The SEC and trends during the 2013 season

[MNIrishman]

So we've been discussing media bias for the SEC a lot on the board lately, and I wanted to analyze trends in rankings to see what happened week to week. The trend value I REALLY wanted was to look at how the rankings changed when two ranked teams from the same conference played each other, but unfortunately I couldn't find a good data set that included the schedules that way and I wasn't about to research and type up everything for the whole season, so my dataset was just the AP poll for 2013.

The first graph is the general trend of rankings by conference average. I had wanted to show that the SEC gradually moves up every week, but that wasn't the case. Really, this data is just noisy because of the low number of ranked teams each week. Some conferences would lose their lowest ranked team from the rankings, then their average rank would benefit a great deal:

So I decided to take a look at the number of ranked teams by conference:

The most interesting part about this was that conference strength isn't determined during the season. It's determined before the season ever starts. There's barely any change in the number of ranked teams each conference has from start to finish. What change there is, however, seems to validate the hypothesis that the SEC gets the benefit: They start with the greatest number of ranked teams and add one during the course of the year.

If anyone was curious...

PS: I know the y axis should be inverted. I didn't feel like messing with the code to do that.

 

[BGIF]

So we've been discussing media bias for the SEC a lot on the board lately, and I wanted to analyze trends in rankings to see what happened week to week. The trend value I REALLY wanted was to look at how the rankings changed when two ranked teams from the same conference played each other, but unfortunately I couldn't find a good data set that included the schedules that way and I wasn't about to research and type up everything for the whole season, so my dataset was just the AP poll for 2013.

The first graph is the general trend of rankings by conference average. I had wanted to show that the SEC gradually moves up every week, but that wasn't the case. Really, this data is just noisy because of the low number of ranked teams each week. Some conferences would lose their lowest ranked team from the rankings, then their average rank would benefit a great deal:

So I decided to take a look at the number of ranked teams by conference:

The most interesting part about this was that conference strength isn't determined during the season. It's determined before the season ever starts. There's barely any change in the number of ranked teams each conference has from start to finish. What change there is, however, seems to validate the hypothesis that the SEC gets the benefit: They start with the greatest number of ranked teams and add one during the course of the year.

If anyone was curious...

PS: I know the y axis should be inverted. I didn't feel like messing with the code to do that.

So, the SEC is the axis of evil.


I've known that since the SWC immolated itself.

 

[T Town Tommy]

So we've been discussing media bias for the SEC a lot on the board lately, and I wanted to analyze trends in rankings to see what happened week to week. The trend value I REALLY wanted was to look at how the rankings changed when two ranked teams from the same conference played each other, but unfortunately I couldn't find a good data set that included the schedules that way and I wasn't about to research and type up everything for the whole season, so my dataset was just the AP poll for 2013.

The first graph is the general trend of rankings by conference average. I had wanted to show that the SEC gradually moves up every week, but that wasn't the case. Really, this data is just noisy because of the low number of ranked teams each week. Some conferences would lose their lowest ranked team from the rankings, then their average rank would benefit a great deal:

So I decided to take a look at the number of ranked teams by conference:

The most interesting part about this was that conference strength isn't determined during the season. It's determined before the season ever starts. There's barely any change in the number of ranked teams each conference has from start to finish. What change there is, however, seems to validate the hypothesis that the SEC gets the benefit: They start with the greatest number of ranked teams and add one during the course of the year.

If anyone was curious...

PS: I know the y axis should be inverted. I didn't feel like messing with the code to do that.

So... the SEC gets the benefit of the doubt because their preseason conference strength had one less team than their end of the season conference strength based on ranked teams? I fail to see how that is a benefit of the doubt. To me, that would show that the conference was actually stronger - by one team - than their preseason predictions for conference strength.

But to look at the polls from this week... it appears the recepients of benefits of the doubts were Arizona and TCU. Arizona goes from being ranked 29th in the polls all the way to 10th with a win against Oregon. TCU - who beat Oklahoma goes from 25th to 9th. I am just reminded of all the posters out here claiming SEC bias when Miss State went from 28th to 14th in the polls a few weeks back.

Going back the last 20 years or so, it would be safe to say that there are a lot more non-SEC teams that have received preferential treatment in preseason polls and have failed to live up to the hype. The data is out there that breaks down which teams have failed to live up to those expectations, and I think some of the Irish faithful would be disappointed with how the Irish have performed relative to their preseason rankings during that timeframe. That benefit of the doubt bias is sometimes more truthful than we want to acknowledge.

 

[MNIrishman]

So... the SEC gets the benefit of the doubt because their preseason conference strength had one less team than their end of the season conference strength based on ranked teams? I fail to see how that is a benefit of the doubt. To me, that would show that the conference was actually stronger - by one team - than their preseason predictions for conference strength.

But to look at the polls from this week... it appears the recepients of benefits of the doubts were Arizona and TCU. Arizona goes from being ranked 29th in the polls all the way to 10th with a win against Oregon. TCU - who beat Oklahoma goes from 25th to 9th. I am just reminded of all the posters out here claiming SEC bias when Miss State went from 28th to 14th in the polls a few weeks back.

Going back the last 20 years or so, it would be safe to say that there are a lot more non-SEC teams that have received preferential treatment in preseason polls and have failed to live up to the hype. The data is out there that breaks down which teams have failed to live up to those expectations, and I think some of the Irish faithful would be disappointed with how the Irish have performed relative to their preseason rankings during that timeframe. That benefit of the doubt bias is sometimes more truthful than we want to acknowledge.

I think you're misinterpreting. The increased count of ranked teams only occurs as the SEC runs through the conference schedule. That would seem to imply that conference on conference games result in increased team valuation, whereas out of conference games go less well for the SEC. However, this overall correlation is relatively weak. What is most significant is that the media decides a priori how strong a conference is, then votes to validate that perception during the season.

 

[T Town Tommy]

I think you're misinterpreting. The increased count of ranked teams only occurs as the SEC runs through the conference schedule. That would seem to imply that conference on conference games result in increased team valuation, whereas out of conference games go less well for the SEC. However, this overall correlation is relatively weak. What is most significant is that the media decides a priori how strong a conference is, then votes to validate that perception during the season.

Then how do we determine if a conference is strong or weak? What is the measurable? The only thing I can think of that would quantify if the media thinks a conference is strong or weak would be how they perform in bowl games. If that's the only measurable, then the SEC has done very well in holding up what the media thinks.

The other measurable may be how each conference performs in the regular season against other conferences. Again, the SEC has done very well in that measurable as well.

 

[MNIrishman]

Then how do we determine if a conference is strong or weak? What is the measurable? The only thing I can think of that would quantify if the media thinks a conference is strong or weak would be how they perform in bowl games. If that's the only measurable, then the SEC has done very well in holding up what the media thinks.

The other measurable may be how each conference performs in the regular season against other conferences. Again, the SEC has done very well in that measurable as well.

0-2 in BCS games during the 2013 season. Losing record all time vs. Big East. Few meaningful out of conference games. Very few true road games OOC. The only teams I've been genuinely impressed by in an objective way of the SEC in the last few years were 2011 LSU and 2012 Bama.

 

[T Town Tommy]

0-2 in BCS games during the 2013 season. Losing record all time vs. Big East. Few meaningful out of conference games. Very few true road games OOC. The only teams I've been genuinely impressed by in an objective way of the SEC in the last few years were 2011 LSU and 2012 Bama.

That's the basis for the suppossed SEC bias? That's fairly weak IMO as I could throw out overall BCS record, BCS title game record, bowl records, records against top 25 opponents, etc all day long and show exactly how the SEC stacked up against the rest of the country. I think that would speak for itself.

Not trying to argue the point... I just fail to see where there is quantifiable data that suggests the SEC receives some sort of bias. The play on the field is what matters. If there is data out there that supports this preception, then someone should bring it forth. Not what we perceive or what we think... what can actually be shown and proven.

 

[NDWorld247]

Then how do we determine if a conference is strong or weak? What is the measurable? The only thing I can think of that would quantify if the media thinks a conference is strong or weak would be how they perform in bowl games. If that's the only measurable, then the SEC has done very well in holding up what the media thinks.

The other measurable may be how each conference performs in the regular season against other conferences. Again, the SEC has done very well in that measurable as well.

IMO, this is the best measurable as bowl games are, more often than not, meaningless exhibition played in SEC friendly locales. I've looked into this before, and posted the link below in another thread some time ago, and found the SEC hasn't performed as well as you Southerners like to think.

In the BCS era (1998-2011), the SEC has an overall record of 81-77 (.513) against Power 5 conferences in the regular season and 60-44 (.577) in bowl games. The SEC was 7-7 in regular season games in 2012 and 6-2 in bowl games and 12-6 in regular season games in 2013 and 7-3 in bowl games. I couldn't find records vs. specific conference in 2012 and 2013. Their overall BCS era record in the regular season is .526 (100-90).

Why SEC Isn't As Great In Football As You Think | ThePostGame

Conference records from 1998 through 2011 season:
SEC vs. PAC-12 regular season: 10-12
SEC vs. PAC-12 bowl games: 1-0
SEC vs. Big 12 regular season: 6-10
SEC vs. Big 12 bowl games: 21-8
SEC vs. ACC regular season: 42-36
SEC vs. ACC bowl games: 16-9
SEC vs. Big 10 regular season: 7-4
SEC vs. Big 10 bowl games: 19-19
SEC vs. Big East regular season: 16-15
SEC vs. Big East bowl game: 3-8

As for media bias...I believe it exists. It seems fairly obvious it exists. It's a tough thing to quantify, but I believe it's there.

 

[T Town Tommy]

IMO, this is the best measurable as bowl games are, more often than not, meaningless exhibition played in SEC friendly locales. I've looked into this before, and posted the link below in another thread some time ago, and found the SEC hasn't performed as well as you Southerners like to think.

In the BCS era (1998-2011), the SEC has an overall record of 81-77 (.513) against Power 5 conferences in the regular season and 60-44 (.577) in bowl games. The SEC was 7-7 in regular season games in 2012 and 6-2 in bowl games and 12-6 in regular season games in 2013 and 7-3 in bowl games. I couldn't find records vs. specific conference in 2012 and 2013. Their overall BCS era record in the regular season is .526 (100-90).

Why SEC Isn't As Great In Football As You Think | ThePostGame

Conference records from 1998 through 2011 season:
SEC vs. PAC-12 regular season: 10-12
SEC vs. PAC-12 bowl games: 1-0
SEC vs. Big 12 regular season: 6-10
SEC vs. Big 12 bowl games: 21-8
SEC vs. ACC regular season: 42-36
SEC vs. ACC bowl games: 16-9
SEC vs. Big 10 regular season: 7-4
SEC vs. Big 10 bowl games: 19-19
SEC vs. Big East regular season: 16-15
SEC vs. Big East bowl game: 3-8

As for media bias...I believe it exists. It seems fairly obvious it exists. It's a tough thing to quantify, but I believe it's there.

From 1998 till around 2004 I would agree the SEC was fairly average. But there is always the "other side" of the story that some don't want to acknowledge.

Since the suppossed start of SEC "dominance" from about 2004 on, what's their record against the Power Five? Pretty easy. It is miles away better than the other conferences. And, the SEC has played more games against Power Five opponents during that timeframe as well. The SEC has played 111 total regular-season games against Power 5 schools since 2004. Its 69-42 record is the best of the all Power 5 conferences, ahead of the Pac-12 (53-42), the Big 12 (42-42) and the Big Ten (36-45).

Over that time, the SEC has gone 42-23 against the ACC, 12-7 against the Pac-12, 9-8 against the Big 12 and 6-4 against the Big Ten.

Bowl games?

Since 2000, the SEC is 26 games above .500 in bowl games, which is a better win-loss differential than the ACC (minus-5), Big 12 (even), Big Ten (minus-23) and Pac-12 (plus-5).



As far as bias, preseason ranking vs where a team actually ended ranked. From 1989 to present the bottom teams were the following - the number represents total number of positions gained/lost in the polls during that timeframe:

9. Miami (-61)
8. Florida (-69)
7. Fla St (-76)
6. Notre Dame (-78)
5. Oklahoma (-87)
4. Michigan (-88)
3. Nebraska (-88)
2. Southern Cal (-95)
1. Texas (-112)

Conference overall- Power Five + Ind. (1993-present):

6. Pac 12 (+45)
5. Big 10 (-21)
4. SEC (-53)
3. Ind. (-61)
2. ACC (-108)
1. Big 12 (-185)

Based on this data, the SEC had one team that made the top nine teams as being overrated to start the season from 1989 to the present. Don't know if we can say there is an SEC bias when it appears the only team that consistently got "overrated" were the Gators.

On the conference side, it appears the Big 12 and ACC get a lot more bias when it comes to preseason rankings vs where they actually end up.


Alabama was +24 over the timeframe... meaning they finished a total of 24 positions higher over the timeframe than they were actually ranked to start the season.

SEC years in which the conference failed to deliver:

2009 (-24)
2008 (-18)
2004 (-19)
2002 (-20)
2000 (-36)
1995 (-15)


If anything, the data shows that recently the SEC is actually doing better than their preseason rankings are reflecting.

 

[Legacy]

I have some personal objections which these well-researched comments have brought up.

- Preseason rankings - and rankings up until after Week 7 - are best guesswork. Many teams do not play any significant games until they get to conference play. Arizona, both Mississippis, Alabama, Baylor and Texas A&M are examples.

-Week to week rankings in the first half of the season mostly benefit television ratings as does the expansion years ago to the Top 25. I would keep to the Top 15 at most, which would be more sound as a basis for chest-thumping and statistical comparisons in my opinion.

-That said, I have minimal objections with the changes from Week 6 to Week 7 in the AP. Week 7 results will be more indicative of performance and any bias. How far will the loser of Auburn-Miss State loser fall? If Ole Miss loses how far will they fall? Is TCU or Baylor the better team?

-As for current rankings after Week 6, I, too, would rank one loss Alabama and Michigan State over TCU and Arizona at this time. That's guesswork and probably bias against TCU and Arizona, I admit. But if TCU beats Baylor, an undefeated TCU should jump both Ala and Mich St and losers from higher ranked teams, which is where I believe they would start if rankings began after Week 7.

-However, one loss Oregon falling from 2 to 12, after having beaten Michigan State is a red flag. Alabama fell from 3 to 7 with the loss to Ole Miss (#11 last week). This seems to be more of an investment in the prior rankings (Arizona was 29th in voting). Two teams started the season unranked. One jumped to #3 by virtue over a win over the #6. One jumped to #10 by virtue of a win over #2. Which was the SEC team?

-Week 7 games - Will the one loss winner of Oregon-UCLA see a corresponding rise in the polls as Alabama past one loss SEC teams? If Texas A&M beats Ole Miss, will both and Alabama be ranked ahead of Oregon if they win?

-Bowl won-loss records and conference comparisons favor Sunbelt teams due to location and are skewed by contracted conference matchups. The ACC, for instance, has a much better bowl lineup than the Pac 12, which is a better conference. The SEC matches up against the Big Ten in many bowl games. Outside of BCS and now Playoff matchups, the SEC has one bowl game against the Big 12 and none against the Pac 12.

-To primarily play Big Ten and ACC teams in bowl games and then make statistical comparisons with subsequent "best conference" thumping is disingenuous. The SEC primarily plays the ACC in non-conference regular season play and the Big Ten and ACC in contracted bowl games.

 

[Me2SouthBend]

These records are possible when you sign 110 kids over the course of 4 years. It isn't a 4 year schollie, it's survival of the fittest, it's perform or be cut. Not a knock on you TTT. Just calling it as I see it.
Edit: 1 other thing, the way to remove some of the bias or some of the problem is to get rid of all preseason polls. First poll should come out in early Oct.

 

[wizards8507]

The way to remove some of the bias or some of the problem is to get rid of all preseason polls. First poll should come out in early Oct.

Who's going to stop the AP from publishing a poll? The NCAA doesn't control them. Rankings are a media construct that fans buy into, not officially sanctioned "standings" of the FBS. Even if the AP poll were delayed until October, then the networks would just start referencing the ESPN power rankings or some other such thing. The issue wouldn't be early polls, it would be whether those polls influence the committee.

 

[MNIrishman]

Who's going to stop the AP from publishing a poll? The NCAA doesn't control them. Rankings are a media construct that fans buy into, not officially sanctioned "standings" of the FBS. Even if the AP poll were delayed until October, then the networks would just start referencing the ESPN power rankings or some other such thing. The issue wouldn't be early polls, it would be whether those polls influence the committee.

Yeah I'm not sure there's a way to fix the issue. You're right that no matter what, there will always be preseason polls, and even if they don't release the important ones until week X, the influence will still be there. I just like to grumble about the media's influence on things before there's a shred of data in a system designed to reinforce baseless assumptions.

 

[IrishLion]

The SEC's non-conference schedules were a joke this year, but they are already riding their perception to the top anyway by beating each other up.

When Mississippi State and Ole Miss lose games later on but don't take the average drop in the poles, the perception will be that "the SEC is a gauntlet." The reality is that nobody in the SEC, save LSU and Georgia, has played a worthy OOC opponent, so we don't TRULY know how they stack up to the top teams from other conferences THIS YEAR.

The SEC has been phenomenal in bowl games recently, no doubt, and I do think it is the best conference in football. However, last year's postseason results proved that the SEC is not NEARLY as far ahead of the nation as their rankings fluctuations suggest.

 

[wizards8507]

When Mississippi State and Ole Miss lose games later on but don't take the average drop in the poles, the perception will be that "the SEC is a gauntlet." The reality is that nobody in the SEC, save LSU and Georgia, has played a worthy OOC opponent, so we don't TRULY know how they stack up to the top teams from other conferences THIS YEAR.

Thankfully there's a ton of matchups between the top-tier SEC teams to sort that all out between now and the postseason. Even if the voters/committee give a slight edge to SEC teams who lose to each other, I really don't think they'll have the cajones to start leapfrogging teams with objectively better records. I don't think a two-loss SEC team would be ranked ahead of a one-loss Notre Dame, for example (assuming the loss was to FSU and not a total embarrassment).

 

[ACamp1900]

When Mississippi State and Ole Miss lose games later on but don't take the average drop in the poles, the perception will be that "the SEC is a gauntlet." The reality is that nobody in the SEC, save LSU and Georgia, has played a worthy OOC opponent, so we don't TRULY know how they stack up to the top teams from other conferences THIS YEAR.

How is this any different from any random year??

 

[IrishLion]

How is this any different from any random year??

It's really not, but the point I was making is that previous years, perceptions, bowl records, etc. shouldn't matter. There should be no "benefit of the doubt" year to year. But there's nothing to be done because the SEC doesn't schedule huge OOC games too often, so we don't really know how they stack up on any given year until it's time for bowl games.

And what last year's bowl games taught us is that the SEC isn't nearly as far ahead as their rankings advantages suggest.

I really don't think they'll have the cajones to start leapfrogging teams with objectively better records. I don't think a two-loss SEC team would be ranked ahead of a one-loss Notre Dame, for example (assuming the loss was to FSU and not a total embarrassment).

I wish I shared this belief... but I fear that the overarching perception of the SEC would result in just such a situation.

I could definitely see the committee taking a two-loss Bama over a one-loss ND. The argument would be that Bama played a tougher schedule week-in-and-week-out than either ND or FSU, and thus the difference in records is moot.

 

[ACamp1900]

It's really not, but the point I was making is that previous years, perceptions, bowl records, etc. shouldn't matter. There should be no "benefit of the doubt" year to year. But there's nothing to be done because the SEC doesn't schedule huge OOC games too often, so we don't really know how they stack up on any given year until it's time for bowl games.

And what last year's bowl games taught us is that the SEC isn't nearly as far ahead as their rankings advantages suggest.

Bama did play PSU up in Happy Val. recently right??? Other than that i can't think of one top end SEC team that's left the south for any real game in, man, since I was a little kid... I'm sure it's happened here and there, but it must be nice to never have to leave your general locale to play really,... anything (Looking at you Florida)

 

[wizards8507]

It's really not, but the point I was making is that previous years, perceptions, bowl records, etc. shouldn't matter. There should be no "benefit of the doubt" year to year. But there's nothing to be done because the SEC doesn't schedule huge OOC games too often, so we don't really know how they stack up on any given year until it's time for bowl games.

And what last year's bowl games taught us is that the SEC isn't nearly as far ahead as their rankings advantages suggest.

I've expressed my dislike for the playoff but this is one area where the committee will hopefully be able to have a positive impact on the game. There's been an "arms race" of teams scrambling to schedule quality OOC opponents and that's a great thing (see Alabama opening against USC in 2016). Herbstreit and the GameDay guys ripped Baylor to shreds a couple of weeks ago for their OOC schedule, going so far as to say the committee should "punish" them for scheduling such shtty competition.

 

[IrishLion]

I've expressed my dislike for the playoff but this is one area where the committee will hopefully be able to have a positive impact on the game. There's been an "arms race" of teams scrambling to schedule quality OOC opponents and that's a great thing (see Alabama opening against USC in 2016). Herbstreit and the GameDay guys ripped Baylor to shreds a couple of weeks ago for their OOC schedule, going so far as to say the committee should "punish" them for scheduling such shtty competition.

The bolded is ironic because I believe Herbstreit has Baylor as one of his four playoff teams right now haha.

And you're right that the committee should have an impact, and every conference (though I've been singling out the SEC) will have to eventually prove themselves against top teams from other conferences during the regular season, and as you mentioned, the process has already begun, which is AWESOME for college football in general.

But I won't have faith in the committee to look at things without bias until we see their first bracket.

 

[irishog77]

Bama did play PSU up in Happy Val. recently right??? Other than that i can't think of one top end SEC team that's left the south for any real game in, man, since I was a little kid... I'm sure it's happened here and there, but it must be nice to never have to leave your general locale to play really,... anything (Looking at you Florida)

auburn and arkansas both played sc in '03-'06ish...and got rocked.

tennessee played ND.

That about sums it up.

 

[Me2SouthBend]

Who's going to stop the AP from publishing a poll? The NCAA doesn't control them. Rankings are a media construct that fans buy into, not officially sanctioned "standings" of the FBS. Even if the AP poll were delayed until October, then the networks would just start referencing the ESPN power rankings or some other such thing. The issue wouldn't be early polls, it would be whether those polls influence the committee.

No one is going to stop the AP from coming out w a poll, I didn't suggest the NCAA could control any such thing. I'm just stating that the preseason polls fuel the problem. If any polling that was going to be used to decide the 4 playoff teams wasn't released until the first weekend in Oct and it was a coaches poll or something of the like at least you'd get rid some of the posers that inevitably start the year ranked much too high. As a side effect, maybe it would force some of these schools that load up on cupcakes to start the season to change their scheduling model.

 

[IrishSteelhead]

SEC schools who have played in SB (last time in parentheses):

Tennessee (2005)
Texas A & M (2000)
LSU (1998)
Vanderbilt (1995)
Alabama (1987)
Ole Miss (1985)
South Carolina (1984)
Missouri (1978)

 

[Legacy]

SEC schools who have played in SB (last time in parentheses):

Tennessee (2005)
Texas A & M (2000)
LSU (1998)
Vanderbilt (1995)
Alabama (1987)
Ole Miss (1985)
South Carolina (1984)
Missouri (1978)

Tenn also played ND in 1978.

Notre Dame's record in the above games is 8-1. The one loss was to Missouri in 1978 (0-3). A&M and Missouri were not part of the SEC at the time we played them. So, that's a 7-0 record against SEC teams who come to SB. There will be a gap of twelve years between SEC teams who play in SB - Tenn (2005) and Georgia (2017). Texas A&M will be coming to SB in 2025.

Also of note, of these nine historical games, six were in late Oct or Nov. ND's Sept schedule fills up quickly, including Georgia and A&M in future schedules. Any SEC team who wants to play ND home and home in November can move to the front of the line.

 

[irishog77]

Nd has lost at least 2 games to tennessee in South Bend.

 

[IrishSteelhead]

Nd has lost at least 2 games to tennessee in South Bend.

Yeah I just included the last visits.

 

[Legacy]

Nd has lost at least 2 games to tennessee in South Bend.

Nice catch! Irish played Tenn in 1991 and 2001 (Nov both years). Tenn won both games. Corrected record against SEC - 7-2.

 

[irishog77]

Nice catch! Irish played Tenn in 1991 and 2001 (Nov both years). Tenn won both games. Corrected record against SEC - 7-2.

Absolutely brutal 2 games for me.