Some Analytical Thoughts

 When you think of sports, numbers, and analytics, baseball dominates the conversation. It is a stat driven game, from the time the lineup card is made to the Hall of Fame selection voting. I absolutely love this article from Penn University with an accompanying podcast that takes a deeper dive into what impact analytics have had on America's pastime (link at bottom)¹. It's the numbers, stupid.

But what about other sports? The NBA impact is easy; its how a 6'3", Davidson College guard by the name of Steph Curry has become the 2x MVP and the greatest shooter in NBA history. His ability to stretch the floor has every team in the NBA chucking three's, and why wouldn't you? Why take a 15 foot jump shot when you could move back 7 additional feet and get another point? Hell, you can even shoot for a worse percentage AND score more points. Compare these stats: in 2013, the Rockets led the league in 3pt attempts per game at 26.4, while the Grizzlies were last at 13.9. The Rockets won 54 games while the Grizzlies won 50, respectively, which is not a huge difference¹. However in 2021, the Jazz led the league with 43(!!) 3pt attempts per game, compared to the Spurs who attempted the least with 28.4. The Jazz also won 52 games in a shortened season to the Spurs' 33, which is a bigger gap, but the real takeaway is attempts¹. The Spurs would have led the league 8 years ago by a full 2 more attempts per game, yet in 2021, their 28.4 meant last place and fewer wins. The analytics trend is showing us 1) its real and 2) it isn't going away soon. All players on an NBA court HAVE to shoot, its simply the best strategy.

This trend is consistent across all sports, but one sport in America lags behind the rest: the NFL. The NFL's big changes have been simple: pass more and more and more. Quarterbacks matter the most, everyone else is replaceable. The contracts prove this, the average salary for all QB's was $5.76 million while running backs made $1 million.¹ QB's also dominate the draft board and TV commentary. Just turn on a game and the first and last comments from announcers end with the QB. Every team knows no QB, no Superbowl. So okay, seems like simple analytics, get a passing QB, and pay him a hefty salary. 

538, an analytics driven website, takes this into account for its game simulations. I compared the real results from Sunday's games against the 538¹ simulator and found their simulations went 7-7 with Monday Night's game pending. Lackluster to say the least. However, when we used their "QB Adjust Mode," the results jumped to 9-5, a 29% difference. Over the course of a teams season, a 29% shift either direction is more than enough to go from top 5 pick to playoff team.

NFL teams finally understand what fans complain about every year: QB play. The future seems pretty clear to everyone: draft your guy, throw a the ball a ton, and throw the world at him. If that doesn't work, rinse and repeat....


Links:

1. Wharton Business Daily Baseball article: https://knowledge.wharton.upenn.edu/article/analytics-in-baseball/

2. Basketball Reference Standings: https://www.basketball-reference.com/leagues/NBA_2021_standings.html

3. NBA Advanced Analytics: https://www.nba.com/stats/teams/shots-general/?Season=2020-21&SeasonType=Regular%20Season&sort=FG3A&dir=1

4. NFL Salaries: https://www.cnbc.com/2019/02/01/heres-what-the-average-nfl-players-makes-in-a-season.html

5. 538 Analytics: https://projects.fivethirtyeight.com/2021-nfl-predictions/games/

Comments

Popular posts from this blog

The Pendulum Always Sways... or Does It?

We Are the Product

There's a Crash Coming...