The No-Stats All-Stars (August 2018)

Deconstructing The Myth of the “Kardashian Curse”

In the world of Kim Kardashian: Hollywood, relationships are simple. You choose a suitor from your list of love interests, take him out to Oak or La Rue, and try to collect enough hearts to eventually get married and have kids. If you’re not feeling it and ghost the dude, the relationship will fizzle, and a little broken heart will appear next to his name in your contacts. It happens. No harm, no foul.

In the world of Kim Kardashian: human being, relationships are considerably more complicated. There are still overpriced restaurants, ill-advised nuptials, and nagging reminders of exes. But with Kim and little sister Khloe, there’s also the tiny matter of the ladies supposedly incinerating the careers of every pro athlete who dares cross their path. Littler sister Kylie summed it up nicely: “The Kardashian curse is, every male figure who dates a Kardashian, their life just kind of goes downhill after that.”

The Internet listicle mill certainly seems to agree. It’s a truth universally acknowledged that every athlete pulled into the Kardashian orbit will eventually be ejected — battered, bruised and totally off his game. But what Kylie, and everyone else, forgets is that behind every so-called curse is the good luck that makes it possible in the first place.


Every year there are baseball players who refuse to participate in the home run derby. Some, surely, have merely neglected the gospel of Glavine and Maddux, but others seem genuinely worried that, after working hard to put together a first half of the year impressive enough to snag a derby invite, actually competing in the mid-season showcase might send them into a tailspin. The specter of the Home Run Derby Curse has them shook. To be fair, the idea of the derby fading performance isn’t completely out of left field; in its first 25 years, derby participants have seen their home run rate fall 11.8% and on-base plus slugging percentage (OPS) slide 43 points in the second half of the season — which sucks, but doesn’t actually mean that a home run hitting contest caused them to all of a sudden become 11.8% worse at hitting home runs.

We’ve known this for a minute. True Score theory is a concept that sabermetricians use to spell out the relative contributions of luck and skill to any given result. The idea is elegant in its simplicity: the total variation of an outcome (which could be anything from basic wins and losses to arcane advanced stats) is equal to the sum of the variation in skill and the variation in luck. That’s it. According to the theory, if we observe a drop in, say, a player’s home run rate, we know that it’s because there’s been a corresponding decline in the player’s skill or luck or both.

The guys who skip the derby seem to believe that participating will cause their hitting ability — that is, skill level — to bottom out, which implies that they don’t consider luck to be too much of a factor. But there are a couple problems with that. For one, the very best hitters — the ones who earned multiple derby invites and, presumably, wouldn’t need much luck to put up derby-worthy first half stats — experience a much milder dip in the home half of the schedule than one-and-done participants. That’s exactly what you’d expect if luck were the dominant variable; the players who were less reliant on chance to begin with suffered softer declines when lady luck inevitably turned her back. More than that, if this were truly a case of cause and effect, then the events should flow in the direction of the causal arrow — first the cause and then the effect. But it turns out that when you look at the entire careers of all the derby participants, pick out the best second half performances and then compare those to their corresponding first halves, the gaps (11.2% in homer rate and 44 points in OPS) are more or less exactly the same as the ones allegedly caused by the derby, except this time there’s no derby to blame. All signs point away from a mythical curse — and towards a natural correction after an unusually fortunate run.

The moral of the story is something we already know: what goes up must come down. True Score theory is like the gravity of athletic performance — it always pulls you back to the middle. The equation has two independent variables: skill and luck. Typically, in the short-term, the skill contribution is pretty stable, while the luck can bounce around every which way. The randomness of luck leads directly to what statisticians call regression to the mean, which is the idea that any measurement that is above (or below) average is likely to be followed by one closer to the middle. According to True Score theory, that happens because when skill is relatively static, any outlying performance must be driven largely by luck — and since we assume that next time around luck will move back towards neutral, the new outcome should hew closer to the average. The takeaway is that short-term fluctuations are sewn into the fabric of sports performance — and so if we limit the sample to guys coming off career-best showings, then that variance is very likely to knock them down a notch or two.

Kim and Khloe’s dudes are all too familiar with this phenomenon. It really seems like the Kardashian ladies somehow engaged the little-known “coming off a career year” filter on their dating apps. Much like the home run hitters — whose derby year first halves were collectively more than 40 OPS points clear of their career averages — nearly all of the Kardashian athlete suitors were riding personal bests when they got called up. The bae sample size is too small for formal statistical analysis, but the roll call is still mighty impressive: Tristan Thompson (Khloe, 2016) and Lamar Odom (Khloe, 2009) had each snagged their first NBA title a few weeks earlier. Reggie Bush (Kim, 2007) was a year removed from rewriting USC’s record books. Kris Humphries (Kim, 2010) had just finished setting personal bests in points, rebounds, assists, blocks and minutes. James Harden (Khloe, 2015) had booked new career highs in points, assists, rebounds, and Player Efficiency Rating. Scott Disick (Kourtney, 2006) had been balling so hard. Even ex-Dallas Cowboy Miles Austin (Khloe, 2010) had just logged his second consecutive 1,000-yard receiving season. Arguably, only baseball’s Matt Kemp (Khloe, 2013) was clearly in decline and even he was only 18 months out from the second of his two All-Star Game nods. Not a wasted konnection in the bunch.

Then, predictably, regression reared its ugly head. None of these guys were fated to suffer a run of bad luck, but it would have been reasonable to expect that they’d stop uniformly experiencing outrageously good luck, which is pretty much how it played out. Thompson and Odom remained useful ancillary pieces on strong teams, though both were also sidetracked to varying degrees by off-the-court scandals. Bush never quite fulfilled the promise of his college years. Austin’s production fell steadily and was out of the league within six years. Kemp bounced around the majors, but has hung on long enough to experience a bit of a renaissance in 2018. Humphries actually set career-best PERs in three of his first four post-Kim seasons before settling into a gentle decline. Disick still Disicks. Only Harden stands out as the ultimate exception that proves the rule, with his Khloe dalliance looking in retrospect like a stepping-stone on his inexorable rise to all-world status. The divergence in the various post-relationship career arcs highlights that the operative characteristic of the Kardashian cohort isn’t where they ended up, but where they started. Their stratospheric initial positions and the cold truth of regression combined to guarantee that most of these guys were headed for some kind of fall, no matter who they hooked up with. Far from casting a nefarious spell over these dudes, KK’s culpability here seems limited to their actually-pretty-reasonable habit of dating successful guys. The whole sequence is a lot like buying a stock when it’s at an all-time high, only to see it then lose some of its value: the result might be disappointing, but nobody blames the shareholder.

Even if regression to the mean debunks the notion of a Kardashian curse as thoroughly as it vaporizes the myth of a Sports Illustrated jinx, it’s not the only reason that Kim and Khloe’s time in the spotlight has largely outlasted their sweethearts’. For starters, they inhabit completely different time horizons. TV stars regularly hold their perch for decades; no athlete comes close to staying on top of their game for that long. Bush may have managed to overcome his two-year tryst with Kim to carve out a decent 11-year stretch in the NFL, but he’s an outlier. The average halfback career lasts 2.5 years; more than four have come and gone since Kim’s sex tape set her on the path to ubiquity in 2007. Really, it’d be exceptional for Kim or anyone else to date a running back for more than a year or two, and not watch his career go to shit. Same deal in basketball. When Khloe hooked up with the Cavs’ Thompson — NBA power forward, median career length: 2 years  — he’d just wrapped up his fifth year in the league. By the NBA’s brutal standards, he was into the gravy phase well before he knocked her up. For the vast majority of athletes the stopwatch starts at 14:59.

But the deeper reason why the Kardashians continue to burn bright as their exes fade away is that the probative force that numbers wield in sport is largely muted in reality TV land. Think of it this way: if Bruce Jenner scored 2% fewer points in the decathlon at the 1976 Olympics, he would have fallen out of the medals, never made it to the Wheaties box, and left the family in dire need of a new origin story. The idea of Kris Jenner falling 2% off her momager game, on the other hand, is basically meaningless. Numbers tell us nearly everything we know about athletes, but very little about entertainers.

Of course, the fame game has its own yardsticks, too. Social media popularity, in particular, is an all-purpose stat that roughly summarizes the reality star box score and looms large in our collective understanding of who’s queen of the castle and who’s played out. But that measure is both largely decoupled from day-to-day job performance and much more resistant to variation than any catchall sports statistic. Because of that, one of the critical numbers defining success in the Kardashian’s world — number of social media followers — is bound to the past in a way that, in sports, just wouldn’t play.

All of which means that, even when Kim and Khloe stumble (RIP DASH), it’s not really reflected on their scoreboard. The truth is that the number of Twitter or Instagram followers that Kim will have tomorrow depends almost entirely on how many Twitter or Instagram followers she has today. Starting position is everything, and the Kardashians long ago established themselves as the apex predators of the attention economy. Studies of social media dynamics unequivocally show that the users who gain the most followers are those who already have the most followers. As far as our attention goes, the rich get richer and the poor get unfollowed. There’s a reason that fully nine of the ten most followed people on Twitter in 2014 still hold that distinction today. Compare that to the never-ending flux in sports rankings; among the top ten vote-getters for NBA most valuable player award five years ago, only four (including Khloe’s Harden) made the grade last year. And, by the standards of professional sports, four out of ten is kind of a lot: The NHL, NL baseball and international soccer had two each. AL baseball had just one. The inertia that has allowed Kim and Khloe take — and keep — a hold on our collective psyche just isn’t available to the athletes that they date. Long live the queens.


A useful statistic has two properties. One is that it’s reliable, which basically means that it’s stable over short time periods. We prefer reliable measures because they’re usually dictated by skill, whereas unreliable ones are dominated by luck. A dice throw, for instance, isn’t reliable at all. The outcome is different almost every time, which is another way of saying that it’s based mostly on luck. Sports statistics, like OPS and home run rate, blend skill and luck and so typically grade out as moderately reliable. Celebrity social media presence, on the other hand, is way on the other end of the spectrum. Those rankings might as well be set in amber; they’re very nearly static over not only the course of a year, but the course of many years. They are extremely reliable, which implies that, unlike rolling dice, maintaining an enormous digital following is an expression of a skill — even if we can’t quite articulate exactly what that skill is.

But a useful statistic also has to be predictive. The reason we bother tallying up all these figures in the first place is to shed a bit of light on some greater objective. Athletes chase sabr-approved stats like OPS and PER that correlate reasonably well with victories and defeats. Kim Kardashian: Hollywood collects stars and experience points that chart her progress towards the A-list. But for the non-virtual Kim and Khloe, things aren’t as straightforward. The trouble is that we can’t determine whether social media follows is predictive unless we know what it’s meant to predict. And there, maybe, lies the genius. Because, really, who’s to say what the Kardashian squad goals are, or should be? The answer is probably unknowable, which means we’ll never truly have a useful numerical standard on which to judge Kim and her family. And, because of that, anyone who stands near them is going to have a hard time keeping up.