It’s a typical Sunday afternoon, with NFL fans all over the country tuning in to watch their favorite teams compete for gridiron glory. But this season has featured an unexpected weekly element – everybody’s waiting… everybody’s watching… Will Taylor Swift be in attendance when the Kansas City Chiefs play?
Since rumors have swirled about the pop icon dating Chief’s player, Travis Kelce, fans have noticed a pattern. When Swift is in attendance, the star tight end performs well and the team wins. But, when Swift is absent, Kelce underperforms. Some fans are convinced this “Taylor Swift effect” is real.
Let’s examine whether Taylor Swift’s presence actually has a cause-and-effect relationship with Travis Kelce’s performance, and the Chief’s success.
The Background Facts
There are a surprising number of timelines detailing the Swift-Kelce relationship, but one clear fact is that Swift’s first game-day appearance was on September 24, when the Chiefs defeated the Chicago bears in Week 3 of the NFL Season. Swift was there when the Chiefs defeated the NY Jets in Week 4, but missed the Chief’s week 5 victory over the Minnesota Vikings (October 8). She was present again for their Week 6 (October 12) and 7 (October 22) wins over the Denver Broncos and LA Chargers. In those two games, Kelce recorded season-high receiving yards. But, when Swift was absent in Week 8 (October 29), Kelce underperformed and the Chiefs lost. To some, this was pretty convincing evidence that Swift can affect a team’s end game.
Is It Possible Swift Could Affect Kelce’s Performance?
It is entirely possible that Kelce plays differently knowing Taylor Swift is watching his every move, as her presence may influence his psychological state. There are various theories relating one’s emotional state to sports performance. It has been proposed that there is a “inverted U” shaped relationship between one’s anxiety level and performance. Too little anxiety could yield sub-par results (e.g., giving insufficient effort), while too much anxiety may produce a jittery athlete who commits unforced errors. The right balance depends on the sport, and power sports like football tend to favor a higher arousal for peak performance. Research shows that situational pressure (i.e., a close score) seems to affect NFL kickers’ performance (the “choking” phenomenon), and concussions are more likely in high stakes games. Psychological factors can definitely affect how NFL players play, but I am not aware of any research studies investigating how being watched by one’s newfound partner affects athletic performance.
What Is The Probability of The Current “Streak”?
The supposed “Taylor Swift effect” is based on the results of six games, four where she was present and two absent. While their Week 8 loss was attributed to her absence, they still won without her in Week 4 – so, a close link between her presence and team outcome is suspect from the start.
That said, let’s focus on the Chief’s loss without Swift in Week 8. Theoretically, a given team has a 50/50 chance of winning any game (if we don’t’ consider ties). However, we know that all teams are not equal – if the League’s best team plays the worst team, the best team is likely (but not guaranteed) to be victorious. A number of algorithms have been developed to predict the winner of a sporting event, and the Elo system is among the most widely used for this. In brief, an Elo rating is computed for each team, based on past performances and this is then updated after each sporting performance. The difference in Elo ratings between two opposing teams can be used to compute the probability of a win or loss in an upcoming matchup. Adjustments can be made to account for other factors which can influence the outcome, such as home field advantage or an injury related loss of a key player. A common formula used for NFL game probabilities is that from FiveThirtyEight.com, so we will use that here with Neil Paine’s publicly available data set.
The Chiefs have been a strong team in recent years, appearing in the Super Bowl in 2020 and 2021, and winning it in 2023 – so they have earned high Elo ratings. The Chiefs have been a clear favorite for every game they played this season – including their Week 1 loss to the Detroit Lions (when the Chiefs had a 78.9% win probability), before Swift was part of gameday culture. When Swift appeared at her first game in Week 3, the Chiefs had a 90.5% probability of defeating the Bears. The next week, they had a 75% probability of beating the Jets (and did). So, their wins should be expected regardless of Swift’s presence – after all, in the previous season, they were the winningest team in the NFL.
It does seem like a real glitch that they lost to the Broncos, when they had an 82.5% probability of winning. It’s actually pretty unusual for there to be such a mismatch – when we examine the 2021, 2022, and current seasons together (over 600 games), there have only been 36 games so mismatched that a team had a winning probability of <20%. Of those, the underdog has only emerged victorious on 5 occasions – or 13.8% of the time. The most extreme recent example is the Houston Texan’s victory over the Tennessee Titans, despite their 10.2% pre-game probability coming off an eight game losing streak in the 2021 season. But, that’s exactly why these pre-game probabilities are just that – probabilities, not prophecies. Is it surprising that the Chiefs lost to a much weaker team? Sure – but it’s not unprecedented.
Kelce’s Performance: Now vs. Pre-Swift
There are 22 players on the field for any given play, so Kelce is not wholly responsible for a win or a loss. Rather his performance is one component of the team’s status. Nevertheless, the spotlight has been on how the SuperBowl LVII team captain plays when Swift is in the bleachers.
In Figure 1, we can see that how he performs in a given game, simplified by his total yardage, is pretty unrelated to the team’s probability of winning. Long story short, sometimes he has a big game, others times he doesn’t – regardless of the team’s prowess.
His 179-yard performance in front of Swift on October 22nd is the second best of his career – only bested by his Week 15 performance in 2021. Of course, that pre-dates their relationship. So, Kelce is capable of huge performances with or without Swift.
A widely cited statistic is that Kelce averaged 99 yards at games where Swift was in attendance, and only 46.5 yards when she was not (which includes the Week 2 game, before she entered the picture). When we add the recent Broncos loss, this number moves to 50.3 yards. However, there is a lot of variability in week-to-week performance – and a small sample size. Yes, Kelce averaged 99 yards in the games Swift was at, but this is elevated by his two huge performances in two games. In the other two games she viewed, he had 60 and 69 yards – which is consistent with his median of 69.5 yards over his NFL career.
These numbers are not all that different from his Swift-less 67 and 58 yard performances in Weeks 5 and 8, respectively. When we examine a histogram of all of Kelce’s performances across his NFL career (Figure 1), we can see he has had plenty of >100 yard games (35 before Swift, to be exact), and plenty of <40 yard games (34 total) over the years.