Digital Debunking: Can Data Analytics Software Predict the Results of The Big Game?
On Feb. 12, American football fans in the U.S. and around the world will gather to watch the biggest game the National Football League (NFL) has to offer, and one of the great spectacles in international sports. Thanks to NFL copyright restrictions, we can’t say the game’s official name here (nor can we use team or player names) but we can provide a fill-in-the-blank puzzle; let’s say fans will gather to watch the _uper _owl. From here on out, we’ll call it The Big Game.
After a grueling season full of heartbreak and joy, only two of the NFL’s 32 teams remain, hoping that after 18 long weeks, they will get to hoist the Lombardi Trophy. Only time will tell who will emerge victorious, but we decided to see if Altair’s data analytics software can predict the winner of The Big Game.
To investigate, we used Altair® Monarch®, Altair’s market-leading, desktop-based self-service data preparation solution. Monarch connects to multiple data sources – including structured and unstructured data, cloud-based data, and big data – to turn difficult data into smart data, creating actionable insights that solve complex problems.
Setting Up the Monarch Model
In this scenario, we used Monarch to perform descriptive analytics that profile the performances of prior winning quarterbacks (QBs) during the first 18 weeks of the NFL season, which we then used to predict which of the remaining two QBs will win The Big Game. Typically, a team’s QB performance is a reliable way to determine how well the team is playing during the season – as such, we decided to study a statistical metric called QB rating.
In a nutshell, QB rating incorporates five variables and weighs them on a scale. These variables include pass attempts, completions, passing yards, touchdown passes, and interceptions. Then, using a few mathematical formulas a little too detailed to mention here, it determines how well a QB played, assigning them a rating between 0-158.3. A higher number indicates a better performance; the NFL average QB rating is around 93.6.
For our investigation, we imported three datasets from Kaggle into Monarch:
- NFL Superbowl (puzzle hint!) History 1967-2020
- NFL Statistics_Basic_Stats
We then went to www.nfl.com to collect the regular season stats from the two Big Game teams’ respective QBs and imported this data into Monarch.
The final result is a table that focuses on select stats from prior winning quarterbacks and how the final four compare.
Monarch data table displaying quarterback performance statistics
Based on our findings, the QB from Kansas City (_atrick _ahomes, if you can solve another puzzle) appears to be the favorite since his performances have been comparable, if not better than, the performance of prior Big Game-winning quarterbacks. So our prediction is in – the team from Kansas City appears to have an edge. But odds are odds, and they can be defied. All that’s left now is for Big Game itself to roll around so we can see if our prediction is correct.
Disclaimer: This article doesn’t represent Altair’s official organizational prediction, of which there's none. The company’s official policy is not to take sides in The Big Game – Altair as an organization will be thrilled no matter if the winning team represents Kansas City or Philadelphia. We’re just here for the show!