Super Bowl Spreads — Whose Line Is It Anyway?February 02, 2016
Las Vegas seems to be loving the Panthers. Following a fairytale season blemished with just one loss, Cam Newton and his team are reaping the majority of wagers from the many who anticipate yet another Panthers’ victory in Super Bowl 50.
The Westgate SuperBook famously opened with the Panthers as 4-point favorites over the Broncos. Since then, the Panthers have grown into 5.5-point favorites – a number that is prone to change, as we’re all too aware.
While watching the spreads, our team wanted to know if there was a better way to predict the rise and fall of the betting lines, especially since advanced analytics is becoming more and more a part of sports betting. And after discovering that 47 MILLION people bet on the Super Bowl (that’s nearly 3x the number of people who attended every NFL game this season, according to the American Gaming Association), we couldn’t help but dig deeper.
To no one’s surprise, we “Zoomphed out” and used our platform to find a relationship between a team’s social media activity and Las Vegas line movements.
The more that people are talking about a team on social media, the more that people are placing bets on that team. This should then forecast spread movement against that team up until a certain point, since bookies are less interested in exploiting their actual predictions for the game’s outcome, and more interested in finding the “sweet spot” of greatest profitability.
To test this theory, we created a consistent tracking methodology in which we listened for mentions of each team as “keywords” in Twitter conversations, or as direct mentions of their Twitter handles.
We then compared daily social media data against a 28-day benchmark average, which was calculated from the 28 days before and through the conference championships (Dec. 28 – Jan.24) for each team. Every day afterwards, we pulled in data from 12:00 AM one day to 11:59 PM the same day.
The final step was to compare each team’s performance against the other’s. We looked closely at total social activity for precise measurement of the volume of chatter about each team. Other KPIs, such as ZPoints™, corroborated our findings — but for simplicity’s sake, we decided to focus on this single KPI here.
The following formula gave us a better side-by-side comparison of the teams’ performances:
Team 1’s percentage change – Team 2’s percentage change
[(Panthers daily – Panthers benchmark)/Panthers benchmark] – [(Broncos daily – Broncos benchmark)/Broncos benchmark]
1. On Jan. 25th, Panthers had 30% more overall activity. Westgate line went up .5.
2. On Jan. 27th, Panthers had 9% more activity. Westgate line moved up 0.5 twice for a total of + 1 point.
3. On all days except for
one two, the Panthers had overall higher total activity than the Broncos; in aggregate, Panthers generated much more social activity around their team.
4. The line has stabilized as the difference between Panthers and Broncos percentage changes has started to lessen, and the bookies seemed to have reached a comfortable point whereby they’re profiting from both sides.**
As mentioned in our blog about the Iowa caucus, there are many other metrics that we could add here to help refine and hone accuracy. But after tracking just two of our most objective metrics, our team has indeed found a correlation between social media activity and line movements.
Specifically—there appears to be a positive correlation between activity around the Panthers and line movement.
We expect to see much more activity in both the betting and social media realms as game day draws closer. We in no way encourage you to bet based on our early hypothesis, but it will be interesting to see if this trend continues.
You can bet that we’ll be nerding out as the Super Bowl is nearing and will be the first to get online if another development arises. Stay tuned.
Within the same hour of publishing this blog, the Westgate line moved up another .5 points. Charts will be updated after we are able to pull in 24 hours worth of data and accurately compare new numbers against existing ones.
The above images have been updated to reflect any changes from yesterday. New data seems to support our theory, though the difference between the teams’ social media performances is very minimal (<1% when comparing original numbers that are not rounded off).
As per new data collected yesterday, the line has reverted to 5.5 — back to what it was when this blog was first published. Note that the Panthers’ overall activity increased by 1% from the previous day, whereas the Broncos’ activity dropped by 5%. This leaves us questioning if the bookies brought the line down to compensate for a growing gap, which was influenced by the different trajectories in overall chatter/social activity (up for Panthers, down for Broncos) around each team.
Updated data support the theory that the Panthers have a greater impact on upward line movements; unless the Panthers’ experience a notable increase in social activity, the line does not move up.