NBA Star Treatment Bias and Our Own Cognitive Biases

September 27th, 2017 | by Allen Reihman

Brent Asay/Utah Jazz

Benjamin Franklin once quipped, “In this world, nothing can be said to be certain, except death and taxes.” If dear old Ben were alive today he may have included a third certainly: death, taxes, and NBA fans complaining about star treatment by referees. Everybody thinks star players get preferential treatment, but do they?

There are plenty of anecdotal instances people could point to. For example, let’s think back to the two most memorable plays in Jazz history.

Firstly — and famously — we fondly remember John Stockton’s buzzer-beating-three to vault the Jazz into the 1997 NBA finals. After Rockets coach Rudy Tomjanovich watched Stockton lead a 13-point comeback by destroying Sedale Threatt, he switched Clyde Drexler onto Stock, hoping Drexler’s length would would disrupt the final shot. Enter Karl Malone, who arm-pinned, bear-hugged and Texas two-stepped Drexler from the top of the key to 5 feet above the break, leaving a lonely, helpless and hapless Charles Barkley to run at Stockton from afar. Malone was not called for the moving pick, and the rest is history.

Secondly — and infamously — Michael Jordan sank his own “The Shot” a little over a year later, pushing off against Jazz defender Bryon Russell to sink the jumper that won the 1998 NBA Finals. No call on the MJ clear-out, and the Bulls hoist another banner.

Those examples and a whole lot of conventional wisdom say that referees give NBA stars the benefit of the doubt, especially with Malone and MJ, reigning league MVPs in those respective seasons.

Evidence of a perceived star standard for calls exists in recent history, too. After a play last season where Derrick Favors appeared to get hacked on attempt but didn’t get the whistle, Jazz coach Quin Snyder stomped down the sideline, berating the official: “No respect! No respect!” Implied in this complaint is Quin’s belief that Favors is a star deserving of calls. Snyder might have still protested the no-call had it been a rookie or a scrub, second-class citizens who are expected to be the victim of bias, but he certainly wouldn’t have made his case based on “respect.” To Snyder, star treatment is axiomatic — it is a given. If basketball savant Quin Snyder takes disparate ref treatment as fact, it must be true, no?

Most players, coaches, GMs, fans and writers believe that the star bias of NBA referees is a given, just a part of NBA life. Now that the NBA provides a last two minute (L2M) report detailing call accuracy in close games, we can find out how true that assumption is.

From the NBA’s last two minute report

This data came from the NBA’s official last two minute reports from 2015 through 20171. The data was compiled by looking at only the personal fouls logged in the report2, and “stars” are defined as players who made All-Star teams between 2013 and 20173.

This data strongly refutes the “star treatment” hypothesis. With a standard deviation of ~2%, the likelihood of a call favoring a star is a coin flip — at least in the final two minutes of a close game.

Competing Hypotheses?

Are there competing hypotheses that allow for a star bias to exist? Is it possible that the last two minutes are an unrepresentative sample? In other words, do stars get better treatment in the first 46 minutes and not the last two? Do stars get calls in blowouts but not close games? These hypotheses are possible, but don’t seem likely.

Perhaps there are confounding variables in play that skew the analysis, a hidden factor that offsets the star bias? Perhaps, but none come to mind.

Or maybe the conspiracy theorists could posit that the back-office NBA officials are covering-up referee errors by not giving the objective truth about missed calls. Some hyper-specific L2M entries have been criticized, but in general it would be hard for the NBA to gloss something over to that degree. I have carefully reviewed hundreds of bad calls and the vast majority of corrections are obviously true.

Meta Bias: Biased about the Star Bias

So for now, let’s apply Occam’s Razor and assume these competing hypotheses are incorrect, and that the more elegant explanation is true: NBA referees do not have a star bias. We infer from this that there IS a bias in the NBA community, which believes that a non-existent bias is real. What cognitive biases might explain this mass delusion? Here are a few possibilities to consider.

We can break the biases into two buckets: origination and perpetuation

Origination is the trickier factor to evaluate: how did we first come to believe in the star bias? The genesis is as impossible to predict as the question of life’s origination on earth (pardon the hyperbole). Everyone has a natural bias to notice bad calls against their own team. Vivint SmartHome Arena does not rise in a standing ovation when the Jazz get a favorable call, fans just see it as the right call. But when a bad call goes against the Jazz, fans jeer and shout unrepeatable expletives. We have a negativity bias, meaning that unpleasant things have a greater effect on our psyche than positive things. This can be compounded by a victim mentality: we are more comfortable scapegoating an outside factor (officiating) than to blame the team we love. As when we get beat by a better team with more stars, we remember those calls against us and the perceived slight to our franchise.

Another possibility, one that I believe to be quite plausible, is that star treatment was real when the orthodoxy was born, and it now lives on only because of the biases that perpetuate it long after it has vanished. Remember, we have only 3 years of data, so a bias may have existed back in the MJ and Mailman days that doesn’t appear in this sample of L2M reports.

Anchoring Bias is when our brain relies heavily on the first information we perceive (the “anchor”) when forming our opinions. We are told there is a star bias during our NBA indoctrination and it sticks.

This can be compounded by Authority Bias as we attribute greater weight to the opinion of NBA experts. And further perpetuated by a Bandwagon Bias where a group adopts the same beliefs due to social pressures.

Further reinforcement may come from a Confirmation Bias, our tendency to seek information that verifies what we “know” to be true. So when we see a star receive a favorable call, we say, “See! stars getting calls!” Yet when a rotation player gets a favorable whistle or non-whistle, we dismiss it as routine and random.

The Boiling the Frog Bias suggests that our brains tend to not notice gradual changes, so if the NBA has corrected the problem incrementally over many years, we may have not noticed it. The admittedly disgusting metaphor is one of a frog in a heated pot not noticing gradual temperature change and remain submerged until death. Whereas a frog thrown into a boiling pot would immediately jump out. Whether this is real or apocryphal, please don’t report me to PETA.

Remember, our brains are constantly playing tricks on us and clouding our perception of reality. So the first time you see Ricky Rubio get hammered, followed by a no-call, consider that it could just be a random and honest mistake, and not because the guy who fouled him is named Steph Curry.

Let me know your thoughts and if you would like to analyze the data in different ways, your suggestions are welcome.

 

Allen Reihman

Allen Reihman is a father of two Jazz fans, chemical engineer, business development director, and long-time Jazz fan. He believes that our brains frequently lie to us, and logical analysis can expose cognitive biases and enhance our intuition about the Jazz, the NBA and the world.

Latest posts by Allen Reihman (see all)

8 Comments

  1. Paul Johnson says:

    I perceive a potential problem with your analysis. You defined “stars” as all players who made all-star teams from 2013 to 2017. In my unscientific observation of “star treatment,” not all “all-stars” are treated the same when it comes to “star treatment.” Rather, in my observation, only the biggest stars on the most popular teams generally get star treatment–and as a star’s influence in the league diminishes, he may no longer be getting star treatment, even though he is still on the all-star team.

    For example, in 2013 the Eastern all-star roster consisted of the following:

    Carmelo Anthony (Knicks)
    LeBron James (Heat)
    Kevin Garnett (Celtics)
    Rajon Rondo (Celtics)
    Dwayne Wade (Heat)
    Chris Bosh (Heat)
    Tyson Chandler (Knicks)
    Luol Deng (Bulls)
    Paul George (Pacers)
    Jrue Holiday (76ers)
    Kyrie Irving (Cavaliers)
    Brook Lopez (Nets)
    Joakim Noah (Bulls)

    Out of those 12 players, by my subjective assessment–at that point in their careers and at that point in their teams’ fortunes–certainly only LeBron James and Dwayne Wade, and possibly only Paul George, Carmelo Anthony and Kevin Garnett, actually received much star treatment by referees. By analyzing and averaging data for all of the all-stars on that all star team–rather than analyzing data for only the biggest stars who actually received star treatment at critical times to help their teams advance to the finals or win a championship–you have skewed and watered down the data on star treatment.

    Another example: Last season Gordon Hayward of the Jazz was an all-star and Kevin Durant of the Warriors was an all-star. However, if you ran the data on each player for star treatment, I believe you would find that Gordon Hayward received almost no star treatment (because he was a marginal all-star from a non-contender) whereas Kevin Durant received considerable star treatment (because he was the biggest star on the best team in the NBA). By merging their data together and then averaging it out, it would appear that collectively, the two of them received almost no star treatment, even though by himself, Kevin Durant may have received considerable star treatment.

    • Allen Reihman says:

      Thanks for your thoughts, Paul,

      We can easily test your hypothesis. Provide your list of who you consider to be stars who get favorable calls & I’ll run the data and post the results.

      • Paul Johnson says:

        Last season, in my subjective assessment, these are the primary players who received star treatment: LeBron James, Kevin Durant, Draymond Green, Stephen Curry, James Harden, Chris Paul, Russell Westbrook. Damian Lillard, Karl-Anthony Towns, Kawhi Leonard and Anthony Davis also perhaps received some marginal star treatment

        • Allen Reihman says:

          Thanks Paul. With LeBron, KD, Draymond, Steph, Beard, CP3, Russ, Dame, KAT, Kawhi and AD, 34 favoring and 35 opposing in 2016-2017 season.

          With those same players over past 3 seasons, it was 100 favoring and 99 opposing.

          So it remains a statistical coin flip for those stars.

          Again, thanks for reading and your comments are always welcome!

          • Paul Johnson says:

            Have you run it on any individual players? I think we have all seen (multiple times) Kevin Durant, James Harden, or LeBron James drive through the middle of the key near the end of a close game and get a foul call (which no other player would get), and on which the replay shows no contact whatsoever–what I call a “phantom foul call.” Because of the ability of those players to shoot foul shots that is like handing that player (and his team) a free 2 points. Also, with Draymond Green, we have all seen (multiple times) him get an alleged steal and on the replay we could see he committed an obvious foul by hacking the crap out of the player with the ball, with no foul call.

            In fairness, there are many more times where in real time there appears to be no foul, or appears to be a foul, but the replay shows that the referee made the correct call.

            That’s where the “myth” of star treatment comes from–from TV viewers watching replay of lame calls by referees made to favor star players at critical points in games.

  2. Allen Reihman says:

    Hi Paul, Yes, you can run any individual player. The challenge is that the sample size is not big enough to draw any conclusions.

    What the stats say is that for every one of the missed calls favoring a “star category” there are an equal number opposing them. So we remember the big miss against our team but forget the miss in their favor.

    Our biases are strong.

    If you believe L2M, NBA referees get foul calls correct ~93% of the time. Which, given foul subjective and how hard it is to get calls right, is pretty amazing. In real time refs don’t have multiple camera angles, slow motion and 5 minutes to get it right. Putting 93% accuracy into perspective, think about how easy it is to shoot a free throw and the best in the NBA make ~90%. (WNBA’s Elana Della Donne shoots an amazing >94%, but I digress).

  3. Mitch says:

    One thing you don’t include here is a baseline comparison of any sort. Maybe stars are getting a coin flip, but is it possible that more average players are more unfairly treated? What about a subset like rookies or sophomores, who are generally considered to get less respect? I’m just curious what those comparisons look like.

    • Allen Reihman says:

      Thanks Mitch. The calls are zero sum. So the net calls against stars equals the net calls in favor of non-stars, and vice versa.

      There were 757 calls favoring non-stars and 736 calls opposing non-stars, so 50.7% favorable, not statistically different from 50% coin flip.

      The net is +21 for non-stars and -21 for stars.

      Hope that helps.

      Keep in mind that if there is negative bias against rookies/sophomores, this would imply that there would be a larger positive bias in favor of 3+ year players who are not stars (again, zero sum). I can’t think of a reason why veteran rotation players would get more favorable treatment than stars, so this seems pretty implausible to me.

Leave a Reply

Your email address will not be published. Required fields are marked *