Thursday, August 21, 2014

2013-14 Predictions Recap: First Two Weeks

With another season about to start, it's time to take a look back at some of the many predictions made around here last year. Today, the highlights and lowlights from the First Two Weeks posts, in which I use just two (or three) weeks of ratings to make a renew/cancel call on new scripted series.

Here's the full list of guesses:

Sleepy Hollow - RENEW - CORRECT
Brooklyn Nine-Nine - defer -> CANCEL - INCORRECT
The Blacklist - RENEW - CORRECT
Trophy Wife - CANCEL - CORRECT
The Goldbergs  - defer -> RENEW - CORRECT
Back in the Game  - RENEW - INCORRECT
The Michael J. Fox Show  - RENEW - INCORRECT
The Crazy Ones  - RENEW - INCORRECT
Super Fun Night - defer -> CANCEL - CORRECT
Welcome to the Family - CANCEL - CORRECT
The Millers  - RENEW - CORRECT
Sean Saves the World - CANCEL - CORRECT
The Originals - RENEW - CORRECT
The Tomorrow People - RENEW - INCORRECT
Once Upon a Time in Wonderland - CANCEL - CORRECT
Dracula - defer -> CANCEL - CORRECT
Almost Human - RENEW - INCORRECT
Intelligence - CANCEL - CORRECT
Killer Women - CANCEL - CORRECT
Chicago PD - defer -> RENEW - CORRECT
Star-Crossed - CANCEL - CORRECT
About a Boy - RENEW - CORRECT
Growing Up Fisher - CANCEL - CORRECT
Resurrection - RENEW - CORRECT
The 100 - defer -> RENEW - CORRECT
Surviving Jack - CANCEL - CORRECT
Friends with Better Lives - CANCEL - CORRECT
Bad Teacher - CANCEL - CORRECT

FINAL RECORD: 37-7 (84%)

Once again, the record improved by an insignificant amount, from 82% last year, 80% two years ago and 75% three years ago.

The record was 5-1 on the shows deferred to week three. The deferral was generally a positive here; I think week three added considerable clarity on Super Fun Night, Dracula and Chicago PD. The only one that it really hurt was Brooklyn Nine-Nine. I don't regret saying "cancel" after week three; why defer if I was going to say "renew" after a four-tenths drop in week three? But I was pretty close to saying "renew" after week two, and then week three happened to be probably the bleakest point renewal prospects-wise in the history of the show. So if there was a "mistake," it was in deferring, not in what I did after the deferral. And I don't think deferring was that big a mistake; week three was just kinda unlucky, much like week two of Chicago Fire last year. It happens.

As always, what I'm most interested in is how 84% compares to waiting till May and predicting based on the full body of work. This season had surprisingly few May surprises among the new class. All seven of the misses seemed to be leaning toward the other direction with all the info at the end of the season. The two least certain shows at the end (at least from my view) were actually shows I got right after two weeks: Growing Up Fisher and Friends with Better Lives.

But I think The Michael J. Fox Show was the only miss whose ratings went 100% in the other direction looking just at the ratings and not the various scheduling tea leaves. There were a few other shows (Back in the Game, Brooklyn Nine-Nine and Reign) whose fates were announced in the other direction well before the upfront, but they were not exactly runaway mortal ratings locks at the time of those decisions. Back in the Game didn't even get an episode extension, but it was still stronger than Trophy Wife, and I guess ABC just decided to take a flier on the buzzier show that they owned. And shows like The Tomorrow People, The Crazy Ones and Almost Human were not that far away from the bubble either. I thought Almost Human should've been renewed, for whatever little that's worth, but I was not optimistic about it happening based on the scheduling. 

One thing I've frequently noted in the past is how rotten my record is with CW fall series. For some reason, the narrative frequently seems to shift in the second half of the fall. It happened again with The Tomorrow People and Reign. The 1-2 record on those shows drops me to a truly ghastly 4-7 with CW fall series! If this happens again with Flash and Jane the Virgin, which seem even clearer going in than most of these shows, it may be time to throw in the towel.

Looking Ahead

The whole point of these posts was to push back at the "conventional wisdom" that you need many weeks or months to be able to tell where a show is at. Having been at 80%ish for four years, slightly above that with the deferral option added, I think it's pretty clear that this is about the level of accuracy that I can achieve in two weeks. It's somewhat variable how that compares to the percentage at the end of the year, but overall the percentage at the end of the year is still well shy of 100%. So I've kinda made my point here, however strong you think that point may be.

I would like to switch this up somehow in the coming season, either introducing more of a game aspect (perhaps assigning confidence points) and/or expanding it to create a model of the accuracy across all of the early weeks (week one, week three, week four, etc.). I don't really want to abandon it completely, because I still think the early weeks of new shows are the most interesting thing about day-to-day ratings. Any ideas?


Spot said...

Again, lots of respect for doing this.
Didn't you forget to do a Betrayal post?

Spot said...

As I said on the lineups post, I hoped that one was obvious enough that nobody would call me a fraud for including it in my record!

Spot said...

Thank you kind sir!

Spot said...

I always like the idea of a game, so this is me spitballing...

Do a First Two Weeks post as usual, but leave out your Prognosis and be more neutral in the analysis. Those that comment on that initial post vote renew or cancel and bet either 2, 4, or 6 points based on how confident they are based on just those numbers. To keep the defer option available for some players, update the post with week 3's numbers. If you comment after the update, you can bet 1, 2, or 3 points. A wrong answer deducts those points from your score. It'd be a way to gauge how optimistic the players are and how strongly they feel about it.

And with this structure you could play too, Spot. I would create a dummy Disqus account so that your guesses don't influence the other players, but that's just me.

Spot said...

My understanding (which I could be very wrong about) is that the production deal that Fox made with National Geographic Channel meant that it had to air on Sundays so that NGO could reair the show on Mondays. It could have aired in any hour as far I know, but Sundays were non-negotiable.

Spot said...

Hmmm, I wonder if the reason the CW is so much harder to predict is because it's a small sample size, so it fluctuates more and in weirder ways??

Spot said...

(Sorry for the book.)

Is there anyway to analyze early on in season 2 as to whether s3 and the almost obligatory s4 is going to happen?

I'm to the point that if I really want to watch a scripted show as it airs, but don't want to get burned by 2 or less seasons, I wait until a renewal for s3 has occurred or the tea leaves really show that it's here to stay. This is usually Late Spring s2.

I think for ppl like me who want to be watching Scandal and Chicago Fire or like this upcoming season, Chicago P.D., Blacklist and Sleepy Hollow, would really benefit from a before Christmas Break during s2, s3 prediction. An accuracy rate that is high, that you feel comfortable with, would be much appreciated. It would be even nicer if Two Week into s2 was fairly reliable, so "we" start only 1 season and 2 episodes behind, or even just 1 season if our friends/family wait 2 weeks to start s2.

Heck, you could start this summer going back a few years to get a starting point to work with. Not too many shows get second seasons these days (always?), so it wouldn't be too much work.

Of the shows that get 2 seasons, a lot of them are 2 and done, esp on Cable as it seems for Premium Cable esp, unless they really bomb, s2 is almost a given. That renewal for s3 decision is where the shows must have proven themselves esp since not many shows get full 3 seasons and not a 4th. So, for those that want to wait and don't feel 2 is enough, 2 weeks into s1 is just too early.

I haven't done any real excel spreadsheets on the matter, but I do think early into s2 renew/cancel can be a decently high accurate educated guess. There will always be some things that are hard to predict, esp a Spring drop after the Christmas Break, but if you include some sort of how confident you are indicator, or not, I think it would be well worth it, for those of us who like the 2 episodes into s1, idea but know that s3 and beyond is no guarantee, no matter how confident you are in s2 occurring.

Just thought I'd throw this out there, since you ask and it's highly related.

Spot said...

First, great job, again.

Second, I don't think you should expand predictions over all early weeks. Because that way you'd burden yourself with more posts in project, yet to not much gain.

Third, confidence points. Maybe you could have tiers? Like
renew = 90% or more sure
likely renew = 66.7%-89% sure
probably renew = 55%-66.6% sure
renew, but bubble really = 51%-54% sure
cancel = 90% or more sure
likely cancel = 66.7%-89% sure
probably cancel = 55%-66.6% sure
cancel, but bubble really = 51%-54% sure

Post a Comment

Newer Post Older Post Home
© 2009-2016. All Rights Reserved.