First, can you tell readers a bit about yourself – how you got into sports statistics and sports management?
I got into sports statistics by accident. I was doing economic policy analysis around tobacco control issues at UCSF when I answered an ad for a part time consultant at a startup company call Protrade Sports (later Citizen Sports that was acquired by Yahoo!). They were a fantasy sports company that was looking to do advanced statistical analysis to create better scoring systems.
While there I met Roland Beech (currently with the Dallas Mavericks) and Aaron Schatz of Football Outsiders and learned about the budding field of sports statistics. From there I created the Journal of Quantitative Analysis in Sports and have worked with teams and other startups as well.
You’ve done a great deal of statistical analysis on football. Can you detail some of the bigger misperceptions people make about the game, revealed by your analysis?
One of the major results from my work is that in general, passing is actually less risky than running. The reason for this is that while a team is more likely to gain some yards running than passing, they are more likely to not gain enough yards to make the play a good play.
For example, on first down, if a play does not gain at least 4 yards, the team’s odds of scoring on that drive go down and there is a higher probability that a team will gain at least 4 yards if they pass than if they run.
When you present this data to coaches and players, how is it received?
I have had the pleasure of working with decision makers who are generally very interested in any information that can help them win more games so the fact that my information comes from complex math instead of hours of film study has not been an issue.
What do you think is the biggest gating in why football coaches don’t rely more on this kind of data when making on-field decisions?
I think it is incumbant on the analyst to convince the coach/decision maker that the value is there.
I think the common themes among analysts who have found success is that they have tremendous respect for the knowledge and intelligence of coaches/decision makers and they are able to communicate their analysis in a way that it is accessible, and actionable.
What have been some of the success stories of teams leveraging this analysis to make in-game decisions?
The biggest success I can think of is the Dallas Mavericks. They are the only NBA team that has an analyst (Roland Beech) on their coaching staff. Roland is one of the best in the business and his input was one (and just one) reason the Mavericks are currently world champions.
As the game changes to become a more pass-focused league, how has that changed your models and analysis and have you noticed new trends developing as a result of these changes?
My analysis has been focussed on the modern, more pass oriented era so there have not been major changes to results.
You recently developed the new QBR quarterback rating for ESPN. How do you think this had improved how quarterbacks are evaluated?
I was part of the team that developed the QBR with Dean Oliver and others at ESPN. The most fundamental advancement of QBR is that it measures the impact that the QB has on each play and does not try and decide the impact from yardage totals at the end of the game.
For example, in the NFL passer rating, a pass completion of 3 yards will lower a QBs rating, but if the completion was on 3rd and 2, that was a good play, and QBR recognizes that.
Where in the ranking do/can you factor for issues like strength of competition, offensive line quality – things that can’t be objectively evaluated but influence performance?
I have done a lot of work on the offensive line, and its impact can be measured. Part of what the QBR does is look at the value that was created on a play and then separate that value and give it to the various players (offensive line, receivers etc) who were involved. How much credit each deserves depends on the type of play and was determined objectively through analysis.
Do you feel there’s any negative aspect of the growing role statistical analysis plays in the NFL-fan experience? Why/why not?
I think it can only help because it brings another way for fans to engage with the game. If they are not interested, they can pretty easily ignore it, but for the groups that are interested, it brings new insights and perspective on the game that only make it more interesting.
Any last thoughts for readers?
The field of sports statistics is growing because it is another source of information that teams can use to gain a competitive advantage. Over the next few years, more and more teams will be implementing serious sports statistics programs which makes it an exciting field to be involved with.