Quill Scores with NFL Data

November 4, 2014 Dan Platt

Here at Narrative Science, we’re a group of data junkies who luckily have an awesome tool at our fingertips that allows us to create stories using interesting data sets. That tool is Quill, our automated narrative generation platform that creates data-driven communications at unlimited scale. It’s the perfect recipe for our data-crazed office.

Our latest project focuses on the NFL and culminated in a story that was first revealed last month at Chicago Inno’s Football 2.0: Innovation in Sports event.

Two of my colleagues, Shingo and Craig, wanted to look at the experience of NFL players, specifically, how winning teams in the past compare in terms of the level of experience on their rosters. This idea informed our configuration of Quill – what the content would be and how we would structure our story based on the data in hand. Once we tell Quill what the context of the data is, the platform can produce automatically generated narratives without human intervention.

Some of the highlights Quill exposed were:
  • Since 2000, teams with an average of 4.1 years of experience per player have won the most Super Bowls.
  • Wide receivers have the shortest careers, averaging 4.9 years.
  • Kickers enjoy the longest careers with 7 and a half year average playing spans.
  • The least experienced Super Bowl winner was the 2013 Seattle Seahawks.

So, we used the following data sets, and computed the following analysis for Quill to generate the final story.

 

Data Sets:
  • Player table
  • Start year
  • Birth year
  • Position
  • Schedule table
  • Season
  • Teams
  • Result
  • Week #
  • Game by game offensive stats
  • Game by game defensive stats
Computation:
  • Calculate year-by-year roster by determining players who accumulated stats for team/season.
  • For each roster record, calculate age and years experience based on birth and start years.
  • Calculate average experience of all rosters.
  • Figure out average experience by position based on all rosters.
  • Determine “last year” for each player by figuring out the last year that player appeared on any team’s roster, which also allows for “career span.”
  • Calculate career span by position group.
  • Compute average experience of rosters that won the Super Bowl, and rank each by average experience.
  • Calculate average experience by position among Super Bowl winning rosters.
  • Compare 2014 roster of each team against the Super Bowl winning rosters.

The story reads like something you or I could write, but was generated by Quill (you can read for yourself below).

 

What Does Experience Really Tell Us?

The 2014 NFL regular season is here, and so is the start of another armchair quarterback prognostication period. With absurd amounts of data and the era of modern computing alive, the league is ripe with statistics and historical summaries that can point fans, gamblers, players and coaches in many directions. Of this vast data pool, there is information available that fans can use to increase their chances of rooting for a winning team, that gamblers can use to improve their odds of raking it in in Vegas, and that coaches and owners can use to better ensure a Super Bowl win.

One way NFL followers can determine whether or not a team is on the rise or decline is through a look at player experience and what the mix has been for Super Bowl winners over time.

Statistical analysis reveals that since 2000, NFL first-team player careers average close to 4 years. Kickers enjoy the longest careers with 7 and a half year average playing spans. Quarterbacks typically have 7-year careers while offensive linemen average 6 and a half years on the gridiron. Wide receivers have the shortest careers, averaging 4.9 years. The average career spans for all players and the rest of their respective positions are: defensive linemen (5.8), linebackers (5.3), defensive backs (5.3), tight ends (5.3), and running backs (5.1).

People might initially conclude that teams with the maximum amount of experience are most likely to have a successful season. If that were true, then this year, the New York Giants have the best odds of winning a championship, as they have 4.8 years of experience on average per player.

It is actually the case that since 2000, teams that have an average of 4.1 years of experience per player have won the most Super Bowls.

This year, teams with similar levels of experience to those Super Bowl winners include Baltimore, Washington, Arizona, Carolina, and Kansas City. Of these teams, Washington and Kansas City have running backs with similar levels of experience.

Of those winning teams with 4.1 years of experience on average, their running backs average just over 3 years of experience where their kickers average 7.2 years of experience. The most experienced champions since 2000 have been the 2003 New England Patriots, the 2009 New Orleans Saints, and the 2000 Baltimore Ravens. The least experienced Super Bowl winner was the 2013 Seattle Seahawks.


 

Dan Platt is a Senior Content Architect at Narrative Science. Connect with Dan on Twitter.

Previous Article
3 Women in Tech
3 Women in Tech

A demographic I find inspiring at Narrative Science is that nearly 30% of our employees are women; somewhat...

Next Article
What Kevin Spacey Can Teach Us About Content Marketing
What Kevin Spacey Can Teach Us About Content Marketing

But after hearing Spacey’s story, it all began to make sense. Who else would have better insight on storyte...

×

Get Narrative Science blog posts in your Inbox

Keep an eye out for your confirm email!
Error - something went wrong!