No tags yet.
NBA Watchability Index
One of my favorite pre-NBA season articles is Zach Lowe's League Pass Rankings. This is essentially Zach's take on which teams will be most fun to watch in the coming season. I thought it would be interesting to take a more objective approach to this idea and look back at last season and use stats to develop a "Wachability Index." This is an attempt at ranking teams in terms of how entertaining (or not) they are to watch, based on objective data rather than pure subjectivity. Obviously, what data to use is subjective, but I tried to use criteria that should be both straightforward and agreeable to most basketball fans.
Determining the Criteria
Points per game was an obvious choice. Watching the ball go into the basket is fun, and the primary objective of the game. Poor shot selection is frustrating to watch, so field goal percentage was also included.
One of the most exciting developments of the modern NBA is the increase in 3-point shots, so 3-point shots made per game made the cut as well. On the offensive end, the last thing I included was lobs (alley-oops) made per game. Who doesn't love a good lob? No one.
Defensively, steals and blocks are the two most exciting things to watch. Both are scored in the index. The last metric included in the index is margin of victory/loss. Unless it's your favorite team hammering a rival, you probably would rather spend 2 hours watching a close game than a blowout.
In addition to rewarding teams for the exciting things they do, the Watchability Index also penalizes teams for sloppy and slow basketball. Long games that lack flow are boring, so teams are penalized for personal fouls per game and minutes played per game. Bad passes and poor decisions are no fun either, so turnovers per game negatively impacted a team's score as well.
Here's what some of these stats look like in real life:
3-Point Highlights
Lob Highlights
Defensive Highlights
Pretty exciting stuff, although I'm sure an argument can be made to eliminate one stat or another, or include something I left out. Overall though, I think the stats I chose produced pretty good results. Before we get to the results here is a quick explanation as to how the rankings are calculated.
Under the Hood
In order to rank teams by the above stats, I had to do two things: 1.) compile the per/game stats for each team, and 2.) develop a method for scoring each category and team. I was able to pull most of the stats from Basketball Reference. Anything I couldn't find there, I calculated from play-by-play logs from BigDataBall.
For scoring, I standardized each variable using a z-score. Simply put, a z-score is a statistical way of comparing apples to apples. For example, it's a way of making the Points/Game variable comparable to the Blocks/Game variable, and so on. More technically speaking, the z-score measures the number of standard deviations each observation is from the mean (average). So, each z-score shows how far above or below average a team is in each category.
Results
Before we get into analysis, let me explain these results. Each column represents a stat used to calculate each team's watchability score (z-score). Each cell contains the watchability score for that team and category. For example, the Atlanta Hawks score for Game Length is 1.323. Comparing this to Miami's -1.974 shows that Atlanta plays shorter games and is above average (positive score) in terms of game length. Miami is below average (negative score). The total watchability score is the total of all categories for each team.
Each column uses a green/yellow/red color scale to help quickly identify good/bad/average scores. The darkest green is the highest score and the darkest red is the lowest scores. This is useful in evaluating how each team performed in each category and how each category is represented in the rankings.
Analysis
For the most part, these rankings are what you might expect. Seeing Golden State and Houston at the top makes sense. Seeing the likes of the Phoenix Suns and Memphis Grizzlies at the bottom are not surprising either. This suggests that the Watchability Index is a reliable metric. If I saw the Brooklyn Nets at the top or Golden State at the bottom of the standings, this work would not have seen the light of day. There were some teams that finished higher/lower than you might expect though... at least if you think playing winning basketball should correlate with being fun to watch.
The most interesting result to me was the Boston Celtics. They had the 4th best record in the league but ranked 21st in watchability. The team is a perfect analogue to one of their best players, Al Horford. Al isn't flashy, and only fun to watch if you are a basketball nerd. But he is very productive, just like his team. Perfect! On the flip side, the LA Clippers do many of the things we love to watch (Lob City!), but didn't make the playoffs.
Overall, the Watchability Index shows that being successful and entertaining are mostly correlated, but not always. Some good teams are boring and some bad teams are fun. But mostly, good teams are fun and bad teams are boring.
Applying to the 2018/19 Season
The major downside of doing these ranking in retrospect is that teams change from season to season. Look for the Lakers to jump significantly in the 2018 Watchability Rankings. Look for the Cavs to drop. That said, many teams are mostly running it back this season, so this has some relevance going forward. Once we're a month or so into 2018, I'll produce in-season rankings and will keep them updated throughout the year. The in-season rankings will be more useful in deciding which games are must watch and when it's a good time to catch up on Better Call Saul.