From Football Study Hall, here’s a win probability chart from Saturday night, as the game progressed.
/cdn3.vox-cdn.com/uploads/chorus_asset/file/2463156/UGAAUwp.0.png)
So, for all intents and purposes, Auburn was dead well before the fourth quarter got underway. Yep, that’s about right.
Sometimes, a picture is worth more than a thousand words.
Filed under Auburn's Cast of Thousands, Georgia Football, Stats Geek!
Get The Picture 
You want to see a steep line on a graph: do this chart for the Bama-AU game last year.
LikeLike
I gotta admit though, even after that last score, I remember looking at the clock and there was like 4:47 left or something like that, I turned to my buddy and said “I know it’s not even mathematically possible for them to come back, yet for some reason I’m still worried”.
I had dinner with one of my other friends last night, he wasn’t able to go to the game, and he asked how the crowd was. I said I’d rate it less than LSU last year and definitely nowhere near the first blackout game, but still a good, loud crowd. I described it to him as an “apprehensively excited” crowded…….it seemed like a lot of people were like me, cheering and enjoying the lead as it grew, yet in the back of our minds fearing that the light was about to flip on for the AU offense at any moment. I’m still in shock that we held them to 7 pts. Just a fantastic performance.
LikeLike
*crowd, not crowded.
LikeLike
I wonder if it was the fact that people were so bundled up? I Had a knit cap and hood on and it wasn’t until I watched the replay that I got the force of the crowd.
LikeLike
I was thinking exactly that. I wouldn’t say the crowd was subdued, but it didn’t seem as loud as it should have for a night game. I think it had to do with my knit cap and hood. Although, my section (a UGA section at that !) was full of Barners so it never felt that crazy.
LikeLike
All of the soft, thick clothing everyone had on had the same effect as curtians in your home or soft ceiling panels in your office…they absorb noise and soften it…a lot.
LikeLike
Yeah makes sense. And I wasn’t trying to say that the crowd was bad by any means, the crowd had great energy I thought. But I remember several times during LSU last year I was having to put my fingers in my ears, it was just so loud, and I wouldn’t say the crowd was ever quite that loud last night. But again, wasn’t trying to say that it was a bad crowd at all, it was a great game to be at all the way around.
LikeLike
I thought that the early penalties really dialed back the crowd enthusiasm.
LikeLike
Like a flat-lining EKG…
LikeLike
That is accurate….from an AU perspective.
LikeLike
They need to run the chart for the UF vs USC game.
LikeLike
believe I saw on ESPN that USC had the lowest chance to win of any team to win this season. .05% or something like that.
LikeLike
I’m impressed that none of the first five comments was WELL WHY WAS GURLEY STILL IN THERE??
LikeLike
Because the folks who usually brag about their 20/20, hindsight brilliance and yell “Fire Bobo”, or “Fire Richt” are taking a few days off….again. Rallying the revolutionary troops,no doubt, in case we win by less than 50 this week. But I did see some competitive boards, AU, raising that flag. I understand them wanting to get rid of those two, UGA is 9-5 against the Tigers with some very lopsided wins.
LikeLike
It’s beautiful!
LikeLike
After the Mizzou win I posted that I thought that was the UGA coaching staff’s finest hour. I now believe this win over Auburn was the UGA coaching staff’s finest hour.
LikeLike
Related fun fact: you’ve got to go all the way back to 2009 to find the last time Auburn scored in the 2nd quarter or later of a game in Athens.
LikeLike
Sucks that Gurley got hurt in the 4th.
LikeLike
That chart is nuts.
I don’t know what the machinery is behind those charts, but their estimates are wayyyy too “present biased.” By that I mean, it seems like the win probability at time t is way too influenced by events that occurred at times t-1 and t-2.
Does anyone think that UGA’s probability of winning was 25% because of being down by 7? Or that their probability of winning was north of 80% because of a 10-point halftime lead?
My guess is that these plots don’t take into account the fact that game strategies adjust based on the score. Offenses gamble more. Defenses get more conservative. Your win probability at time t affects your behavior at time t+1.
This chart’s model seems to assume that the underlying game process is the same in minute 5 as it is in minute 55.
Or maybe it’s because, as a UGA fan, I’ve been conditioned to expect heartbreak until the final whistle.
LikeLike