Surprise! ESPN drops us in the FPI for 2023 after finishing 5th last year.

[AstonRoyal]

Yes, no surprise that ESPN is showing their love for us as usual.

Where Tennessee is ranked in ESPN's College Football Power Index for 2023 season

Vols come in at #12
View attachment 547082

8-4 prediction after winning 11-2 last year. Clearly they are not exactly sold on Milton taking the reins just yet. Hope we prove them wrong.

That loss to the cocks last year really screwed us. I hope this year we put a score on them so bad that 50 years from now, old timers who were there will still fall hushed when they mention that game.

Didn't this model expect us to go 8-5 last year?

 

[AstonRoyal]

Didn't this model expect us to go 8-5 last year?

Oh wait they didn't even have us ranking in their preseason top 25 last year haha

 

[Savannahbayvol]

Yes of course they hate us...that's why they're paying us 300 million a year for 10 years. I can actually FEEL the hatred coming through!

Let's see, the Orange Bowl drew 8.6 million viewers to ESPN. Guess that's why they hate us.

 

[Boston Vol]

Oh wait they didn't even have us ranking in their preseason top 25 last year hahaView attachment 547252

I don’t think that they had UT in the top 25 either, That’s the pre-season AP poll, not ESPN FPI.

 

[DawgInterloper]

I am very well aware of what it is and how it works. I am sure you know however that humans absolutely can skew AI results in multiple ways if they intend to.

Do I think ESPN wakes up every day saying piss on UT? no, but I do think their models favor certain things over others (weak ass conferences for example), that do not provide an accurate reflection of true strength between teams.

In my OP, I was having a bit of fun (besides, I said they love us!) but there is no denying they forecast 4 losses and with that schedule, there is no way to arrive at anything less than 9-3 based what is known and with recent history without bias of some sort.

Again, the only reason I can imagine we are where we are at is because of USCjr - otherwise, our loss to GA and wins over UAB, LSU and Clemson could not in any way come out to a #12 without an error.

My understanding is that Bill Connelly's formula incorporates roster talent (using prior recruiting classes as a proxy, not current roster), returning production with various weightings for each position, prior years' records (I believe three years with prorated weightings for earlier years), and probably some other inputs I don't know about. I don't think it considers coaching changes, but could be wrong.

Among those inputs listed above, here are the ones that potentially push Tennessee's FPI rating out of the top 10; prior season records (maybe 2020 -2022), recruit rankings for the last 4 years (I believe). Possibly also returning production but to a lesser extent than the others (most weighty negative being loss of starting QB).

I'm pretty sure Connelly does tweak his formula year to year as outcomes and data from the latest season are added are added to his model and tested. I'm not a fan of his model, way too simplistic IMO, but I don't think he tweaks it for goal seeking purposes other than the goal of improving its predictive value.

 

[TheVolsFrog]

247
SEC football predictions: Projecting every team's final record for 2023 season

 

[AstonRoyal]

I don’t think that they had UT in the top 25 either, That’s the pre-season AP poll, not ESPN FPI.

Oh whoops, just kidding

 

[BigOrangeTrain]

Didn't this model expect us to go 8-5 last year?

I think 7-6