Real-Time Predictive Analytics in Sports Broadcasting and Process Operations

As has been the tradition in recent years, ARC Forum held in Orlando began with a “pre-Forum” Super Bowl Party on Super Bowl Sunday.  Clearly, many folks at the party were intent on watching the game;   while others were mostly there to socialize and network.  Of the former, there appeared to be good representation by both New England Patriots fans and Atlanta Falcons fans.

The New England Patriots had already won four Super Bowls in the past, all with the same quarterback - Tom Brady who is now 39 years old.  The first time I saw Tom Brady close up was in 2002 when he was a rooky practicing at the team’s old training site at Bryant College in nearby Rhode Island.  I was watching the team practice with my daughter and husband and I recall the coach telling him to repeat an exercise.  I distinctly remember my daughter, who was 10 at the time, saying: “Mr. Brady you need to do it right this time.”   Back then, none of us could predict Brady’s meteoric career to become arguably the greatest quarterback of all time.

Moving ahead to the present, as I learned in discussions at the Forum, ESPN makes use of predictive analytics to determine a game’s outcome in real time during the broadcast.  Viewers can follow along with these predictions.  This particular Super Bowl game was very exciting, but for New England fans, the excitement really began to ramp up in the fourth quarter, when the team began an unbelievable and unprecedented comeback from a huge deficit. 

At the start of the game, ESPN’s predictive analytics (ESPN.com/Analytics) gave each team a 50 percent chance of winning.  However, by halftime and into the third quarter the New England Patriots trailed by 25 points.  At this point, the ESPN predictive analytics had adapted to show that the Atlanta Falcons had a 98.7 percent chance of winning.  However, the fourth q

Source: www.espn.com

Toward the end of the edge-of-our-seats fourth quarter, the analytics had changed to now predict that the New England Patriots had a 99 percent chance of winning.   The final score in this, the first-ever Super Bowl overtime was Patriots 34; Falcons 28.  For its real-time broadcast analytics, ESPN.com uses a team of experts and statistical tools and data. In this instance, no matter which side you were on, the predictive analytics appeared accurate at any given point in the game and adapted well to the (unpredictable) game changes.

As we learned during the course of this same ARC Forum, a variety of industrial organizations today also use predictive analytics to predict process behaviors in real time and adapt to those process changes.  Predictive and other advanced analytics are being used to prevent abnormal process behavior, avoid unnecessary asset maintenance, and more.  At the Forum, we saw a number of use cases for predictive analytics and other advanced analytics for both manufacturing and the supply chain.  We heard from diverse companies such as Abbott, ADM, Albermarle, Ascend Performance Materials, BASF, Agco, Evapracool, Johnson & Johnson and more.  Many presenters reported impressive benefits and ROI.

If you have not started using some of the new advanced software analytic tools available today for manufacturing, ARC believes that you should get started right away on this journey.  Remember that big data by itself doesn’t add much value to an organization unless it provides actionable intelligence and insights.

Stay tuned for ARC reports on this important topic over the next few months.

uarter was game changing in all senses of the expression.