion.com/video/x12sekq_the-prisoner-arriva...

BTW, I was born in '84 and not 1884. ;-)

I think with regards to the Simpsons it's just that it was better quality pre-2000 or whatever. http://www.imdb.com/title/tt0096697/reviews?ref_=tt_urv you can read through the reviews and see several good points about it.

Jumping the shark seems to be more perception than reality: http://graphtv.kevinformatics.com/tt0070992

FWIW that is in Season 5. Season 6 went even higher!

It's quite weird;used to remember that the recent highlight in a rather dull decade "Steal this Episode" was being rated higher.

11% 1 star ratings with a pretty remarkable demographic distribution. http://www.imdb.com/title/tt2966850/ratings

In general an alternate IMDB algorithm probably would give a truer image if you clip of the extreme ratings 1 and 10 before averaging, thereby getting rid of most fanboy/rage votes.

IMDb does do some type of weighting to the votes, but they don't release their method.

http://www.imdb.com/help/show_leaf?ratingsexplanation

Law and Order was fairly consistently awesome, as I expected: http://graphtv.kevinformatics.com/tt0098844

OP here, I'd really love to include viewership data if anyone knows how to get their hands on some. I was looking at the Nielsen site but saw no easy way of acquiring or requesting data.

Glad to see "Lisa Goes Gaga" as the lowest rated episode. I used to watch the Simpsons religiously but I stopped ~5 years ago. I turned on FOX the other day and happened to catch that episode ... really glad that's the exception and not the rule because it was unwatchable.

This is great stuff! One that I found pretty interesting was House of Cards http://graphtv.kevinformatics.com/tt1856010.

Notice how first and last episode of season 2 are way up there.

The West Wing: http://graphtv.kevinformatics.com/tt0200276

Sorkin co-wrote the first four seasons. I was expecting the dip in S05 but I did not expect such a dramatic upturn in S07.

Here's a great analysis of what made The Simpsons good and why it started to decline after season eight:

http://deadhomersociety.com/zombiesimpsons/

Another interesting one is Two And a Half Men. You can clearly see when things changed. And it's building back: http://graphtv.kevinformatics.com/tt0369179

Sad the "jumping the shark" theory for Happy Days doesn't really hold up against the data: http://graphtv.kevinformatics.com/tt0070992

Scrubs is the one I've found with the biggest outlier season. And for good reason:

http://graphtv.kevinformatics.com/tt0285403

You know, I actually liked the last season. No, it was not as good as the other seasons, but when I add the handicap for it basically being the first season of a new show, it held up ok. I would have kept watching had it been renewed.

But I understand where people are coming from when they compare it poorly to the other ones. The original Scrubs was a growing up story about JD (and comrades to a lesser extent). The last season was more straight sitcom, less story arc. It wasn't attempting to be as meaningful.

I like how this generally confirms my opinion of shows; I can only infer that I hold meta opinions/no opinion of my own. Season 2 finale of The West Wing as some of the best tv ever, though.

That's how you end a series: http://graphtv.kevinformatics.com/tt0165598 huge bump for the series finale

Funny thing: all star treks except TOS end up higher than they started!

I just worked by way through TNG and DS9. TNG took a while to get it's legs...in today's environment it wouldn't have lasted past season 2. But then it became great television for the rest of the run.

DS9 started pretty solid, and became one of the best written pieces of sci-fi TV quickly for a long time.

Really? I started watching the first two seasons, and it's honestly dull.

DS9? Yeah, it starts pretty soap opera-ish. That never really goes away (they're kind of stuck on the station for the most part), but the stakes get higher and higher and the various players and factions involved become pretty interesting. Not quite Game of Thrones, but way out of character for a Trek show.

Pretty predictable plot of Rick and Morty:

http://graphtv.kevinformatics.com/tt2861424

I'm not seeing Something Ricked This Way Comes, which is definitely my all time favourite. It's present on IMDb, so I guess this site caches data.

It's pretty clear that something fishy is going on with the ratings distributions here.

It's pretty clear that Simpsons fans hate Lady Gaga and Musicals.

Really, the Lady Gaga episode was awful. And not only because it was a very obvious 22 minute advertisement. It was just plain bad, very bad. I absolutely agree that it's the worst Simpsons episode I've watched.

The individual ratings may not be very "accurate" (well, they're going to be subjective no matter what. My favourite episode is near the bottom of its season), but the trends are interesting. The fact that the floor of seasons 1-9 is above the ceiling of season 14 and onwards (save a few outliers) reflects quite well what many fans think.

The pattern seems to be an effective floor of 7.4 until midway through Season 8, a floor of 7.3 until Season 11, and a ceiling of 7.3 from then on.

I can't think of why that would be the case, or how it makes sense.

I agree, no other TV show on this site has this kind of pattern. And the oddest thing is the ratings of season 1. Because, really it was awful.

The distribution is a result of the low number of ratings. Look at the number of ratings for each episode. For all those early 7.3-7.4 episodes it's about 600 ratings, the higher rated episodes have more ratings. Then from about season 12 there's a drop in number of ratings to about 300-400 for most episodes.

Compare this to the Game of Thrones ratings[1], where each episode has about 3000 ratings.

[1] http://graphtv.kevinformatics.com/tt0944947

I really enjoyed playing around with this!

no sqrt(n) error bars?!?

Reply on news.ycombinator.com