Examining Hat Guy’s Tennessee-Georgia pick

Tennessee opened as an 8- to 11.5-point underdog to Georgia this week, with most being around the -8.5 mark. As I’m writing this, the line has settled to Georgia -7.5 to -8.5. Hat Guy’s un-eyeballed numbers had Georgia as a 16-point favorite. Let’s take a closer look at how he got there, why we should probably call in the replay officials, and what the pick should be.

First, how he got there:

From the perspective of Tennessee

Tennessee’s points

The Georgia scoring defense of 10.5 is most similar to the following prior FBS Tennessee opponents this year:

  • Alabama 16.6
  • Kentucky 19.9

Tennessee scored 52 points against Alabama and 44 points against Kentucky, a combined 263% of what those teams usually give up. Applying that to Georgia’s scoring defense of 10.5 results in a prediction of 27.6 points for the Vols.

Georgia’s points

The Georgia scoring offense of 41.8 is most similar to the following prior FBS Tennessee opponents this year:

  • Alabama 43.1
  • LSU 35.1

Tennessee gave up 49 points to Alabama but only 13 to LSU. Taken together, that’s 79% of what those teams usually score. Applying that to Georgia’s scoring offense of 41.8 results in a prediction of 33 points for Georgia against the Vols.

Estimated score: Tennessee 27.6, Georgia 33

From the perspective of Georgia

Georgia’s points

Tennessee’s scoring defense of 21 is most similar to the following prior Georgia FBS opponents this year:

  • Missouri 21.5
  • South Carolina 24.6

Georgia scored 26 points against Missouri and 48 points against South Carolina. Taken together, that’s 160% of what those teams usually give up, which when applied to the Vols results in a prediction of 33.6 points for the Bulldogs.

Tennessee’s points

The Tennessee scoring offense of 49.4 is most similar to the following prior Georgia FBS opponents this year:

  • Oregon 42.4
  • South Carolina 30.3

Georgia allowed only 3 points to Oregon and only 7 to South Carolina, a frightening 14% of what those teams usually put on the board. Applied to Tennessee, that results in a mere 6.9 points for the Vols against the Dawgs.

Estimated score: Georgia 33.6, Tennessee 6.9

Combined Estimated Score

If you put those two score predictions in a blender and mix ’em up, you end up with this:

Georgia 33.3, Tennessee 17.3 (Georgia -16)

Difference between Hat Guy’s prediction and the Vegas opening spread (UGA -8.5): 7.5

That makes this one an uneasy one for Hat Guy, meaning he’s only kinda confident about his conclusion that Georgia will cover at -8.5 points.

Eyeball adjustments

I honestly can’t tell whether its my eyeballs, my heart, my mind, or my hunger that object so strongly to the idea that this Tennessee offense is going to score only 7 points. Or 17. Or even 28, although that is at least approaching reasonableness. So let’s look closer at those numbers.

First, the analysis from Georgia’s perspective predicts only 6.9 points for the Vols. Why? Because they held an Oregon offense now averaging 42 points to a single field goal and a South Carolina offense now averaging 30 points to a single touchdown. As Hat Guy points out, the Georgia defense limited the two best offensive comps to 14% of their usual capacity. But if you don’t limit the comps to two games and instead include all games, that 14% number goes up to 41%. That’s still good, but it’s not a bankruptcy sale. Applied to Tennessee, that would mean a little over 20 points for the Vols, not 7. It would also change the combined estimated points for the Vols from 17.3 to 24. So call it Georgia 33, Vols 24 (UGA -9).

But what about that prediction of 27.6 points for the Vols from their own perspective? Is that correct? The math is right, and Tennessee’s offense scoring 263% of what the two next-best defenses usually give up sounds impressive. But when it’s applied to Georgia’s current average points allowed per game of 10.5, it still only adds up to 27.6. The fewest points Tennessee has scored this season is 34 against Pitt in Week 2. Alabama’s defense was allowing only 12.5 points per game when they played the Vols, and the Vols put 52 on them. Kentucky’s defense was allowing only 16.4 when they ran into the Heuper Drive and gave up 44. So forget the percentages and just look at what Tennessee did against really, really good defenses (hopefully) not that much different than Georgia’s: They averaged 48 points. That would make the combined points prediction for Tennessee 36 instead of 24. Even if you scraped another 10 points off the top just because it’s Georgia, you’d get a combined points prediction for the Vols of 31, three points less than any game they’ve played so far this season.

I think Tennessee’s offense has exposed disembodied Hat Guy’s Achilles Heel, namely applying percentages to super low numbers. I do think Georgia gets 33 points, give or take, and Tennessee ends up somewhere between 31 and 38. My mind says Georgia 33, Tennessee 31. Everything else in me is saying, “Nope.” Final answer: Tennessee 38, Georgia 35.

Other predictions from other systems

Vegas mostly had Georgia as an 8.5-point favorite when the lines opened this week. With an over/under of 65-66, that translates to something like Georgia 37, Tennessee 29.

Bill Connelly’s SP+ likes Georgia by 9.2 (Bulldogs 33.5, Vols 24.3) and gives the Vols a mere 14% chance of winning.

Summary

  • Vegas: Georgia 37, Tennessee 29 (Bulldogs -8.5)
  • SP+: Georgia 33.5, Tennessee 24.3 (Georgia covers)
  • Hat Guy: Georgia 33.3, Tennessee 17.3 (Georgia covers easily)
  • Me: Tennessee 38, Georgia 35 (Vols win a close one)

What do y’all think?

Comments are closed.