SP+ Trends Michigan, Alabama, Ohio State

Submitted by RealElonMusk on December 5th, 2023 at 10:10 AM

IMHO SP+ is the best predictor of college football games with the sometimes exception of Vegas.  SP+ is not a ranking by resume like the AP Poll, it's designed to make weekly predictions by looking at play by play success to filter out the effects of tempo & opponent variations.

The overall rating is how much better the team is than the average D1 team-  Michigan's SP+ rating is 31.1 -  Michigan should be the average team by 31.1 points.   Alabama is 22.7 SP+ so on a neutral field Michigan should win by 8.4 points.  But wait, this is Alabama we are talking about-  how is this possible?

Well-  this is not prime Alabama-  Michigan played Alabama 2020 and lost 16-35 -  that year, Michigan's SP+ rating was 10.7 while Alabama's was 28.3 a 17.6 point difference-  Michigan lost by 19  

Below are trend charts for Alabama and Michigan since 1990.  This shows why Alabama dominated from 2010 to 2020: their SP+ rating was 29 or better from 2009-2019. Fortunately for Michigan Alabama has declined significatly from their peak of 36.3 in 2018.

The SP+ chart of Michigan and OSU very clearly shows the decade of pain for Michigan from 2010-2020:   From an SP+ perspective, since 2010 the better team has won every year except 2016 (ref game & injured Speight) & 2021 (Snow :  ).    

Also, the Big10 average hasn't been very good since 2005 and is down about 3pts from last year.

 

Wally Llama

December 5th, 2023 at 12:32 PM ^

This was a great moment - thanks for bringing it back up as it's not discussed much in the pantheon of great UM football moments.

To this day, however, I'm still a bit disturbed about Moeller having that clipboard stuffed in his pants....

H/T to Wolverine historian for these recap videos:

1988 HoF Bowl Part 1: https://youtu.be/Y47jhOfdnic?si=uB4JH2xbLq1WTPPG

1988 HoF Bowl Part 2: https://youtu.be/waGp-Mr6wTY?si=pC5vfurF5zl1w7Wu

Perkis-Size Me

December 5th, 2023 at 1:21 PM ^

I wonder how much this year Alabama tends to play up to or down to its opponents, or if they just don't "get up" for everyone the way a vintage Alabama team seemed to always do.

Saban's prime Alabama teams, like the teams from probably 2011-2020, they NEVER experienced a true upset. I mean yeah they got beat from time to time and they were technically considered upsets, but it was always by opponents of equal or damn near equal caliber. They laid waste to everyone else. That's what I always remember being what separated Alabama from everyone else.

OSU could get upset by Purdue, Iowa, or Virginia Tech. Clemson could get upset by Pitt. Oklahoma could get upset by a Texas Tech or Kansas State. Alabama NEVER experienced a true upset under Saban during the 2010s. Ever. And they almost never played down to their competition. They just pistol-whipped everyone. 

I guess what I'm saying as I come back to my original thought: I'm not putting a lot of stock into what they did against Arkansas or Auburn, as those are terrible teams this year that Alabama simply may not have "gotten up" for (if that's the case against Auburn, then shame on them for doing that in a rivalry game but still, Auburn sucks this year). But then they turn right around and knock off the #1 team in the country. Limited sample size perhaps, but whatever the best version of this particular Alabama team is, I'm betting that's what we get at the Rose Bowl. Even if we don't get that, we 100% won't be getting the team that played Auburn or played Arkansas or South Florida.

MGoHomeUrDrunk

December 5th, 2023 at 10:40 AM ^

Here's a primer from the source (Bill C.): https://www.sbnation.com/college-football/2017/10/13/16457830/college-football-advanced-stats-analytics-rankings

"It takes a while for a college football team to create data of statistical significance or establish reliable priors. Therefore, for predictive purposes, you have to lean pretty heavily on preseason projections.

My S&P+ projections incorporate recent performance, recent recruiting, and returning production; early in the season, they carry significant weight, and they are phased out with increasing speed. After a team has played seven games, its preseason projections are completely phased out of the S&P+ equation.

There is a decent case for keeping the preseason numbers involved for much longer than that, but maintaining anything beyond seven weeks has not, in my experience, changed predictive accuracy to any major degree."

EGD

December 5th, 2023 at 10:46 AM ^

Alabama may have "looked" better but considering it was just a couple weeks ago they needed a miracle pass on 4th & a cab ride to beat a wretched Auburn team and less than a month since they squeaked by terrible Arkansas by a FG, I don't know how much the computer models out to be treating them as a vastly improved squad. 

gwrock

December 5th, 2023 at 10:55 AM ^

If Alabama doesn't complete that 4th and 31 touchdown pass, but still beats Georgia in the SEC championship game, who would we be playing on Jan 1st?

(Also, it's still funny to me that Alabama is now "one of the best 4 teams in college football" because they completed that one crazy pass, but wouldn't be if they didn't).

WampaStompa

December 5th, 2023 at 11:05 AM ^

I think they would have still screwed over FSU in favor of Georgia in that scenario 

Also, your comment reminds me about how 1997 Nebraska needed an illegal last second play (that the WR admitted he intentionally kicked in the air) in order to avoid losing to a 7-5 Missouri team. But because they got that assist from the refs against a mediocre team, we have to constantly hear about how awesome they were and their fluke share of the national title

LostPatrol14

December 5th, 2023 at 11:18 AM ^

I think they would have still screwed over FSU in favor of Georgia in that scenario

100% This.

Joel Klatt laid it out perfectly on what the playoff committee does: Path to least resistance. If there is, at least, a 1-loss SEC team, they will not be telling them "No", because they're the SEC. FSU gave the committee an out with the injury. If Georgia only had one loss, they would have just fell to 3 or 4, depending on the other teams/outcomes. It just shows how broken the system is.

EGD

December 5th, 2023 at 12:40 PM ^

I like Klatt, but his position on the CFP has been inconsistent and incoherent. For weeks he was criticizing the committee for taking the "path of least resistance" by ranking the undefeated teams in a top tier, then the one-loss teams after that--even though arguably some of the one-loss teams were "better" than the undefeateds. Then he started (rightly) complaining about the committee propping up the SEC by overrating Tennessee and Missouri, but then he himself got on the "you have to take Alabama and Texas over FSU because nobody wants to see FSU with an injured QB" bandwagon. Then when the committee did exactly that, he called that the "path of least resistance" and started saying how much it sucks for FSU. 

LostPatrol14

December 6th, 2023 at 8:54 AM ^

I get that and I think that's because he is more of "Best 4" than "Most Deserving" crowd. He knows that a Power 5 undefeated champ should be in, and he also knows that, with Jordan Travis' injury, the committee wasn't going to pick FSU. To me, and you hit this point, that Klatt sounded hypocritical in his remarks due to what he said. I like Klatt, as well, but yeah, I got that same vibe, too.

J. Redux

December 5th, 2023 at 11:27 AM ^

FSU would definitely have been in with a two-loss Alabama SEC champion.  Texas would have been out.  The only reason that Texas is in the field is that head-to-head fetishists decided that if they beat Alabama, and Alabama is a playoff team, then Texas must also be a playoff team.

AlbanyBlue

December 5th, 2023 at 2:01 PM ^

RE: 97 Nebraska -- whenever anyone mentions that travesty, I immediately bring up two things: that illegal kick play and then also Tom Osborne retiring (since that "share" was a clear gift to him). Shuts virtually everyone up quickly.

Michigan would have crushed that Nebraska team has they played on the field, and everyone knows it.

EGD

December 5th, 2023 at 10:57 AM ^

I also think S&P+ may be underestimating Michigan this season. Although S&P+ adjusts for garbage time, it does not adjust for teams substantially treating entire games as garbage time as Michigan seems to have done several times this season. 

Also, I don't believe S&P+ does anything to try and adjust for injury effects. M has had a number of key players either miss games or play hurt this season so those guys having several weeks to heal up should give M a boost. 

I am not sure if Alabama has key guys playing with nagging injuries or expects to get anybody back for the Rose Bowl--if so, the second factor could even out. But given the number of close games Alabama has played against mediocre competition this season, it's doubtful they've had very many opponents they didn't need to take seriously like M has had. 

4th phase

December 5th, 2023 at 11:39 AM ^

yeah a good question is how has Michigan performed against the SP+ spread this season.

 

edit: so they didnt cover at all in the non conference, which goes to your point. Otherwise they were 6-3-1 against SP+ spreads. The three they didnt cover were Maryland, Purdue, and Rutgers. The "tie" was OSU, where they won by 6, exactly in line with SP+. 

WampaStompa

December 5th, 2023 at 11:03 AM ^

That OSU graph is pretty interesting to me. First off I'm a little surprised their 2012 team isn't higher than it is, same with their 2014 team when they actually won the CFP. Second, I feel like OSU fans always moan about what their 2015 team could have been, and how they feel like it was their best team in forever. But according to SP+ it was only their 4th best team of the CFP era.

Also lol at the Big Ten average nosediving this year. It kinda feels like the Big Ten average corresponds with the valleys and peaks of OSU. I wouldn't mind seeing Michigan assist them with a metaphorical push over that cliff.

Sione For Prez

December 5th, 2023 at 12:07 PM ^

I'm not all that surprised by their 2014 team. That team lost by 2TD at home to a terrible VT team, needed 2 OT and one of the most ridiculous review scenarios I've ever seen* to beat a bad PSU team, only up 1 against a 3 win Indiana team going into the 4th and the Michigan game against our garbage 2014 team was tied until almost the end of the 3rd quarter. 

They were obviously talented and could bomb some teams out of the building but they had their fair share of struggles throughout many of those games. 

*Hackenberg was "intercepted" but replay on TV and in the stadium very clearly showed the ball wasn't caught at all. But the officials claimed there was "technical difficulties" with the equipment resulting in the play staying an INT even though 100k in the stadium were watching it not be caught. https://www.sbnation.com/2014/10/25/7071291/ohio-state-interception-gif-penn-state

ShadowStorm33

December 5th, 2023 at 12:05 PM ^

Don't forget the overturned Roman Wilson TD that turned into a fumble on the goal line. 21 points right there. And when you think about how the game was turning in the second half, and how they scored that backbreaking long TD late when we were pressing to get a stop (i.e. probably wouldn't have been pressing so hard if the score was closer), flip just one of those three, and we probably win...

WampaStompa

December 5th, 2023 at 11:25 AM ^

I can't find the exact numbers leading up to the TCU game, but Brian's preview said Michigan was favored by 9 points over TCU per SP+.

The only actual SP+ rankings I can find are the final rankings after the national championship, where it says Georgia won by so much that the entire SEC rose in the rankings and everyone else fell. Michigan ended at #3 and TCU #8 per SP+, so I assume both were ranked higher before the CFP started. This year Michigan is currently #1 to SP+ and Alabama is #7. 

M Go Cue

December 5th, 2023 at 11:13 AM ^

The Alabama graph is interesting.  Saban took over in 2007.  In the five years prior their recruiting average ranking was 24.6, but mostly in the mid teens.

From 2007-2016 their recruiting average shot up to 2.4, with the number one class in 7/10 years.

This all appears plain as day in the graph.

MgofanNC

December 5th, 2023 at 11:43 AM ^

Games are decided by unpredictable things like injuries, turnovers, individual matchups and in-game plans and adjustments. 

The UM vs PSU game I think is a good example of this. 

  • Injuries: not many in this game but JJ did get injured in the second half (ankle) kept him out of the run game mostly and limited his passing abilities
  • Turnovers: PSU's lost fumble out of the half and their fail on 4th down put the game out of reach
  • Individual Matchups: Barnhart's mismatch on the edge was a massive problem for UM. Forced us to play differently than we likely wanted to. 
  • In-game Adjustments: UM adjusted to its Mammoth package and stuck to a choke the game out plan limiting possessions (and scoring opportunities) for PSU. 

These things are mostly unpredictable. Games against outmatched opponents usually can be predicted just on the individual matchup metric because there will be so many wins for the more talented team that the other factors get overwhelmed (usually). In games where the matchups/talent is much closer to even these other factors play a much bigger role. There are no doubt a lot of models and educated guesses as to how the game will go or who has the upper hand. I bet it will come down to the factors listed above.  

Blue Middle

December 5th, 2023 at 11:45 AM ^

I'm not afraid of 'Bama.

In fact, I'm glad we got this game.  It sucks that FSU got screwed, but if we beat FSU and then the other winner, SEC slappys would be talking about how we didn't beat the "best" and 'Bama would have been the champs.

We are the best team in the country.  I'm not guaranteeing victory nor am I 100% confident in a win.  

But I feel like this is a game we can and should win.  We're not a plucky underdog taking on a juggernaut.

I don't get the hand-wringing over 'Bama.  Yes, they beat Georgia.  But they lost to Texas, scraped-by Auburn (rivalry games can be weird), and don't have the death star defense or Heisman candidate RB/WR/QB they seemed to have for so many years.

It's time people started fearing and respecting Michigan as one of the elites.  And I think team 144 will prove that.  Maybe even decisively.

I think it's much more likely Michigan wins by 10+ than the other way around.  I think the most likely outcome is Michigan by a TD or so.  Vegas is pricing in Saban's earned reputation, but Michigan is the better team by a significant margin.

Bluetotheday

December 5th, 2023 at 11:47 AM ^

I’m guessing this statistical analysis will not be discussed on the roll tide blogs!

this is an important game, this win will be huge for Michigan. Going through Alabama is important to demonstrate you are elite.