FanPost

Overrated - the Follow-Up




I recently wrote an analysis of how the pre-season AP poll compares to the final poll and which teams actually tend to be the most overrated. Because Hoegher was insistent in the comments, I did some more work.

Background

Read my previous post for a longer explanation. The short story is I wanted to see how overrated teams actually are in the pre-season poll. I saw Chris Stassen's work on the subject and didn't like parts of his approach, so I set out to correct these issues:

  1. He didn't normalize for the number of times a school made the pre-season poll, so teams that appeared more frequently were punished.
  2. He compared the finish of a team to where it started. I felt that was unfair because #1 can't move up, only down while #25 has lots of room to improve. Any team that is normally ranked low gets an unfair benefit.
  3. He used a consensus pre-season ranking but the AP final poll.

Approach

The first thing I did was determine the average finish for a team for each pre-season rank. Using the same time period as Stassen (since 1989, when the poll went to 25), I calculated these averages. Then I compared where teams finished to where they were expected to finish based on those averages. I deviated from Stassen's work in one way. He used a consensus pre-season ranking while I stuck to just the AP. That way the same voters determine the starting and finishing spots, so their bias should be the same. It also made the work easier for me.

Interesting Results?

I think so. You really should read my previous post and decide for yourself.

So What's Hoegher's Beef?

One other problem I noted with Stassen's work is that any team that finished the year unranked was given a ranking of 26. While that makes things easier, it also skews the results because a team may have really been #40 instead. Fourteen places is a huge difference. This really is important for a low-ranked team. Say you start #25. You can improve to #1 (+24), but the worst you can do is drop to #26 (-1). That screws up the averages considerably.

If You Knew It Was a Problem, Why Didn't You Fix It Last Time?

Frankly, because it entailed a lot more work and I wanted to get a result. I also wanted to match Stassen's approach so that our results would be comparable. To fix it, I had to go back and look up 221 teams to see if they also received votes in the final poll. If so, I had to assign them their new final rank. If not, I had to move them below every team that did receive a vote.

So Was It Worth the Effort?

I think so. Let me know what you think.

New Average Finishes

Rank Average Change
1 6.54 -0.46
2 6.46 -0.38
3 12.25 -0.38
4 13.75 -0.50
5 10.29 -1.21
6 12.25 -0.58
7 14.00 -0.83
8 14.63 -1.42
9 18.79 -3.04
10 19.92 -1.83
11 18.33 -2.33
12 21.58 -3.88
13 22.42 -4.88
14 18.38 -3.25
15 27.04 -6.08
16 26.50 -5.25
17 21.71 -3.00
18 30.33 -7.46
19 25.71 -4.38
20 23.04 -4.58
21 21.67 -3.63
22 23.63 -4.58
23 30.00 -6.38
24 27.46 -5.54
25 32.17 -7.67

You can see the considerable change in the average finishes for teams low in the polls.
The best linear fit is y = 8.29 + 0.897x (vs y = 8.51 + 0.623x before). That makes a lot more since as the average finish decreases almost 1 position for each rank.

Anything Else of Interest?

Sure. Here are some fun facts:

  1. The lowest finish for any team was #43
  2. On average, 36 teams were ranked in the final AP poll
  3. On average, 9 teams would finish outside the top 25
  4. On average, 6 of those 9 would actually also finish without receiving votes
  5. The worst year was probably 2010, when 9 teams finished outside the top 39. 9 teams also missed in 2008 and 1998. 8 teams missed the top 40 in 2002.

So How Did the Schools Do?

The big result that I found last time was that the kings outperform most other schools, with the notable exception of Notre Dame. That held up after making these changes.

Basic stats:

68 schools were pre-season ranked at least once. 25 were ranked 10 or more times. I'm going to focus on those 25 since the larger sample size helps reduce the noise from one fluke season.

School Average
OSU 5.25
VT 4.94
AL 4.37
UGA 3.92
PSU 3.44
UF 3.13
Auburn 2.48
CO 1.87
MI 1.50
TN 1.41
FSU 1.41
Miami 1.40
OU 1.30
NE 0.76
UT 0.11
IA 0.01
UW 0.00
LSU -0.33
USC -1.26
Clemson -1.93
TAMU -2.28
WI -3.08
WV -4.19
ND -5.15
UCLA -9.03

Look at all those kings in positive territory. So despite what fans like to say, they really aren't overrated compared to other schools. On the other hand, everyone laugh at Notre Dame for being so bad (not so fast, Badger fans).

But What About My School?

Yeah, yeah, yeah. Here's the whole list.

School Average Count
Utah 13.38 2
HI 11.00 1
WSU 8.33 1
OSU 5.25 24
VT 4.94 16
AL 4.37 17
OR 4.08 9
UGA 3.92 16
UH 3.90 3
BSU 3.86 6
SC 3.67 4
PSU 3.44 18
TT 3.14 3
UF 3.13 21
KSU 3.07 9
Stanford 2.95 5
Auburn 2.48 16
BC 1.97 3
CO 1.87 12
TCU 1.85 8
Marshall 1.71 1
MI 1.50 20
UMD 1.44 3
TN 1.41 19
FSU 1.41 23
Miami 1.40 18
OU 1.30 20
UVA 1.21 8
NE 0.76 20
OkSU 0.69 4
NW 0.42 2
UT 0.11 19
IA 0.01 10
UW 0.00 12
UL -0.24 5
LSU -0.33 16
AR -0.54 7
BYU -0.58 8
PU -0.68 5
USC -1.26 21
Clemson -1.93 14
TAMU -2.28 13
MN -2.83 1
WI -3.08 15
SU -3.13 7
MSU -3.52 8
GT -3.76 6
WV -4.19 12
Fresno -4.54 1
ND -5.15 16
AZ -5.23 6
Cal -5.45 6
UNC -5.52 7
NCSU -5.63 3
Pitt -6.25 6
IL -6.73 5
MO -7.25 3
MS -7.52 2
S Miss -7.67 2
MS St -7.96 4
ASU -8.00 7
WF -8.00 1
UCLA -9.03 10
CSU -9.50 3
KU -12.00 3
USF -12.29 1
RU -12.50 1
OrSU -17.60 2

Anything else you want to know? Ask me.

Hoegher Wants Standard Deviation Info

Here it is.

Rank Average St. Dev.
1 6.54 7.71
2 6.46 7.31
3 12.25 11.04
4 13.75 10.74
5 10.29 10.33
6 12.25 9.42
7 14.00 10.12
8 14.63 11.96
9 18.79 12.95
10 19.92 11.53
11 18.33 11.61
12 21.58 12.40
13 22.42 14.42
14 18.38 13.15
15 27.04 12.05
16 26.50 10.95
17 21.71 11.19
18 30.33 11.04
19 25.71 10.76
20 23.04 13.26
21 21.67 11.71
22 23.63 12.76
23 30.00 9.44
24 27.46 10.54
25 32.17 8.26

I assume hoegher knows I'm just kidding with him. It was a valid point and we discussed it for a while. I always intended to do this part, but his interest spurred me to do it sooner.

X
Log In Sign Up

forgot?
Log In Sign Up

Please choose a new SB Nation username and password

As part of the new SB Nation launch, prior users will need to choose a permanent username, along with a new password.

Your username will be used to login to SB Nation going forward.

I already have a Vox Media account!

Verify Vox Media account

Please login to your Vox Media account. This account will be linked to your previously existing Eater account.

Please choose a new SB Nation username and password

As part of the new SB Nation launch, prior MT authors will need to choose a new username and password.

Your username will be used to login to SB Nation going forward.

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Off Tackle Empire

You must be a member of Off Tackle Empire to participate.

We have our own Community Guidelines at Off Tackle Empire. You should read them.

Join Off Tackle Empire

You must be a member of Off Tackle Empire to participate.

We have our own Community Guidelines at Off Tackle Empire. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9341_tracker