Using Math and Logic to Predict Future Super Bowl Halftime Performers

I felt a small tinge of excitement when I found out that Shakira is performing at the halftime of Super Bowl LIV, indicating that I’m the kind of slightly-nostalgic viewer with a cursory acquaintance with popular culture that the NFL was targeting with the pick. Finding myself at all interested in the Super Bowl Halftime show was a new experience and got me thinking about what kind of performers get booked for the only football halftime show anybody watches and who might be future candidates for the prestigious position of keeping viewers from switching to the Puppy Bowl while the Super Bowl teams take a break. I wanted to score potential Super Bowl halftime performers on how well-qualified they are for the gig, so I collected data on all the halftime performers since 1998 (the year commercial sponsors started running the halftime show), excluding the years 2005-2010. I skipped these years since they took place in the shadow of Nipplegate, the 2004 Super Bowl Halftime catastrophe in which millions of unsuspecting Americans were exposed to a female nipple (which is totally unlike the wholesome male nipples viewers would be exposed to during the Red Hot Chili Peppers’ 2014 performance). These six shows featured older rock artists and seem likely to be a relic of a particular time and not indicative of what we should expect from future Super Bowl halftime shows.

Super Bowl Halftime performers are people you know

Super Bowl halftime shows center around well-established superstars. The major Super Bowl halftime performers (the ones getting their names on the commercials) were, on average, about 37 years old and 20 years into their careers. There seems to be a big focus on getting performers nearly all viewers would recognize — every performance in my data featured at least one artist who had had at least one hit that had peaked at #1 on Billboard’s top 100, and every main headliner had had a #1 hit. In light of this, I used having a top hit as a necessary condition for inclusion in my predictions — I’m really only interested in predicting the main act, anyway. This restriction has the additional benefit of allowing us to use the time passed since performers has had a #1 hit as part of the performer’s score.

While Super Bowl halftime shows feature established stars, they are also promotional events for the performers — I learned that they don’t even get paid for their performances from Josh Hermsmeyer’s article on the topic. So there isn’t much value in performing if you aren’t trying to promote your still-active career. While the average age of performers’ most recent rise to the top Billboard spot was about 7 years, this average is inflated by supporting performers like Martha Reeves (1998) and Lenny Kravitz (2015); the median time between an artist’s halftime performance and their last Billboard-topping hit was 4 years, and the mode was only one year.

So, in general, we want: (1) Really big stars who are (2) well-established and deep into their careers but (3) are still active and popular. I came up with an appropriately unscientific metric to score the appropriateness of potential halftime performers on a scale that maxes out at 50. Let’s first look at the results before briefly describing how it works. The most likely Super Bowl Halftime performers, as of January 2020, for the next five years are:

My first thought when I saw Drake topping the list was, “Oh, that makes sense,” so this metric is clearly doing a fantastic job. Drake, Rihanna, Bruno Mars, Beyoncé, Lil Wayne, and Taylor Swift appeared in each year’s top 10. (Bruno Mars and Beyoncé have each already performed at multiple Super Bowl halftimes.) However, these lists will change as artists release new songs; having more top 10 hits and a number 1 hit in the last seven years helps one’s score, so artists have time to bolster their chances, and more active artists will move up the rankings of future years if their careers are successful.

The metric looks only at artists who have had a #1 Billboard hit. Each artist starts at 50 points, with points being deducted for a few reasons. Their score is penalized if their most recent peak at the top of Billboard’s Hot 100 was more than seven years ago, with the penalty increasing with the age of this peak. Their score is also penalized due to how much their age (or the average age of their members, if the artist is a band) and the length of their career differs from those of the average Super Bowl Halftime performer. And finally, their score is penalized for ranking lower than their competitors in total top 10 Billboard hits.

Data on Billboard hits was taken from Billboard’s website. Data on artists’ ages and the ages of their careers were taken from Wikipedia. I can send the data and equation I used to get scores to whomever wants to improve the metric or is just curious.

Here are the complete 2021 rankings:

For Something about the Super Bowl Itself

3 thoughts on “Using Math and Logic to Predict Future Super Bowl Halftime Performers”

  1. There isnt anything to take into account the performers style, really. For example, while I love Ed Sheeran, I am not sure he would have the stage show style that is likely required for this show. Thoughts?

    1. I think that’s one of the major factors that this kind of analysis can’t catch easily. (At least, I have no idea how to capture that idea objectively.) For instance, Adele also does really well on this metric, but I would be shocked if she got a Super Bowl Halftime show. This is probably an area where using subjective information (like Ed Sheeran’s music being a bit too snoozy for a bombastic spectacle) would allow us to outperform a mathematical model.

Leave a Reply