Wednesday, May 23, 2018

Renewology: A Look Back at the Second Season

My recap of the Renewology maiden voyage in 2016-17 was actually published in late September, but I'm doing this one four months earlier because there is a lot of stuff fresh on my mind! I'm also feeling better about how things went for the model this year, which may be another motivating factor...

As you might have noticed during the renew/cancel rush, this year full of network decisions that went against the "conventional wisdom" about how these decisions are made. It was not a great year for the people that tell you that a bunch of shows have their fates locked in because of non-ratings factors. A lot of those so-called locks didn't get there this year, and you should take notice of that. For me anyway, being extremely confident and missing on multiple occasions is a huge deal. The conventional wisdom certainly had some "wins" compared with Renewology, with shows like Blindspot and Madam Secretary and Elementary. But if you're going to be super-confident about the shows that you think will be renewed regardless of ratings, you better have a very good track record. And supposedly back end-motivated shows like Lucifer, Quantico, even stuff like Designated Survivor and Scorpion that R% had in "only" light green territory... those aren't wins.

For Renewology, a model that is based only on ratings and how many shows get renewed in a very broad sense, the biggest miss was Kevin Can Wait, which closed at 81%. If not for the Chris Rock/Adam Sandler bump on finale night, it might have snuck into light green territory, and we would not have had a single miss in dark red or dark green (pending Champions). So none of these decisions (so far) were truly gobsmacking from a pure ratings perspective. I think this year helps illustrate the usefulness of a tool like this. It tells you what the situation is with the hardest data we actually have, while others are ignoring the ratings and making strong promises based on much squishier evidence.

Just to clarify that last point, it's not that I don't "believe" in the non-ratings factors; costs and back-end revenue surely matter a lot. I just don't believe in our ability to know very much about those things before the fact. It's easy to fit explanations to the decisions after the fact, but those explanations have a history of not warranting so much confidence when applied in future situations. Maybe ratings matter less over time, but how much predictive value can really be added beyond what they tell you? That is sort of the fundamental question of the Renewology experiment.

In the end, there were still a lot of surprises no matter what approach you used. That's why the percentages are there; no matter how much you know, unpredictable things happen! It's part of the beauty of this time of year. If the model had been 100% correct in terms of win-loss record, I would be more critical because it'd be clear it was way too under-confident. But it seemed like it was pretty decently calibrated this year. Here's a breakdown of the renewal percentages on broadcast within each of the five colored tiers:

TierAverage R%Actual Renewal RateRenewals/Total
Dark Red (0-19%)11%0%0/12
Light Red (20-39%)32%43%6/14
White (40-59%)50%45%5/11
Light Green (60-79%)71%73%8/11
Dark Green (80-100%)95%98%42/43

(For this table and all the rest of this post, I'm using what the R% was at the time of the decision. That means shows whose fates were announced before their premieres are not included. The shows not yet decided as of this posting (Code Black, Champions, Law and Order True Crime, Timeless and Ghosted) are also not included.)

At least so far, the model ran the table on the dark red shows (and Law and Order True Crime, not included, would be a 13th). But that's not to say this tier should not be considered a certainty; the R% for all these long shots still adds up to 1.35 renewals, and typically there is something totally ludicrous like Galavant or Scream Queens that pulls through. A lot of people thought it was gonna be Rise this year, but alas. Maybe a Champions renewal is still coming......

Though the networks were ruthless with the mega-bombs, they made up for it with one or two extra of the "minor reaches" in light red territory. The biggest reach renewal of the season was For the People at 28%. That's still higher than many would've given it, but the True formula is kind to 10:00 shows and Renewology is kind to ABC dramas. Throw in fellow freshmen A.P. Bio and Instinct, plus a few others that likely fall under the "non-ratings reasons" banner in Blindspot, Madam Secretary and Elementary. Timeless is still pending and could make this look even worse if it pulls through, but at least the entire tier would still tip narrowly toward cancellation.

The true bubble shows tipped from an ideal 5/10 slightly in the cancellation direction with Monday night's axing of LA to Vegas. Still, it is good that the final percentage fell somewhere between 40 and 60, and would do so even if you count Ghosted as a cancel. (I will probably not "count" Ghosted either way if Fox actually waits to see summer ratings.)

The light green territory looks really good: an average guess of 71% and a renewal rate of 73%. Scorpion, Brooklyn Nine-Nine and Designated Survivor were the misses. It is a bit annoying that Designated and Scorpion both fall under the same umbrella of getting a lot of credit for airing at 10:00, and many people were more down on them ratings-wise than this formula. But the 10:00 factor has saved enough shows in previous seasons that I don't anticipate a big change on that front. Scorpion seemingly couldn't get a syndication deal, and Designated was plagued by behind-the-scenes turmoil pretty much throughout its run. These things happen. Code Black is pending, but like with Ghosted I probably will not "count" it if the decision comes after the summer airings begin.

And nearly half of primetime series ended up in dark green territory. Like the dark red, this tier actually may have been a bit under-confident if anything, because the only miss out of 43 shows was barely dark-green Kevin Can Wait. Usually we would expect one or two shows at lock ratings levels to fall through due to costs like Last Man Standing last year, and I thought Criminal Minds could be the one this year. Maybe I should just stock up on the wins for now, because there could be some 100% misses ahead if the networks drag their feet on ending stuff like The Big Bang Theory and Modern Family...

OK, if you're still somehow reading... I want to shift from a breakdown by R% tier to a quick breakdown by network, which I did in the post last year. For me, a big part of the evaluation of Renewology is comparing the total number of actual renewals with the total number of renewals expected by R%, because setting the renewal bar is obviously a very important part of this kind of pursuit. The news on this is better for the Renewology model than it was last year, but there's still some room for improvement.

NetSum of R%Renews

ABC: Renewology is intentionally generous with ABC, but this was the second year in a row it was too generous. It wasn't as big a miss compared with last year, when the American Idol revival probably wiped a couple renewals off the board. I was right in the ABC Renewology post that one of the low-end dramas would get renewed, but Renewology still would've expected at least one more show to make it, either Designated Survivor or a second of the pity dramas. Perhaps it's time to move past the era of wild renewals under Paul Lee and look at ABC as a little bit harsher under Channing Dungey.

CBS: If only CBS had waited till June to renew Elementary, or cancelled Code Black now, this would only be off by half a renewal or less. Throwing out Elementary and Madam Secretary (two shows that are time-tested cases of "non-ratings reasons" winning out), it's very close to 15 on both fronts. The network made quite a few "reaches" in terms of ratings strength: Kevin Can Wait and Scorpion on the green side, Instinct and Madam/Elementary on the red. But they came reasonably close to cancelling out in terms of total renewals.

NBC: Definitely the worst performance of the season for Renewology. As I've said a couple times, I took the gamble that NBC would revert toward its harsher behavior in the years prior to 2016-17, and they pretty much just renewed generously again. They were not as extreme as some were dreading, as they did cut all the dark red stuff like Rise and Great News (unless of course Champions pulls through). But they handed out a renewal to pretty much everything in white and light red territory (Good Girls, Blindspot, A.P. Bio) once again. If Timeless and/or Champions get renewed, it'll look even worse... and even if both are cancelled, it'll only add about 0.4 points to the "Sum of R%" column, still making it our biggest miss. And that's not even counting the pickup of Brooklyn Nine-Nine from another network. So yeah... it seems NBC is more generous than this formula says.

Fox: The Monday night cancellation of LA to Vegas kicked Fox into the "a little harsher than expected" column. (If it had gotten renewed, 11 renewals with 11.19 total R% would be excellent.) Even with that cancellation, Fox was the biggest win for this model relative to "conventional wisdom." It turned out to be correct to have stuff like Lucifer, Gotham, LA to Vegas and The Mick as toss-ups at best, after even I personally thought most to all of them would make it out alive. Despite doing better with Fox this year, trying to quantify how harsh they will be in the move to "New Fox" next year will be a whole new challenge...

CW: Seems like it did well with the CW again. iZombie appeared to be a late-breaking toss-up that depended on pilots, and we would be half of a renewal off in either direction depending on how that shook down. Dynasty and Valor were toss-ups (at least in the fall), and one indeed got renewed. And everything else was in dark green/dark red territory and played out as expected. Hard to complain.

By the way, if you add up the entire big five, you get 59.87 renewals expected and 61 actual renewals. So NBC needs to be a little more generous and maybe ABC a little harsher, but it seems that the model in general is pretty darn close to capturing how many shows actually get renewed and cancelled in a broad sense.

And if simple "win-loss record" is your metric of choice... Renewology went 79-12 (87%) in terms of getting the right result above or below 50% at the time of the decision. That includes a rather annoying "miss" on LA to Vegas which closed at 50.02%, but I shouldn't complain since I almost certainly overachieved by going 9-2 on the shows in the 40-60 range. While "win-loss record" is not really a good metric for these evaluations, 79-12 appears to be every bit as good as some of the other guys, despite this being a model that openly ignores most of the non-ratings stuff. Thanks for following!

No comments:

Post a Comment

© 2009-2022. All Rights Reserved.