Stats models and rankings

[slugboy]

Jackets are now #18 in Total Offense at 459.2 ypg. We've scored 33 offensive TD's. The only team left on our schedule that's ahead of us is UGAg. That'll be a tough game.

Defensively, well, we need mo offense!

Scoring offense is the good one, though

 

[ThatGuy]

If anyone had told me at the end of last season that we would improve from bottom quartile to top quartile in total offense I would not have believed them. Very impressive. Hope this continues.

Agreed. Our players have been awesome and have certainly done the work. But IMO, Buster Watson (that lowly "Quality Analyst" from uGA that so many on this board scoffed at), Chris Weinke, and Geep Wade - and Brent Key, for hiring them - deserve a TON of credit.

It's an amazing turnaround, and it places another brick in the wall that both X's and O's AND Jimmy's and Joe's are crucial to having a successful team.

 

[684Bee]

Agreed. Buster Watson (that lowly "Quality Analyst" from uGA that so many on this board scoffed at), Chris Weinke, and Geep Wade deserve a TON of credit. It's an amazing turnaround, and it places another brick in the wall that both X's and O's AND Jimmy's and Joe's are crucial to having a successful team.

Good ole Buster Watson. Not sure who that is, but give it up for him.

 

[slugboy]

College Football Geek has the best “straight up” forecasting percentage of the people that CFB Picker tracks

Here’s an EPA offense v defense chart (FBS only). We’re below average for the P5, to the left on the x axis. UVA is to our left.



And, the ACC highlighted

 

[slugboy]

FEI projects UVA over GT by 1.9 (25-23.1)

BCF Toys - 2023 Game Projections

www.bcftoys.com

SP+ picks us over UVA by 0.4.

CBC R2 is the best model against the spread this season according to CFB Picker. It has UVA 33.7 - GT 32.5

CFB Projections (Sam) | R2 Sports Metrics

www.r2sportsmetrics.com

It looks like a nail biter in Charlottesville

 

[stinger78]

On your O/D EPA graph, consider that 25 teams are in the Halp! quad and 22 are in the Good/Good quad. That's 47 of the 65 teams. So, their linear regression fit is pretty good. However, their categories might be debatable. If their axes off by just about 5%, GA Tech falls into the good O, bad D quadrant. I assume those axes represent the means for both?

I see the issue being more our Jekyll and Hyde act on O, not so much whether or not we're good or bad on O. You don't get to be #18 in total O, against a fairly tough schedule (overall W/L 32-16 FBS) by having a bad O. In 7 FBS games we've averaged 33 points in 3 wins, and 27 points in 4 losses. That's not a bad O considering that we left points on the field trying to catch up in a couple of the games we lost.

I have no issue with their metric, but maybe so with their buckets? I dunno.

 

[slugboy]

On your O/D EPA graph, consider that 25 teams are in the Halp! quad and 22 are in the Good/Good quad. That's 47 of the 65 teams. So, their linear regression fit is pretty good. However, their categories might be debatable. If their axes off by just about 5%, GA Tech falls into the good O, bad D quadrant. I assume those axes represent the means for both?

I see the issue being more our Jekyll and Hyde act on O, not so much whether or not we're good or bad on O. You don't get to be #18 in total O, against a fairly tough schedule (overall W/L 32-16 FBS) by having a bad O. In 7 FBS games we've averaged 33 points in 3 wins, and 27 points in 4 losses. That's not a bad O considering that we left points on the field trying to catch up in a couple of the games we lost.

I have no issue with their metric, but maybe so with their buckets? I dunno.

Either means or medians for P5 teams. If you’re in the bad/bad quadrant, you’re below average in both. Typically, you’d set that as the average, which makes the buckets easier.

Total offense can be misleading. Under Johnson we were murder on the offensive efficiency stats, but there were a lot of pass-happy teams with gaudier yardage numbers. Punting from the opposite 40 bulks your yards up, but it isn’t what you need. And sadly, if you give your opponents a lot of fast drives, you get more opportunities with the ball (I’ll have to check that stat for us).

We’re #47 in points per game, and we’re much worse in points per drive

Rec​	FBS​	NPD​		OPD​	Rk​
5-7​	4-7​	-1.11​		1.15​	125​

4-4​

3-4​

-.39​

2.43​

51​

We get a lot of shots, so we have great highlights, but we don’t make the most of our shots. We’re very inefficient

Edit: I must have grabbed last year’s stats for offensive points per drive. Should have caught that—the record is a tip off
This year’s numbers just added—we’re up a ton, but now in the middle

 

[yeti92]

Good ole Buster Watson. Not sure who that is, but give it up for him.

Maybe he is our version of Connor Stallions

 

[stinger78]

Either means or medians for P5 teams. If you’re in the bad/bad quadrant, you’re below average in both.

Yes, below average, but there's also a pretty wide spread, therefore a fairly large sigma. So, if you're close to the mean, it may not be quite so distinct as a line seems. Anyway, our O is on and off, so the bigger issue (IMPO) is which we are going to be for a particular game.

ETA: GT is #31 in O Overall EPA and Haynes King is #31 in Total EPA in their raw stats. Those are both top half.

 

[4shotB]

Yes, below average, but there's also a pretty wide spread, therefore a fairly large sigma. So, if you're close to the mean, it may not be quite so distinct as a line seems. Anyway, our O is on and off, so the bigger issue (IMPO) is which we are going to be for a particular game.

Or even in a particular quarter within a game (refer to L'ville and UNC).

 

[slugboy]

Yes, below average, but there's also a pretty wide spread, therefore a fairly large sigma. So, if you're close to the mean, it may not be quite so distinct as a line seems. Anyway, our O is on and off, so the bigger issue (IMPO) is which we are going to be for a particular game.

GT is #31 in O Overall EPA and Haynes King is #31 in Total EPA in their raw stats. Those are both top half.

Yeah, we’re really volatile on a game by game basis.

But, this is just P5, so being #31 in offensive EPA is going to be around average. Where did you see 31?

If you used this to power rank the ACC, it would look something like

FSU
Miami
Duke
UNC
Louisville
Clemson
BC
VT
NCST
Pitt
GT
Syracuse
UVA
Wake

 

[stinger78]

Yeah, we’re really volatile on a game by game basis.

But, this is just P5, so being #31 in offensive EPA is going to be around average. Where did you see 31?

If you used this to power rank the ACC, it would look something like

FSU
Miami
Duke
UNC
Louisville
Clemson
BC
VT
NCST
Pitt
GT
Syracuse
UVA
Wake

I sorted by Overall EPA and used my fingers and toes. Ha!

Good stats, though. Our inconsistency makes us hard to rank.

 

[iceeater1969]

My eye ball tests says when the opposing team has tough DT we have inconsistent on Offense. Mia, BC and Louisville.

Buster has been getting better at schemes- quick pass to flats to make DT run.

My eyes say
Uva has pretty good defense , but our OL has improved, our receivers are healthy, our qb is still tall , and last game he made the important rushing first downs - even w sore foot.

Qb s dad (Over 200 hs wins in Texas) who has good football eyes was at game and said - he" liked gt toughing it out".
Haynes said he likes winning but really really hates losing.

My stats add up we improve on D and get passing going and then run the ball in second half.

What do the stats say?

 

[roadkill]

FEI projects UVA over GT by 1.9 (25-23.1)

BCF Toys - 2023 Game Projections

www.bcftoys.com

SP+ picks us over UVA by 0.4.

CBC R2 is the best model against the spread this season according to CFB Picker. It has UVA 33.7 - GT 32.5

CFB Projections (Sam) | R2 Sports Metrics

www.r2sportsmetrics.com

It looks like a nail biter in Charlottesville

Early in the season most of us seemed ready to pencil in a W for UVA, but expected Syracuse to be a tough out. Based on the season thus far, I have those predictions flipped. Despite their W/L record, UVA has been in most of their games, have played good teams close, and are trending up. Syracuse on the other hand has seen their margin of victory steadily decline through the season and go negative as soon as they entered ACC play. Syracuse is winless in the ACC and their losses have been by multiple scores. Their average score in ACC games is in the single digits.

You might think we should be able to outscore UVA, but we've been so inconsistent that any specific game prediction is a crapshoot unless you know which GT team will show up.

 

[ThatGuy]

Good ole Buster Watson. Not sure who that is, but give it up for him.

God, I love autocorrect.

How "Faulkner" became "Watson" is a good question for the people who make the iPhone. Sheesh.

 

[ibeattetris]

This is the FBS v FBF epa diff from https://www.cfb-graphs.com/
Has us at rank 84 primarily from our atrocious D.

Offense definitely looking good.
Here is a PPD info

I'm guessing our offensive EPA is being pulled up by the fact that our offense is over performing in long drive situations. It's actually pretty insane that our OLD is higher than out overall OPD.

 

[CEB]

View attachment 15023
This is the FBS v FBF epa diff from https://www.cfb-graphs.com/
Has us at rank 84 primarily from our atrocious D.

Offense definitely looking good.
Here is a PPD info
View attachment 15024
I'm guessing our offensive EPA is being pulled up by the fact that our offense is over performing in long drive situations. It's actually pretty insane that our OLD is higher than out overall OPD.

I am not familiar with this.... Am I interpreting correctly that our "Short Drive Offense" is as bad as our defense? Thats kind of crazy... maybe we haven't had very many short field opportunities due to our poor D so short field missed opportunities have a bigger effect on us?

 

[leatherneckjacket]

I am not familiar with this.... Am I interpreting correctly that our "Short Drive Offense" is as bad as our defense? Thats kind of crazy... maybe we haven't had very many short field opportunities due to our poor D so short field missed opportunities have a bigger effect on us?

Yes, our short field offense is pretty poor, especially compared to every one else.

 

[CEB]

Yes, our short field offense is pretty poor, especially compared to every one else.

How many short field drives have we had compared to everyone else? I feel like we would be pretty poor in quantity as well…

 

[ibeattetris]

I am not familiar with this.... Am I interpreting correctly that our "Short Drive Offense" is as bad as our defense? Thats kind of crazy... maybe we haven't had very many short field opportunities due to our poor D so short field missed opportunities have a bigger effect on us?

Yeah. So bcftoys denotes a short drive as any drive "that begin less than 60 yards from the end zone". Sampling probably influences this a bit. For example, in the boston college game, we had 13 drives, but only one of them was "short". The Wake game had the most short yards. (11) FG, (38) Punt, (51) Down, (26) FG. So out of the 4 drives we got 6 points.

We had none against Miami. Having a defense that rarely forces a punt leads to our offense having a lot of drives in the longer areas.