Friday, April 19, 2013

Career A18-49+: Evaluating Full Series Runs

Since this has been a very slow week in entertainment TV land, I've been doing some old data housekeeping on War of 18-49 posts. I've created new SpotVault pages for a select handful of ended shows (see the "Results" column in the first table on each of those pages), and all the "ended" pages in the War of 18-49 have now been updated with the current table format. And there's been one other addition, as one intrepid commenter has already noticed: the "CAREER" number newly included at the bottom of the A18-49+ column on the ended pages. What does this number mean???

It's my stab at a stat that can evaluate a full series run, taking into account both the show's ratings production and its longevity. While most of my A18-49+ stuff so far has focused on putting individual seasons on a level playing field, the number should also be able to, say, compare the ratings value of a full show airing from 2003-2007 with that of a show airing from 2009-2013. In other words, it should be able to combine seasons together. And that's what "Career A18-49+" does, in pretty much the simplest way possible: A show's "Career A18-49+" = the sum of each individual season's A18-49+.

As with the original season-long A18-49+, the next step is to come up with a frame of reference. And we can't really do that without giving up some of the goods. Here are all of the scripted shows with Career A18-49+ of over 1,000 in the last decade:

Show Career A18-49+ Status
Grey's Anatomy 1717 ongoing
CSI 1685 ongoing
Two and a Half Men 1604 ongoing
Desperate Housewives 1499 ended
House 1332 ended
CSI: Miami 1245 ended
The Simpsons 1196 ongoing
NCIS 1187 ongoing
Law and Order: SVU 1140 ongoing
Family Guy 1136 ongoing
The Office 1069 ongoing
The Big Bang Theory 1035 ongoing
How I Met Your Mother 1002 ongoing

I want to make labels that provide context for this along the lines of the "megahit / big hit / hit / solid / marginal / flop" ones used in regular A18-49+. Where should the cutoff for the top tier (which I'm tentatively calling "icon") be? My thought is somewhere in the 1300 to 1500 range. This is probably not very high-integrity, but I would hate to make it 1500 if only because that would mean leaving Desperate Housewives out by a single point. Desperate Housewives feels pretty iconic to me; eight seasons isn't a stellar total length, but four of them were megahits and another two big hits. House is more of a borderline case with just two megahit seasons (neither of which were as big as the first two Housewives seasons). I'm inclined to lean on the side of making this a truly elite club and putting "icon" in the vicinity of 1400, if not higher.

But I'm going to leave the tier-making for a separate post. Right now I'm just going to lay out some concerns about this number in general. Perhaps someone will have an idea about how to improve upon some of these problems.

1. Missing info. This number is most useful for shows that premiered in 2003-04 or later, since there are no A18-49+ numbers from prior to that season. But the good news is that many of the older War of 18-49 shows can assuredly be put in whatever number is decided on as a "top tier" (like ER, The Simpsons, CSI, Law & Order, SVU, American Idol and Survivor), and there are only a few others that aren't quite as abundantly obvious.

2. Multiple-season, multiple night shows. There's the question of how to come up with one number for a show like Survivor, which airs multiple seasons within one year, or a show like American Idol, which airs on multiple different nights, or a show like Dancing with the Stars, which does both. My first instinct is just to average all the seasonal/nightly A18-49+ numbers from within one TV season.

3. Season lengths. Though it'd probably be a lot of extra work, it might be worthwhile to base this stat on the number of hours aired rather than the number of seasons. But this is perhaps more complicated than it might sound, because it might give a large edge to bloated reality shows that they don't particularly deserve because they never air in repeats.

4. Is longevity rewarded too much? In the baseball world, sabermatricians recoil at the notion of using some arbitrary number (like 3,000 hits or 300 wins) to decide whether a player belongs in the Hall of Fame. I don't mind the general idea, because I think a number like that does a decent job of combining longevity and production. But does it favor longevity too much? Generally speaking, getting to these thresholds is much more about playing for a long time than playing at an extraordinary rate. It's not like you can get there in five seasons.

To use a practical example, Cold Case, with a 619 career across seven seasons, was a more valuable show than Heroes, with 528 across four. If Fox were to cancel Glee after four seasons (which probably won't happen), it'd have a smaller number (551) than Cold Case as well. It doesn't seem "right" because we can all remember times when Heroes and Glee were tearing it up in the ratings, and Cold Case never tore it up. (In fact, it never even hit the league average in seven seasons.) We tend to think that "greatness" should be about great production, not a long period of good/decent production, because great production is more memorable.

So I guess the question on this front is: do we want a stat that tells us something we don't know, or do we want one that simply puts numbers to things we already know?

Anyway, this is all just a brief, rambling intro. Nothing is particularly set in stone yet, but I wanted to get something out there for those who might have stumbled upon this number on a War of 18-49 page.


Spot said...

This. Is. Amazing.

Spot said...

Perhaps a separate Career Average of the 18-49+ number for each episode, so that Cold Case situations would have context.

Spot said...

"And there's been one other addition, as one intrepid commenter has already noticed"


Spot said...

This is great! If I may suggest just one small thing, is that maybe it worth be worth it to have each season weighted by the number of episodes. Like for instance, Scandal's season 1 average is lower to what is currently scoring in Season 2, but season 1 also has a whole lot less of episodes, so it should count less. In most cases it won't matter of course, but in a few cases maybe it will! Just a suggestion, feel free to ignore it! I will follow this regardless!

Spot said...

In sports, while the flashy new thing is great and gets all the headlines, isn't it always the lower-tier but steady utility player who ends up being more valuable to the team if that flashy new guy doesn't pan out long-term? Going off of that, I think Cold Case having a higher Career 18-49+ than Heroes is actually a fairer representation of the legacies of the two shows. After all, when people think back on Heroes now, all they seem to see is a smoking head wound that basically knocked NBC down for the count when it fell apart.

Spot said...

On the other hand, Rock Center could be cheap filler for another 14 years.

Spot said...

Yes, I meant to make a note of that wrinkle in that section. Though I tend to agree with James' general principle, all of the 20/20's, Datelines, America's Funniests and Cops of the world would probably get into the top tier or very close if we could examine their full runs. But perhaps it's better to just exclude that whole category from being a part of this.

Spot said...

Thanks. This is basically what I'm getting at in #3... I'm just afraid of there being an "unscripted bias" toward shows that air a ton of originals and no repeats. If I get into weighing seasons based on runtime, which I agree seems like a good idea on the surface, I think the repeat value has to be incorporated too somehow.

Spot said...

Yeah, I wasn't really thinking about the newsmags and other such "Sundays @ 7" stuff. Given that, 20 years of Dateline at circa 95 would be around 1900 for the career, which is higher than anything listed in the article.

Spot said...

Maybe instead of using career totals you should use a show's career average. You wouldn't need to make new labels and shows that are valuable than others but has less seasons and career A18-49 points are judged accordingly.

Spot said...

I have an idea. Maybe, let's say, a show lasted for five seasons. The A18-49+ labels could be adjusted for each five season show's Career Total. For example:
0-349 is a flop for a five season show.
350-499 is marginal.
500-624 is solid.
625-749 is a hit.
750-999 is a big hit.
1000+ is a megahit.

The labels would be adjusted according to how many seasons the show had.

Spot said...

Isn't that basically what you just suggested, though? :-P It's just grading based on the average over the five-year period.

An idea I had was to keep the current equation (average * number of seasons), except put a cap on the "number of seasons" part. Maybe 10 or so. That would keep the low-priority stuff from getting into the top tiers.

Post a Comment

© 2009-2022. All Rights Reserved.