Data Science

Data science says the winner of Super Bowl LIV will be…

By Travis Henry for Galvanize

Data science, at its core, could be described as extracting useful information from available data to make an informed decision or prediction.

Las Vegas bookmakers have been doing it for decades.

In years past, collecting that data included reading newspapers, talking to people, and watching games on television. Before a big game, bookies would collect as much information as they could and then set the odds on who would win, who would lose, and by how much.  

But in 2020, we live in a predictive analytics world. And to no surprise, Vegas is right in the thick of it.

Casinos use predictive analytics to set betting lines and big data to support new marketing campaigns targeted at customers beyond gaming. Data scientists are now on staff and work directly with the bookmakers.  And the data being collected is enormous.

“The breadth of data in sports now allows data scientists to do more things,” according to sports fan, data scientist, and NuView Analytics founder Nick Newby. “Every single play is tracked.”

There are pros and cons to using data science to pick which team will win the next big game. Pros include predictions not swayed by emotion. Cons include predictions that can’t weigh in momentum.  An example of that, according to Newby, is the NFL’s Tennessee Titans’ recent two upsets to reach the AFC Championship game.

“You can’t capture momentum with a number,” Newby said. “There is an intangible you can’t measure with Derrick Henry running for 180 yards in each of the last three games. Those numbers would have been washed out in a season of data.”

On the flip side, however, data science can weigh in reality. Data will remember that the Titans lost to the Denver Broncos earlier in the season.

Recently, Galvanize asked Newby to model out the possible Super Bowl scenarios based on data science, but not until he gave us a head/heart prediction.

“My heart says the 49ers are going to win, my head says that it’s going to be the Chiefs over Packers,” Newby said. “My heart is with San Fran.”

For his prediction based on data science, Newby modeled the NFC Championship Game, the AFC Championship Game, and possible Super Bowl scenarios.

To create the models Newby said he would look at the different games each team won and lost and especially how they performed against similar teams to their upcoming opponent. 

Below are the results:

Predicting the 2020 Superbowl through Data Science

AFC Championship Game

Jan. 19, 2:05 p.m.
Kansas City Chiefs vs. Tennessee Titans
Arrowhead Stadium
Kansas City, Missouri

Chance of winning based on Newby’s model

Kansas City – 55% Tennessee – 45%

Vegas Line

Vegas odds suggest that Tennessee only has a 26% chance of winning.  This means that they are not giving enough credit to Tennessee or as much credit as our model thinks they deserve.

While we aren’t advocating for betting, it’s curious to compare the data science numbers against the chances Vegas are giving the teams.

“The best Vegas bet (in the AFC Championship) is Tennessee since they have a better chance than Vegas thinks,” Newby says. “If you are just trying to pick a winner  it’s Kansas City over Tennessee.”  

Data Science predicts fantasy football

NFC Championship Game

Jan. 19, 4:40 p.m. 
San Francisco 49ers vs. Green Bay Packers
Levi’s Stadium
Santa Clara, California

Chance of winning based on Newby’s model

San Francisco – 60%  Green Bay – 40 %

Vegas Line

Vegas odds for Green Bay are the same as Tennessee, 26%.  Meaning they feel Green Bay has as good a chance to beat San Francisco as Tennessee does of beating Kansas City.  Our model thinks both have a better shot of winning.

“Again Vegas has better odds on the underdog, so the best Vegas bet (for the NFC Championship game) is on Green Bay,” Newby says. “The best pick to win is San Francisco.” 

Super Bowl LIV

February 2, 2020 4:30 p.m.
Hard Rock Stadium
Miami Gardens, FL

Chance of winning based on Newby’s model

Kansas City – 55%  San Francisco – 45% 

“We won’t know the Vegas odds until the next two games are played, but the team with the best chance to win is Kansas City,” Newby said.


Interested in learning predictive analytics, or just how to get a leg up in your fantasy league? Take a look at Galvanize’s Data Science bootcamp and start learning today.

0 comments on “Data science says the winner of Super Bowl LIV will be…

Leave a Reply

Your email address will not be published. Required fields are marked *