As we kick off another batch of True Power Rankings, here's a brief intro to a new number called "Heat" that will be included.
The True Power Rankings have always included only the last third of results as a way of throwing out inflated early numbers that quickly become irrelevant. While early-season results may not be as important as more recent ones, they can still be useful as a point of comparison. Are a show's recent ratings stronger or weaker than the earlier ones?
I've always wanted some measurement of "momentum," a way of saying which way a show is trending after adjusting for the changes in timeslot difficulty. The previous edition was "Last," which compared the current rolling average vs. the rolling average one episode ago. This simply wasn't a robust enough number to have any real meaning; in essence, it was just comparing the newest episode with the one episode that had just left the average.
So now I'm introducing something that I hope will be both more meaningful and easier to understand. Enter "Heat." It's pretty simple: "Heat" is the percent difference between 1) the rolling True average and 2) all post-premiere episodes before the rolling True average. For example, if a series has 15 True results, "Heat" compares the rolling average (episodes 11-15) with the average of everything else beyond the premiere (episodes 2-10).
It's then classified one of five ways with a color-coded background:
Heat +10% or higher: H O T
Heat +5% to +9%: W A R M
Heat -4% to +4%: No discernible momentum
Heat -5% to -9%: C O O L
Heat -10% or lower: C O L D
This number points out some pretty interesting things, but it still may be tweaked somewhat in the future after I see it in action for awhile. Not every single one of these things will probably get solved, but here are some things to keep in mind.
1) In general, there's a lot more cold than hot.
1a) Most new shows tend to be cold. At least in its current form, this number is kinda depressing. I put these two notes together because 1a) is a significant part of the explanation for why there's a lot more blue than orange in these tables. I threw out the season premieres because they're inflated for all series, but the early-episode inflation often goes far beyond the first episode for newbies.
I've tried a couple different things to try to balance it out a bit. I've tried comparing the last-one-third average with only the next-one-third, rather than going all the way back to episode 2. I found that helped with the new shows a bit but tended to make returnee results more volatile, so I opted for the episode 2 version. I also nearly extended the Cool/Cold borderline to -15%, since a large majority of the Cold shows fell from -10% to -14%. Both those ideas may be revisited, but not in this initial edition.
2) For unscripted shows, Heat can be useful with standalone series, but much less so with ones that significantly change within a season. For example, an audition-based talent show like The Voice will almost always go down as Cold because the show naturally drops a lot after the blind auditions, while a series like The Bachelor that grows significantly by season's end will almost always qualify as Hot. But this is less a problem with Heat itself and more about the fact that the whole last-one-third approach should be ditched with these shows anyway.
3) Reminder: it's about momentum within a season, not on a year-to-year basis. The best example of this is Once Upon a Time, which is the Coldest returning scripted series on broadcast because it dropped so much late in the fall, but it still has one of the very best year-to-year trends. So it's probably helpful to consider both "Heat" and "y2y" in concert.
There are probably even more caveats, but maybe we'll get into some of those over the course of the Power Rankings, which debut momentarily. Heat will make its debut on the Weekly Power Rankings pages starting next week (on February 18).