TCU has played much better in the second half of games this year. What other schools have seen dramatic shifts of play after halftime?
22 Jun 2004
interviewed by Aaron Schatz
We've written a good deal about the Tennessee Titans and New England Patriots here at Football Outsiders, but the Green Bay Packers are another team on the forefront of using objective and statistical analysis in the NFL. The leader of this effort is the Packers' director of research and development, Mike Eayrs.
Eayrs (pronounced AY-ers) has spent 24 years developing computer applications to study football, and served as director of research and development for the Minnesota Vikings from 1985-2000 before Mike Sherman hired him away and brought him to Wisconsin to work for Minnesota's NFC North rivals.
One of the most important aspects of Eayrs' job is the use of digital video to track which players are performing in which specific game situations. You can learn more about this system in this article from Ziff-Davis' line of computer magazines. Here's a good example of how it works:
Packer running back Ahman Green [...] fumbled the ball seven times in the first nine games last year. In his case, the team queried plays in its database by his No. 30 jersey over the last two years. A compilation of plays involving Green took minutes, instead of daysÃ¢â‚¬â€?which is what it took three years ago when team staff had to sort through piles of tapes.
Coaches reviewed the compilation of plays and determined that Green fumbled when his elbow wasn't horizontal to the ground as he was hit. When he cradled the ball with his elbow in a horizontal position, Green didn't fumble. Using that business intelligence, the coaches could point out elbow positioning to Green, who made a mental note and made the adjustment. Green only fumbled once (he recovered it) during the team's last seven games.
I recently had a chance to share some of our Football Outsiders analysis with Eayrs, and I asked him a few questions about the work he does to help get Green Bay back to the Super Bowl.
Football Outsiders: I understand that you began as a regular on-field coach before becoming involved in research and statistical analysis. How did you become interested in doing statistical research, and how did your career develop to where that became your primary position?
Eayrs: Necessity is the mother of invention. I was extremely happy as an assistant football coach in the NCAA College Division. College Division contracts are often split, with a percentage being for coaching duties (40% in my case) and for academic responsibilities (60%). In grad school I had received good grades in research methods and statistics, plus some of the work had also been published, so the Dean asked if I would teach these classes within the HPER curriculum. Since the majority of my students were going to be future coaches I decided to use sports statistics as the basis for project and lab work within the classes. Originally we did a descriptive study of the programs within our conference and established a set of recommendations, which became the basis for our program's recruiting and football philosophy. We eventually did studies on major college conferences and the NFL and sent the results to several of those programs. Some organizations responded and I started doing consultant work for them on a project basis. One day, after presenting the results of a project requested by the Minnesota Vikings, they literally asked me if I would like to work for their organization full time, which I accepted.
Football Outsiders: How do you use your video system to complement statistical analysis in improving the Packers?
Eayrs: Databases and digital video are the two fundamental building blocks for any game analysis program. The strength of a computer database is the ability to sort and analyze categorical and numerical information. I always like to say the database answers the questions of "WHO", "WHAT", and "WHEN". You also locate those appropriate plays on the video system and view them, which provides answers to the "WHY" and "HOW" questions. By identifying valid information in the database and viewing the video image of the plays you have a valid and complete answer to questions about whatever player or team you wish to study.
Football Outsiders: Tell us a little about some of the new statistics you've created using the video system and analysis, such as "dominating blocks."
Eayrs: Well, I can't talk specifically about new statistics, which are proprietary information. We do a great deal of research to determine valid measures that contribute to winning. We then determine what an individual's contribution realistically needs to be within the team framework for us to be successful -- a context measure kind of like your VOA. The goal of winning the game is broke down by what would probably be required by individual and collective performances. Some of our measures have correlated 100% with winning games during the past four seasons with the Packers.
Football Outsiders: Does your job also include work on draft prospects? If so, how do you use analysis to help the Packers make better draft choices?
Eayrs: Yes, we have done historical and descriptive research on the college draft. One example is trade research -- what are the values for our selection numbers in terms of locating great players (especially by position), when might it be advisable to trade up or down, and what would be a good value for trading that pick.
Football Outsiders: As research director, what are your responsibilities on game day (as opposed to during the week preparing for the next game)?
Eayrs: During the week I prepare a profile on our opponent based on what they have done versus similar teams and situations. During the game I chart opponent calls to determine if we are seeing what we expected, or are we seeing something different, and point that out to the coaches. If we face a situation with a significant probability trend (offensive play or defensive call) I also remind the coaches and players of that information.
Football Outsiders: Do you think that most other NFL teams use statistical analysis as much as the Packers? Do you regularly discuss ideas with other coaches and research directors around the league?
Eayrs: Everyone is interested in information that can contribute to winning. People are reluctant to share sensitive competitive information, and close fraternization between staffs is discouraged by the Commissioner's office, but on occasion we have some great dialogues about football.
Football Outsiders: Why do you think this kind of research has received so little attention in football compared to baseball?
Eayrs: I have no idea, but the minimal attention is fine with me.
Football Outsiders: How has working for the Packers been different from working for the Vikings?
Eayrs: The two teams are very similar in the weekly routine, practices, travel, and games. They are very different in tradition, ownership, and organizational philosophy. I have greatly enjoyed working for each organization. They are both fantastic places to work.
Football Outsiders: Our research shows that the Packer defense took a big step forward with the acquisition of Grady Jackson. Does that statement agree with the research you've done? If so, why do you think the addition of Jackson had such an impact on the Packer defense?
Eayrs: The acquisition of Grady Jackson last season was very positive -- his quickness and size created problems for opponents on both running and passing plays. Michael Hawthorne, Larry Smith, and Curtis Fuller were also post-training camp acquisitions that made a positive impact, plus several other players had great games during the last part of the season. Your work on DVOA illustrated this as well.
Football Outsiders: Finally, do you ever have trouble tracking video because Al Harris and Mike McKenzie's hair gets in the way?
Eayrs: It hasn't been a problem to date!
Thanks to Mike Eayrs for the interview. If you are a Green Bay fan checking out the site for the first time, you might be particularly interested in a couple of recent articles: Michael David Smith's timeline of the best quarterbacks of each half-decade (featuring a couple of Packer greats) and the newest upgrade to our innovative DVOA statistics which rate the Packers as the best NFC team of 2003.