Kirby Smart details difference balance between probability, reality in analytics
Analytics have been a part of football as long as people have tracked third down tendencies or broke down game film. But as more modern analytics have taken hold in sports, it’s on coaches to marry advanced modeling recommendations and their gut feeling on what their team can do — and Georgia head coach Kirby Smart seems to know this.
Georgia football surely deploys analytics in myriad ways, but Smart explained that whatever models or information are at their disposal are to inform decision making in the moment. They can tell him what is more or less likely to happen if he makes a certain call, but Smart is the one tasked with factoring in the human element and owning what happens next.
“I don’t control that stuff,” Smart said. “Probability is all that stuff that’s out there, the noise, the people that think that a teams not — why is Michigan favored by 20-something and then in a tight ball game? Why is Ohio State favored by whatever they’re favored by? Why is Florida favored by whatever they’re favored by? Why is UCF? Why is Tennessee favored by whatever, then it changes?”
CLICK HERE to subscribe for FREE to the On3 YouTube channel
With point spreads and Vegas — the thing Smart was mainly calling out above — the figure is the result of a multi-variable calculation involving how the teams play on the field but also what will entice bettors to get in on that game.
“It’s not reality,” Smart said. “It’s what some person in Vegas or some person sitting in this room thinks.”
Top 10
- 1New
CFB Hall of Fame
2025 class announced
- 2Hot
Lou Holtz
ND legend jabs Ryan Day
- 3
CFP beer prices
Concession prices for title game
- 4Trending
Carson Beck
UGA myth dead, Bama pursued
- 5
Quinn Ewers
Texas QB declares for NFL
Get the On3 Top 10 to your inbox every morning
By clicking "Subscribe to Newsletter", I agree to On3's Privacy Notice, Terms, and use of my personal information described therein.
As for probabilities themselves, those are reality, but any notion that they’re a flawless predictive tool is, itself, a flawed notion. The models — they vary as much as the people making them — usually build in past results and statistics plus other quantifiable or absolute factors like the weather or whether a game is home or away.
One can make a rather effective model showing how various decisions and their outcomes affect win probability. Smart still knows he has to live with it if he makes a smart analytical decision that falls apart. While his argument at times strayed the way of making a straw man out of analytics, there are plenty of people willing to lambast him on the worn-out ‘analytics are the boogeyman’ trope, no matter what Smart thinks.
At the end of the day, he only cares about what happens from snap to whistle.
“Probability doesn’t equal reality,” Smart said. “Reality is what you do on the field. And you control that by you prepare. So I’m a lot more worried about preparation than I am probability.”