Real-life superteams may not live up to the same hype in fantasy.
When NBA stars pair and triple up in hopes of earning a title, their skills are often (but not always) compatible. But the ramifications in fantasy basketball haven't been so positive.
With Paul George joining Russell Westbrook in Oklahoma City and Chris Paul joining James Harden in Houston, we figured it was worth asking how the superteam design has panned out in the fantasy world in the recent past.
For the sake of this investigation, we'll call the birth of a superteam the moment when a star -- a subjective concept that we'll define here as a player who has made an All-NBA team in the past three seasons -- joins a team that features at least one other player who meets the same criteria. By that definition, 14 "superteams" have been created in the 10 offseasons prior to this one, though we disqualified three of those teams because one of the stars did not play for his team long enough to qualify for statistics leaderboards. We also ignored superteams created midseason.
Some of the 11 remaining superteams, like the 2016-2017 Warriors, went on to accomplish great things. Others, like the 2012-2013 Lakers, didn't exactly live up to expectation.
But regardless of the team's ultimate performance, the question we are really seeking to answer is this: Does pairing stars together help or hurt their numbers?
To do this, we contrasted each star's six standard fantasy counting statistics (points, assists, rebounds, three-pointers, blocks and steals) in the season before the creation of the superteam with his performance in those same categories a year later. In order to account for pace, we actually looked at those statistics per 100 possessions. Of course, adding Dwight Howard to your team presumably affects the dynamic differently than, say, Deron Williams. Ultimately, then, we considered a player's statistics only after he was paired with a new star teammate ranked in the top-20 in one of those pace-adjusted statistics the previous season.
The most noticeable change came when stars joined up with a scorer. On the superteams in question over the past 10 seasons, 20 stars began playing with a new star teammate who ranked in the top 20 in points per possession the year prior. In 18 of the 20 cases, the teammate of the scorer saw their points per possession drop. On average, the scorer's All-NBA teammates scored 3.41 fewer points per 100 possessions in their first year playing together compared to their previous season. If both (or more) stars joining up were top-20 scorers the year before, their numbers dropped even further -- 3.95 points per 100 possessions, on average.
That is not altogether surprising. The cliché that there's only one basketball to go around is actually relevant here.
That same idea applied to several other statistics, notably rebounds and assists. After starting to play with a new teammate who was a top-20 rebounder the previous year, stars lost 1.07 boards per 100 possessions. Assists dropped by .62 per 100 possessions when stars began taking the court alongside a top playmaker.
Somewhat surprisingly, there was a negative correlation between playmakers and scoring and scorers and assists. In other words: While one might think that adding a player who racks up a ton of assists might help drive up his star teammates' points, it actually had the opposite effect. When a star was joined by a teammate who ranked among the top 20 in assists per possession the year prior, the former dropped 2.85 points per 100 possessions the following season. And the reverse was also true: Assists went down by .58 per 100 possessions when a top-scoring star became the playmaker's new teammate.
Granted, all of those findings came via an incredibly small sample size. That is a function of the fact that All-NBA players aren't constantly swapping teams to play with each other, as much as it might feel that way at times. But for fantasy purposes, whether or not a squad is really a superteam doesn't matter all that much. Getting a new teammate who excels in one or more areas of the game related to fantasy stats -- regardless of how good they are at all other aspects of basketball -- does.
So we lowered the bar a little bit by defining stars as merely those players who made an All-Star team in any of the previous three seasons, increasing the sample and allowing the 2017-2018 Celtics to enter the club.
The results? They remained roughly similar, though slightly muted.
Points dropped by 2.37 points per 100 possessions when one All-Star joined up with another All-Star who was a high-volume scorer -- and again, dropped a little further if both players were top-20 scorers.
Rebounds, in a sample of 17 players -- still very small -- were down .64 per 100 possessions. Assists dropped by .65. Assists when adding a scorer -- together with the inverse -- remained negatively correlated. Even with the expanded sample, there were no discernible takeaways from the other three fantasy counting statistics.
At this point it is worth noting that these observations are simply a description of what has happened in the past. They aren't necessarily predictive. And while this kind of year-to-year comparison can be useful, it leaves out plenty of details. Like the quality of a superstar's teammates on his old team. Or the caliber of non-star teammates on his new team. Or the age of any of the players in question. All we're doing here is taking a look at the output changes as a result of two or more stars teaming up.
That being said, if this latest batch of paired stars followed the same path as those that came before them, what would happen to their stats this year?
Let's start with the Celtics. Both Isaiah Thomas and Gordon Hayward finished among the top 20 in points per possession last year, meaning both -- along with Al Horford -- will likely lose a couple of points per 100 possessions in 2017-2018.
The same can be said for the Rockets' Paul after joining Harden, the sixth-ranked scorer per possession last season. The Thunder's two superstars would drop points as well. And in the case of both teams, the combination of adding either an accomplished scorer or distributor doesn't bode well for any star's assist numbers (based on superteams of the past).
Maybe these premier players will buck the trend. But if you draft these stars expecting repeat numbers from a year ago, just know you're betting against superteam history.
For more from ESPN Analytics, visit the ESPN Analytics Index.