Statistics and Sports: “Deflategate”

On September 3, 2015, US District Court Judge Richard M. Berman granted the motion of Analysis Group’s client, the NFL Players Association (NFLPA), to vacate the four-game suspension of New England Patriots’ quarterback Tom Brady in the so-called “Deflategate” case.

The widely watched court case focused on the arbitration process that culminated in Mr. Brady's suspension following the use of allegedly under-inflated footballs during the 2015 AFC Championship Game between the Patriots and the Indianapolis Colts. In his decision, which was largely premised on legal deficiencies in the NFL's case, Judge Berman referenced the findings of Analysis Group academic affiliate Edward A. Snyder, Dean of the Yale School of Management.

Dean Snyder had analyzed the statistical evidence put forward by experts for the NFL and had presented his findings at an earlier appeal hearing before NFL Commissioner Roger Goodell. Supporting Dean Snyder in that analysis were Senior Lecturer Michael J. Moore of Northwestern University and an Analysis Group team led by Principal Jimmy Royer, as well as President Pierre Cremieux and Managing Principal Paul Greenberg. 

At the appeal hearing, Dean Snyder referred to the “impromptu protocols” used at halftime of the AFC Championship Game to measure the pressure of the footballs. He emphasized that the pressure measurements (in PSI) were taken sequentially, first for the relatively cold and wet Patriots’ footballs and then for the Colts’ footballs. Although this time difference was not accounted for in the NFL experts’ statistical models, it greatly affected the findings upon which the NFL’s conclusions were based. PSI increases the longer footballs remain in a warm and dry locker room, where the PSI measurements were taken, relative to cold and wet game-day field conditions outdoors.

Dean Snyder concluded that after properly accounting for timing, there was no statistical difference in the relative pressure drop of Patriots’ and Colts’ footballs, thereby negating the statistical underpinnings of the NFL’s case against Tom Brady. ■