Last Wednesday, the NBA released its 2014-2015 schedule. To the disappointment of Memphians and fans of the Grit and Grind everywhere, the Memphis Grizzlies will appear in exactly one nationally televised game. That means that the Grizz, 50 game winners a year ago, will be featured either on ESPN, TNT, or ABC the same number of times as the Bucks, 76ers, Jazz and Magic combined. The cumulative win total from last year of those four teams? 82.
Before returning to grad school this fall, I wanted to continue applying regression to research and make sure I didn't lose what little statistical skills I've accumulated. Analyzing basketball is more light-hearted than the democratization process of autocratic regimes. And, it's infinitely less depressing than researching gun violence (spoiler alert: the USA has a lot of gun deaths). I wanted to ascertain which factors influenced the number of times a franchise was showcased on national television, and determine the influence of each factor. So, I collected data from the last three seasons and ran a regression analysis using three main factors. The regression analysis will demonstrate how important each factor is in determining the national TV schedule, and a produce a predictive model implying how often each team should be featured on ESPN, TNT, or ABC.
- Win Total From Previous Season: Measures how good a team was the year before. This was done for each team for the 2012-13, 2013-14, and 2014-15 seasons. Because the year prior to 2012-13 season was a lockout induced 66 game season, each team's winning percentage from that year was multiplied out to produce a win total equivalent to the standard 82 game schedule.
- Market Size: Measures the size of each NBA city, and presumably how many people in each city have access to watch their hometown team play. The Nielsen Television Market was cataloged for each of the 29 American NBA cities for the years 2012-2014. Exact figures for Toronto's market size were difficult to obtain, so thesportmarket.biz estimate of 2,703,000 homes with televisions was logged for each of the three years.
- Team Facebook Likes: Rather than measuring the size of the city in which a team plays, measuring the number of Facebook likes each team has accumulated offers unique insight into the popularity of each team. The numbers were derived from the current number of team likes, as well the number of likes in 2013 and in 2012.
These three variables were then compared to the total number of non-NBATV nationally televised games for each of the 30 teams over the last three seasons. Regression analysis allows you calculate what percentage of the change in the Y variable (in this case, national tv games) is caused by changes in the X variable (in this case, wins, market size, and Facebook likes).
The results from this analysis proved informative and instructive. Win total, market size, and Facebook likes were calculated to be responsible for 65% of the number of games each team had on national television. The P-values were all significant at the .00 level, meaning there is less than 1 in 100 possibility these calculations occurred by chance.
Although disheartening to Grizzlies fans upset that their perennial playoff team will only show up once in prime-time television, the previous season's win total was shown to be a driving factor behind the national schedule. On average, a team could expect to see 1 game on national TV for 2 games they won. Market size was found to have some effect, though less than anticipated. For every one million homes with a television, a team is expected to be rewarded with 1 national game. Facebook likes, measuring overall popularity, proved to be more potent a factor than traditional market size. For every 50,000 Facebook likes a team managed, an increase of one national tv game was expected.
What About Star Power?
More than any other league, the NBA is driven by individual stars. Fans want to see their favorite players, and network executives want fans to watch their channel. An initial analysis of this subject included the most popular player from each team determined by their Facebook popularity. Unfortunately, that limited the data set to only the current year, as past statistics on player pages proved too difficult to find. A data set comprised of previous season win totals, market size, team Facebook likes, and player Facebook likes for just one season was found to be statistically insignificant. However, there is a strong correlation (70%) between current top player likes and current team likes.
In order to calculate how much each team would be expected to be on national television, you multiply each team's previous season win total, market size, and Facebook likes by their respective coefficients. Then, you subtract the sum total from the Y intercept of -12.17784305. Essentially, if a team had 0 wins the previous season, had 0 homes with a television, and no one liked them on Facebook, it is assumed they would be on your television set -12 times a year. Although this makes little real world sense, that's how this stuff is calculated statistically.
Believe it or not, 18 of the 30 NBA teams will be showcased on national TV within 3 games of what would be expected by this model. While it may seem absurd that the Los Angeles Lakers will be on TV twenty times, they play in a top market and remain the most popular team according to Facebook.
Despite playing in the country's biggest market and boasting three certifiable stars in Deron Williams, Joe Johnson, and Brook Lopez, the Brooklyn Nets have the worst actual vs. expected differential in the entire league. This could be in part to the fact that they were featured 29 times on ESPN, ABC, and TNT in the past two years combined. Viewer fatigue may have set in. Additionally, last summer's additions of Paul Pierce and Kevin Garnett, along with first year head coach Jason Kidd cultivated an atmosphere of excitement that simply isn't there this year.
Perhaps owing to the unexpected loss of Paul George and the unforgivable loss of Lance Stephenson (seriously, Lance forever), the Indiana Pacers will only be on display five times this season. That's nine times less than the model predicted. Most experts forecast an average season for Indiana. With their smaller market and pedestrian overall popularity, it's no surprise network heads chose not to focus their prime-time resources on Indiana.
Driven by consistent winning and fueled by the popularity of Kevin Durant and Russell Westbrook, the Oklahoma City Thunder will be shown nationally 25 times this season. That's tied for most overall, and yet still nine more games than what was predicted. I would imagine the popularity of KD and Russ is responsible for the differential.
I'm not sure if you guys were aware or have heard about it, but LeBron James returned home and signed with Cleveland this off-season. Couple that with the agreed in principle trade that would bring Kevin Love to the banks of the Cuyahoga, the Cavaliers unsurprisingly will receive an enormous amount of attention. The addition of LeBron and Love make this team fundamentally different from the one that won a paltry 25 games in 2013-2014. However, even if this Cavs team won 55 games last year, they would still be 12 games over the number predicted by model. None of that matters, because the King has returned, Simba has come back to reclaim Pride Rock, and the Cavaliers will dominate the media market.
How did our Grizzlies fare? Well, if you want to talk about getting screwed with your pants on, the Beale Street Bears got rogered but good. Despite winning 50 games the last two seasons (three if you extrapolate the lockout shortened season to 82 games) and making the playoffs four straight years, the Grizzlies get one national game. One. That's nine below what was expected, or the second worst differential in the league. Contrary to popular belief, it isn't because the Grizzlies play in the second smallest market in the NBA. And it isn't only because the team is in the lower tier in terms of Facebook popularity. Unfortunately, the combination of both happening at the same time has doomed the Grizzlies to a life of borderline non-existence in the eyes of the national television market. Without a name brand star of which to speak compounded by a lack of mainstream popularity, the Grizzlies don't have a realistic chance at consistent national exposure.
The lesson learned here: life's tough in the aluminum siding business. Capitalism may dictate that the Cavs should be on TV every week, and so should the Thunder and even the Clippers. Personally, I expect the Grizzlies to have some flex games added on ESPN or TNT as the season progresses. Until then, the Lakers can have their 20 nights to shine. As long as the Grizzlies are playing nationally televised games in late April and May, I'll be just fine.