Perhaps the BEST B1G Forum anywhere, here at College Football Fan Site, CFB51!!!

To test my theories I also looked at the SEC for a comparison so @OrangeAfroMan , @rolltidefan , other SEC fans, you might be interested in this.  Here are the SEC results:

For the SEC I used the 12 teams that made up the league from the 1992 football season up until the recent expansions.  Looks similar but the SEC is a bit flukier because they've had divisions longer so the non-protected non-divisional match-ups were only played about 8-10 times in the 32 seasons from 1992-2023.  That small sample size gets even smaller when you cut it in half for H/A.  Also, the SEC has a LOT more neutral site games such as the WLOCP which throw things off a bit.  

In each league the H% was almost exactly 0.100 above the A%.  That is interesting.  

The Bobs did a similar analysis years ago using point spread in conference games and he noted the SEC had the smallest HFA by that metric of any major conference, but a fair bit.

FWIW:
That is way more data than I wanted to mess around with.  I just used straight up W/L so it makes no difference if a team won by the skin of their teeth or blew their opponent out, just W's and L's.  

I've got to think there's another aspect to this kind of thing.  I recall UGA under Richt having a stellar road record.  Doesn't your HFA seem weakened by your above-expected performance on the road?

OSU has a ceiling on HFA because look at it, their road win% is above anyone's home win%.  Shit.  If their HFA was any higher, they'd have won a dozen NCs, lol.  So somewhere in the ingredients, there's got to be an aspect we're failing to incorporate.  I haven't ever spent much time on it, so idk.  

Also, 

.731 Florida
.732 Alabama

God damn, Spurrier was a dude!!!

I've got to think there's another aspect to this kind of thing.  I recall UGA under Richt having a stellar road record.  Doesn't your HFA seem weakened by your above-expected performance on the road?

You are right about Richt having a stellar road record.  I noticed when I was putting this together, check this out:
Home winning % in games among the 1992 SEC teams:

• .785 Florida
• .768 Bama
• .717 Georgia

Road winning % in games among the 1992 SEC teams:

• .731 Georgia
• .697 Bama
• .676 Florida

So Bama was neither the toughest place to play (Swamp) nor the toughest team to host (UGA) but they were the best overall team because they were second in both of those things. 

My guess is that a big part of Georgia having a negative HFA is the fact that they play a big rivalry game at a neutral site (your Gators obviously).  Granted that wouldn't have mattered much in a lot of years because for large chunks of the 32 years from 1992-2023 Florida was substantially superior and then for other chunks Georgia was substantially superior.  The other weakness of that explanation is that obviously Florida also played in that game and they have a pretty strong HFA of .109. 

OSU has a ceiling on HFA because look at it, their road win% is above anyone's home win%.  Shit.  If their HFA was any higher, they'd have won a dozen NCs, lol.  So somewhere in the ingredients, there's got to be an aspect we're failing to incorporate.  I haven't ever spent much time on it, so idk.

This is definitely true and the converse is also true.  Ie, in the SEC Vanderbilt's Home win% was lower than any of the other teams' road win% so it just couldn't get much worse in road games for Vandy just like it couldn't get much better in home games for tOSU.  I think we are incorporating that with the theory that teams that are really good or really bad tend to have minimal HFA.  Ohio State's is 0.083 and Vandy's is 0.005.  Similarly, Bama's is 0.070 and Illinois' is -0.018. 

This is the point that @betarhoalphadelta and I have been making for a while.  HFA isn't just about who you play where but also who you are.  UCLA looks to be one of the worst teams in the B1G.  For them an advantageous schedule is one where they get the other bottom feeders at home because, realistically, they are extremely unlikely to knock off tOSU/PSU/Oregon no matter where the game is played.  The Bruins are better off to lose in Columbus, State College, and Eugene but then get Northwestern, Maryland, and Purdue at home.  Conversely, Ohio State looks to be one of the best teams in the B1G so for them an advantageous schedule is one where they get the other contenders at home because they shouldn't and realistically probably will not lose to UCLA, NU, UMD, or PU even on the road.  The Buckeyes are better off to win (even if one or two are close) in Westwood, Evanston, College Park, and West Lafayette but then get PSU, Oregon, Illinois, and Michigan at home. 

Also,
.731 Florida
.732 Alabama

God damn, Spurrier was a dude!!!

Yeah, you got to enjoy watching one of the all-time best as the coach of your school.  I've always loved the story that Spurrier's first game as HC at Florida was on September 8, 1990.  The unranked Gators demolished the unranked OkSU Cowboys 50-7.  It was the ONLY game in which Spurrier's Gators were unranked.  The Gators were then ranked in 209 consecutive polls from 1990-2002 which encompasses the rest of Spurrier's tenure and slightly beyond.  That streak of 209 consecutive AP Poll appearances is the 4th longest such streak of all-time trailing only Nebraska's 1981-2002 streak, Bama's current streak, and FSU's 1989-2001 streak. 

What makes Spurrier's accomplishments more impressive to me is that Florida was very much NOT an established power when he arrived.  Florida had only appeared in 196 of the 670 AP Polls from 1936-1989 (29.3%) and they had only finished ranked 10 times. In Spurrier's 12 years as HC in Gainesville the Gators had more AP Poll appearances, more final AP Poll rankings, and more NC's than they had in their entire history before he arrived. 

The other weakness of that explanation is that obviously Florida also played in that game and they have a pretty strong HFA of .109.  This is definitely true and the converse is also true.  Ie, in the SEC Vanderbilt's Home win% was lower than any of the other teams' road win% so it just couldn't get much worse in road games for Vandy just like it couldn't get much better in home games for tOSU.  I think we are incorporating that with the theory that teams that are really good or really bad tend to have minimal HFA.  Ohio State's is 0.083 and Vandy's is 0.005.  Similarly, Bama's is 0.070 and Illinois' is -0.018. 

Yeah... Imagine that you break any 12-team conference into 3 tiers.

When playing another team within your own tier, HFA is significant. You're closely matched enough that playing at home or away makes a meaningful difference. I.e. if you're projected to win 50% of these games on a neutral field, that might be 60% at home and 40% on the road (20% swing). 

When playing any team one tier above or below you, HFA is a factor but it's minimal. These are teams that you should either beat or lose to consistently, but HFA can make it a little bit more likely to have a positive or negative upset. I.e. if you're projected to win 15% of games against a team a tier above you, but that number becomes 20% at home and 10% on the road (10% swing). 

When playing any team two tiers above or below you, HFA really isn't a factor. You're maybe going to win that game 1 out of 20 times (5%). At home that might be 1 out of 15 (6.7%) and on the road 1 out of 30 (3.33%). So you really only have a very minimal swing in those games home vs away. 

So if you break down the tiers:

• Top tier: 3/11 (27%) of their games (since they can only play the three, not four, other teams in their tier) have a meaningful HFA. 4/11 (36%) of their games are going to have minimal but still relevant HFA. And 4/11 (36%) of their games will have almost zero HFA effect.
• Bottom tier: Same as top tier, just the opposite direciton. 
• Middle tier: 3/11 (27%) of their games have a meaningful HFA, and 8/11 (73%) of their games will have minimal but still relevant HFA. They have zero games with almost zero HFA effect. 

So as a result, HFA will most affect the middle tier teams, and those are the ones that logically should have the largest HFA effect. 

You also want to take into account the layout of the home stadium itself.

Iowa's Kinnick Stadium is well known for having tight areas between the sidelines and the stands,  where the fans are extremely close to the field.  This helps crowd noise affect the play on the field  a little more,  and sometimes Iowa fans can even affect the visiting team while they are standing on the sideline,  close to those annoying fans.

On the other hand,  I have had people tell me how surprisely quiet 100,000 fans can sound in some of the Big Ten's bigger stadiums.

Among the 11 teams of the Big11Ten there are 55 possible match-ups (10 for the first team, 9 for the second . . . =55).  Three of those have the exact same winning percentage regardless of location:

• Ohio State / Indiana:  The Hoosiers are 0-fer regardless of location.  
• Penn State / Purdue:  They played 18 times between 1993-2023 with Purdue once at home and once in State College.  Penn State leads 8-1 in both home and road games against Purdue.  
• Wisconsin / Indiana:  They played 20 times from 1993-2023 with Wisconsin holding an 8-2 advantage both in Madison and in Bloomington.  

In 15 of these, there is a statistical HFD or Home Field Disadvantage:

IMHO, you can just throw the bottom eight out as "noise" due to small sample-size.  Specifically, if you look at #8, Minnesota had one win over Ohio State from 1993-2023 and that one win happened to occur in Columbus so the Gophers went 1-9 in Columbus and 0-9 against Ohio State in Minneapolis.  

The other seven are harder to toss aside but I still think most they are 'noise' issues.  

• Like Minnesota, the Illini didn't beat Ohio State in Champaign between 1993-2023 but they did win four times in Columbus.  
• Northwestern went 3-9 against the Spartans in Evanston but 6-5 in East Lansing.  
• The Illini were 2-10 against the Boilermakers in Champaign but 5-8 in West Lafayette.  
• Just like with Ohio State, the Illini beat Michigan four times between 1993-2023.  In this case it wasn't all four but still three of the four wins were in Ann Arbor so the Illini went 1-8 against Michigan in Champaign and 3-7 in Ann Arbor.  
• Minnesota beat Michigan twice between 1993-2023 and both of those were in Ann Arbor.  0-11 at home, 2-9 in the Big House.  
• This one is a one-game difference.  The Illini went 3-13 against the Spartans but two of the three upsets were in East Lansing so they end up 1-8 against MSU at home and 2-5 in East Lansing.  
• The Spartans beat Ohio State four times (in true H/A games) from 1993-2023 and three of the four were in Columbus so they ended up 1-11 against Ohio State in East Lansing and 3-9 in Columbus.  

Here are the 37 match-ups with a statistical HFA:

I will not go through all of these but most of them make sense.  The top 10:

• Iowa was the 5th and MSU the 6th best team in the league so you would expect them to be evenly matched such that HFA would be more frequently decisive.  
• Minnesota was the 9th and Indiana the 11th best team in the league same.  
• This one makes no sense.  Ohio State was the best (by a LOT) and Purdue was 8th so you wouldn't expect HFA to matter much in this series but they call them the Spoilermakers in Columbus for a reason.  Purdue went 0-9 in Columbus but 5-6 against Ohio State in West Lafayette.  
• Wisconsin was the third and MSU the 6th best team so this makes sense.  
• These were the worst two teams in the league so this makes sense.  
• Purdue was the 8th and Minnesota the 9th best team so this makes sense.  
• This one seems odd because NU was the 7th best and IU was the worst team but I think NU's highs and lows make this one make sense.  Most of the time NU was a bottom-feeder right with IU.  
• Penn State was the 4th an MSU the 6th best team in the league so this makes sense.  
• Michigan was the second and Wisconsin the 3rd best team in the league so this makes sense.  
• These were the best two teams in the league so this makes sense.