Putting ths season into context, Part II: looking at the past to predict the future

Submitted by taistreetsmyhero on

This is Part II of the Season Score Metric v2.0 diary series. Part I is here.

With the [dry] explanation of the Season Score Metric methodology out of the way, we can now plug this season into the metric and see how it compares to previous years.

But first, let's start with a broad overview and look at the distribution of records over the last 49 years:

We can see that 8-4 and 9-3 are among the most common season results. Interestingly, there was one team that finished 8-3: the 1969 Michigan team. Bo's first year.

Now, let's look at this season's current Season Score.

2017 Season Score Breakdown


Season Score
Preseason AP Record E&I Score Win Score Loss Score OSU MSU Bowl B1G Champ Nat Champ
1.31 11 8-3 -0.66 2.28 0.14 NA -0.45 NA 0 0

From the previous post, the performance quartiles are:

  • Bad, < 2
  • Fine, 2-6
  • Good, 5.5-8.5
  • Great, >8.5

So, this season is currently in the "Bad" quartile, but it is a win against OSU away from jumping into the "Fine" one.

This is how the Season Scores break down across coaching tenures:

We can see that Harbaugh isn't quite in the same tier as Bo, Moeller, and Carr, but he's also not Hoke. So please, for the love of God, don't call him the $9 million Hoke.

That important distinction acknowledged, let's return to this season.


The 8-4 Season

If Michigan beats OSU, the rest of this discussion is entirely moot. It is the prototypical situation where beating OSU can make the season--or break an already broken one. But, in the interest of preempting the doomsayers should we fall next week, I'm going to assume the most likely scenario happens and we finish the regular season 8-4.

A loss to OSU (*assuming the results of all of our previous opponents' games go according to Massey Win Predictor*) would drop the Season Score to -0.16. Here is a chart of all of the 8-4 seasons to put this record in context:

We can see that, clearly, not all 8-4 seasons are the same. This one would fall in the bottom half. Interestingly, if you look at the Win Score (which is Wins x Win % x Defeated Opps' Win % w/ Mich game removed), it is the lowest score. Alternatively, if you look at the Loss Score, it is also the lowest. That means that these would be the most hollow 8 wins out of all the 8 win seasons, but also they would be the 4 most "quality losses" of any 4 loss season. Well, that...sounds about right.

Now that we have identified 3 very comparable seasons to this one, let's take a closer look at how those teams arrived at 8-4, and see how they rebounded the following year.


Rebound Seasons: 3 case studies

Note: I didn't watch any of Bo's seasons, so I would love for anyone who did to add more perspective to the following high level season reviews.

1. 1979

Schedule Highlights:

  • Lost to #9 ND by 2
  • Won the next 8 games in a row against a bunch of body bags
  • Lost @ #14 Purdue by 3
  • Lost to #2 OSU by 3
  • Lost to UNC in the Gator Bowl by 2

[Offensive] Roster Highlights:

  • QB was a JR John Wangler
  • Leading WRs were a SR Doug Marsh and FR Anthony Carter
  • Leading RB was a SO Butch Woolfork

The Rebound: 1980

Season Highlights:

  • 10-2
  • #4 Final AP Rank
  • Outright Big Ten title
  • Rose Bowl Champion
  • Season Score of 12.52, #2 ranked season

Schedule Highlights:

  • Back-to-back OOC losses #8 ND and unranked South Carolina
  • Undefeated in Big Ten play, featuring a 26-0 demolition of #16 Purdue, followed by a 3-point win @ #5 OSU
  • Rose Bowl victory over #16 Washington

Offensive Roster Highlights:

  • QB was a SR John Wangler, who threw for 7 more TDs on only 2 more INTs compared to the previous season
  • Leading RB was a JR Butch Woolfork, who had a very comparable season
  • Leading WR was a SO Anthony Carter, who doubled his yardage and TD production from the previous season

2. 1987

Schedule Highlights:

  • Lost to #16 ND by 19
  • Lost @ unranked MSU by 6
  • Lost @ #15 Indiana by 4
  • Lost to unranked OSU by 3
  • Beat unranked Alabama by 4 in the Hall of Fame Bowl

Offensive Roster Highlights:

  • QB was a JR, first-time starter Demetrius Brown
  • Leading WR was a SO Greg McMurtry
  • Leading RB was a SR Jamie Morris

The Rebound: 1988

Season Highlights:

  • 9-2-1
  • #4 Final AP Rank
  • Outright Big Ten title
  • Rose Bowl Champion
  • Season Score of 10.01, #9 ranked season

Schedule Highlights:

  • Started season with back-to-back, OOC losses @ #16 ND and #1 Miami
  • Undefeated in Big Ten play, with one tie @ unranked Iowa
  • 25 point win over #14 Indiana
  • 3 point win @ unranked OSU
  • Rose Bowl victory over #5 USC

Offensive Roster Highlights:

  • QB was a two-headed attack with JR Michael Taylor and an incredibly efficient SR Demetrius Brown
  • Leading RBs were new from the previous season: another two-headed beast with JR Tony Boles and JR Leroy Hoard
  • Leading WR was a JR Greg McMurtry, who posted nearly identical numbers as his previous season

3. 2001

Schedule Highlights:

  • Lost @ #15 Washington by 5
  • Lost @ unranked MSU by 2
  • Beat #22 Illinois and #17 Purdue
  • Lost to unranked OSU by 6
  • Lost to #8 Tennessee by 28 in the Citrus Bowl

Offensive Roster Highlights:

  • QB was a SO John Navarre
  • Leading WR was a SR Marquise Walker
  • Leading RBs were a JR BJ Askew and SO Chris Perry

The Rebound: 2002

Season Highlights:

  • 10-3
  • #9 Final AP Rank
  • Season Score of 6.78, #18 ranked season

Schedule Highlights:

  • Beat #11 Washington by 2
  • Lost @ #20 ND by 2
  • Beat #15 Penn State, then lost to #13 Iowa by 25
  • Lost @ #2 OSU by 5
  • Beat #22 Florida by 8 in the Outback Bowl

Offensive Roster Highlights:

  • QB was a JR Navarre, who had major bumps in his yardage, TDs, and efficiency
  • Leading RB was JR Chris Perry, with SR BJ Askew as the change-of-pace back
  • Leading WR was a SO Braylon Edwards, who posted a 1,000 yard, 10TD season

Summary

Unless Michigan beats OSU, I think it is more than fair to say that this season was a bust. Even if Michigan performed to the floor of expectations going into the season, given how hollow these 8 wins ring, the floor expectations were for a bad season relative to the context of the last 50 years of Michigan football.

However, as history shows us, an individual bad season is not uncommon. Furthermore, Bo and Carr consistently followed bad seasons up with good or great ones. In the grand scheme of things, one bad season is largely meaningless.

This season matches bits and pieces of each of the 1979, 1987, and 2001 seasons. There is no reason to think that Michigan can't swing back in a big way over the next two years like those teams managed to do. By doing so, Harbaugh would really start to cement his place amongst the Michigan greats.

Comments

Eye of the Tiger

November 20th, 2017 at 1:53 AM ^

...in my opinion. 

I mean, the team we fielded against Wisconsin featured: 9/11 defenders who didn't start in 2016 (counting Hurst as a starter) and 8/11 offensive players who didn't start in 2016. Add in critical injuries, for a team where many backups are true freshmen, and it's not a great recipe for success. 

In 2018, we'll likely feature 2/11 new starters on defense (or 3/11 if Winovich goes to the NFL); and 2/11 new starters on offense. Unlike many positions in 2017, the guys stepping into those roles will have significant game experience already. Also unlike 2017, we'll have credible backups at most if not all positions.

All those dudes will have an extra year of natural physical development, an extra year in S&C, greater mastery of the playbook and more discipline in executing their roles. And despite the fact that your system calls our season "bad," it's not 2014 or 2010 bad--and in both those cases we improved mightily the following year. Yes, coaching changes mattered quite a bit in those cases, but the principle of young players improving dramatically over a single year holds.

I did not expect us to be very good this year--I expected us to max out at "pretty good," which is about right. We've beaten every not good team we've played and lost to all the decent ones. Next year, however, we should be straight up good, even very good. Unlike this year, I expect us to contend for the conference championship in 2018. We may not win it, but we should actually be in contention for it. 

   

Ghost of Fritz…

November 20th, 2017 at 9:02 AM ^

great post.  Thanks for putting in the work.  Factual info is always much appreciated.

The core of this team is very young.  Stands to reason that as the 16 and 17 classes mature, the team will get much better results.

Still some ifs...  QB situation has to be better next year (Peters shows real promise); o-line (who knows? hope so). 

befuggled

November 20th, 2017 at 9:18 AM ^

The 1979 team struggled to find its offensive identity and three different quarterbacks started games (BJ Dickey,  John Wanger and Rich Hewlett). The defense was good enough to keep the team in all of its losses, but special teams were a big factor in all four of the losses. Notre Dame blocked a last minute field goal attempt to preserve their win, Purdue blocked a punt, Ohio State blocked a punt for the winning score, and a missed extra point against North Carolina led to a failed two-point conversion after they scored what could have been the tying touchdown. (No guarantee that they wouldn't have gone for two if they had made the extra point, though.)

By the end of the season it seemed as if Wanger had established himself as the starter at QB, but then Bo surprised everyone and started Rich Hewlett against Ohio State. They had some initial success but it didn't lead to any points, and Hewlett went out with an injury in the second quarter. Wangler himself was hurt in the Gator Bowl (on a tackle by none other than Lawrence Taylor), and Hewlett wound up struggling in a couple of starts the next year before Bo gave the job back to Wangler.

I think Doug Marsh was a tight end instead of a wide receiver. At least that's what he played in the pros. 

Demetrius Brown in 1987 was something of a turnover machine. All of the losses except the initial loss to Notre Dame were close; Brown threw 7 interceptions against Michigan State and struggled in a lot of games (finishing with a completion percentage of 47.6%). He lost the job to Taylor in 1988, but after Taylor was hurt Brown took over again and didn't turn the ball over once. 

Brown got into some kind of trouble after the season and wasn't on the 1989 team. I do wonder what would have happened if he'd been around, as Taylor was hurt against Notre Dame in a close loss and was rumored to have been playing hurt against USC in the Rose Bowl (another close loss).

LeCheezus

November 20th, 2017 at 9:23 AM ^

Outside of the complete and total NFL factory football teams, there seems to be far more parity in CFB than there used to be.  There may be a lot of factors in this: Scholarship limitations, lots of TV exposure even for non blue blood teams and more frequent coaching changes to name a few.  As a result teams with good but young talent or ones missing a key piece or two seem more likely to take a big tumble than 10-15 years ago.  Notre Dame was awful last year but I want to say 7 of their 8 losses were by one score or less - they're not going to make the playoff but they definitely turned things around in a hurry.  MSU managed to put up 3-9 with 6-6 talent and will probably swing all the way back to 9-3 this year - the only major difference I see this year with that team is a competent quarterback and safeties who are less bad in coverage.  Georgia went 6-6 last year and is contending for the CFP this year, even with a true freshman at QB. 

Florida, Tennessee and FSU are all having downright awful seasons by their standards and I expect at least 2/3 to bounce back next year.

The combination of new starters and injuries at the skill positions has just been too much to overcome this year.  Call that excuses if you want, I don't care.  Should they probably still have beaten MSU? Probably? Maybe?  Losing to the 4 teams on the schedule that are actually decent may be disappointing, but would you feel better if we'd won one of those games but dropped one of the games we did win?  The base level for a good team is beating the teams you should beat, and at the very least we did that.  I think it's reasonable to be both dissapointed and while acknowledging that a lot of things (mainly injuries) just didn't go our way this year.

tybert

November 26th, 2017 at 7:13 PM ^

2016 3-9 season - Sparty had major locker room issues (leadership) and the QB was chosen as a captain because of the criticism of not having Cook name a captain. JOK was worse than O'Connor but not that much worse. 

They won Iowa, PSU, and UM this year with a +9 TO margin (10 vs. 1) - that will regress to norm next year.

The OSU game showed everyone that they were really a 7-5 or 8-4 team that pulled off two big wins. 

DrewGreg

November 20th, 2017 at 9:26 AM ^

...helps to put some perspective around this season, compared to other "like" seasons.

That said, I would be curious to see what the correlation between season scores and the injuries (or lack-there-of) in a given season looks like. Certainly, this season is the exception not the rule, but losing 2 QBS, Multiple WRs, Multiple RBs, #1 CB, #1LT (Newsome) and #1RT among others I am forgetting for significant time/snaps, has got to play a role in how this season is graded or perceived. I don't want to make excuses for the on-field performance of this team, but the above injuries combined with the natural attrition from the previous year shouldn't be ignored.

 

 

You Only Live Twice

November 20th, 2017 at 9:56 AM ^

You did some great work here.  I struggle a bit to find comparisons to the Bo years in today's football world. Yes, we had wins that were definitely more comfortable to experience.  Most teams we played, including in the conference, were cupcakes.  It always came down to The Game and maybe one other strong contender or occaisonal flukey upset.  We had big, punishing lines and less QB injury.  In line with the cupcake reference, if a team wasn't ranked, they were certainly in that category, and the drop off after the #10 position or so in the AP was steep.  Today, as LaCheezus was noting above, the landscape is so much different.  Most any team among the top 40 or so teams has a legitmate shot at upsetting any other team, on most any given Saturday.  

The USC game will always burn in memory as the nation watched Charlie White lose the ball on the one yard line and the refs called it a TD.  

taistreetsmyhero

November 20th, 2017 at 11:18 AM ^

By the end of the regular season, we will have played 4 good-great teams this year, and 8 cupcakes. In comparison, those "bad" 8-4 Bo teams played 2-4 solid-great teams a year, and  8-10 cupcakes. You can argue about the relative horribleness of those cupcakes, but they are all squarely in the bad-at-best category. So the win qualities are basically equally hollow, but the losses are more respectable now compared to back then.

SeattleWolverine

November 20th, 2017 at 12:21 PM ^

The 2001 season is a good comparison to 2017. We had lost basically the entire offense off of the 2000 team except for Walker and Walker was basically our only good player on offense that year. Well, Goodwin was still around I think? But anyway, yeah obviously the guys on the 2000 offense were more talented (but underutilized...which is another thread) than the 2016 group with ATrain, Terrell, Henson, Hutchinson, Backus, Mo Williams, Brandt etc IIRC so there was a lot to replace and the offense truly struggled. We made it worse by going super conservative at times. Some building blocks like Pape, Askew, Perry, Joppru but they needed more seasoning. 

 

Likewise, the defense was still good in 2001 and particularly the LB play with Foote and Hobson. Shantee Orr. Rumishek, Eric Brackins, Marlin, Diggs, LeSeuer, Todd Howard, the much maligned Cato June, Charles Drake etc. Pretty solid defense. 

 

I mean honestly, there was nothing magic about the improvement in 2002. Players got better, particularly Navarre as effectively a 2nd year QB (though he played like 4 games in 2000 including the UCLA implosion). Losing Walker hurt but Braylon emerged as a more athletic and similarly productive sophomore. Baas, Petrezullio, Pape etc were all better with more experience. Joppru was good. Bellamy was eh, just a guy. Askew and Perry better. That's pretty much the obvious path forward here. Although the records may not show it, but that team wasn't actually really good until 2003 so maybe not such a positive sign for 2018. 

 

Of course, with the lack of junior contributors on the depth chart, the QB/OT development cycle, and the heavy road/home split for the next year few years, it is looking like 2019 is more likely to be our year than 2018. :( 

taistreetsmyhero

November 20th, 2017 at 3:04 PM ^

I think the quality of the team will improve drastically next season, but given how tough the schedule looks, the record may only improve by 1-2 wins. 2003 was a great team and much better year than 2002. I think this team follows a nearly identical trajectory, but hopefully 2019 team puts it allllll together and reaches the promise land.

ak47

November 20th, 2017 at 12:37 PM ^

If we have a 2002 type season next season are people going to be happy? Some ranked wins, some ranked losses, lose to osu on the road and have a good season during which nobody considers you a champioship contentder? 

taistreetsmyhero

November 20th, 2017 at 1:08 PM ^

Our schedule next year is brutal. That level of improvement would set 2019 up for another 2006/2016-esque shot at THE YEAR.

Another reason why I think this season failed to live up to expectations is that the timeline has been pushed back a year. People knew this would be a rebuilding year, but expected it was building up to a potential title run for next year. Obviously it still could happen, but this team has done nothing to make that a reasonable expectation for next year.

The key to next season is to at least show enough improvement to make that a reasonable expectation for the following year, when the home/road schedule flips back in our favor.

SeattleWolverine

November 20th, 2017 at 1:26 PM ^

Yeah, another thing about next season is that it has to be successful enough that the narrative about the program doesn't change. When the narrative starts to change from one of building a program to coming up short of expectations, then the recruiting starts to falter. Perception can be reinforcing and become a self-fulfilling prophecy. And we can't have a down year in recruiting. So date, 2019 looks promising, but after having a less than spectacular class this year we need to have a great class next year.