UM/MSU/ND: A Study in Mutual Opponentism

Submitted by Brhino on September 10th, 2012 at 4:49 PM

INTRODUCTION

This diary is an attempt to answer the following question:

Can the results of the Michigan/Notre Dame and Michigan State/Notre Dame football games be used to predict the results of the Michigan/Michigan State game?

INSPIRATION

State plays Notre Dame before we do this year.  I thought that was unusual, but I wanted to see just how unusual it was.  From there, I thought it would be interesting to compare the results of the games between the three teams on any given year, and see how closely related UM's and MSU's performances against Notre Dame can be matched to their performances against each other.

RAW DATA

I decided to look back at the last 20 years in which Michigan, Michigan State, and Notre Dame all played each other.  In 2000 and 2001 MSU played ND but Michigan did not.  In 1995 and 1996 neither MSU nor UM played ND.  Therefore the data spans from 1988 from 2011 with a few gaps in between.

Year First ND-UM Result ND-MSU Result MOV Diff UM-MSU result MOV
2011 UM UM-35 nd-31 ND-31 msu-13 22 MSU-28 um-14 -14
2010 UM um-28 ND-24 MSU-34 nd-31 (OT) 1 msu-34 UM-17 -17
2009 UM UM-38 nd-34 ND-33 msu-30 7 MSU-26 um-20 (OT) -6
2008 UM ND-35 um-17 MSU-23 nd-7 -34 msu-35 UM-21 -13
2007 UM UM-38 nd-0 msu-31 ND-14 21 um-28 MSU-24 4
2006 UM um-47 ND-21 nd-40 MSU-37 29 UM-31 msu-13 18
2005 UM nd-17 UM-10 msu-44 ND-41 (OT) -10 um-34 MSU-31 (OT) 3
2004 UM ND-28 um-20 nd-31 MSU-24 -1 UM-45 msu-37 (OT) 8
2003 UM UM-38 nd-0 msu-22 ND-16 32 um-27 MSU-20 7
2002 UM ND-25 um-23 nd-21 MSU-17 2 UM-49 msu-3 46
2001 n/a DNP msu-17 ND-10   MSU-26 um-24* -2
2000 n/a DNP MSU-27 nd-21   UM-14 msu-0 14
1999 UM UM-26 nd-22 msu-20 ND-13 -3 MSU-34 um-31 -3
1998 UM ND-36 um-20 MSU-45 nd-23 -38 UM-29 msu-17 12
1997 MSU UM-21 nd-14 msu-23 ND-7 -9 um-23 MSU-7 16
1996 n/a DNP DNP   UM-45 msu-29 16
1995 n/a DNP DNP   MSU-28 um-25 -3
1994 UM um-26 ND-24 nd-21 MSU-20 3 UM-40 msu-20 20
1993 UM nd-27 UM-23 ND-36 msu-14 8 MSU-17 um-7 -10
1992 UM ND-17 um-17 nd-52 msu-31 21 UM-35 msu-10 25
1991 UM UM-24 nd-14 ND-49 msu-10 49 um-45 MSU-28 17
1990 UM ND-28 um-24 nd-20 MSU-19 -3 msu-28 UM-27 -1
1989 UM nd-24 UM-19 ND-21 msu-13 3 um-10 MSU-7 3
1988 UM ND-19 um-17 nd-20 MSU-3 15 UM-17 msu-3 14

terms:

First notes which team (MSU or UM) played Notre Dame first that year.

HOME TEAMS are in bold.  Each series alternates sites yearly, and the three series are synced up such that each team has one home and one away game.  Rather well done on the part of whoever made that work.

MOV Diff is the difference in margin of victory of the two Michigan schools over Notre Dame, with MSU's MOV subtracted from Michigan's MOV.  A higher number means Michigan did better against Notre Dame than State did.  A negative number indicates State did better against ND.  Example:

1997: UM defeats ND 21-14: MOV 7.  MSU defeats ND 23-7: MOV 16.  MOV Diff = 7-16 = -9.

MOV is Michigan's margin of victory over Michigan State in the UM-MSU game.  A negative number indicates Michigan lost sad

DNP means the teams in question did not play each other that year.

* This game gets an asterisk because it was BULLSHIT.  That is all.

I PROCESS INFORMATION BETTER GRAPHICALLY THAN I DO TEXTUALLY DO YOU THINK MAYBE WE COULD HAVE A SCATTER PLOT?

HEY THAT'S TRENDING UPWARD

Just barely.  It's more like a shotgun blast than a linear trend.

  • When Michigan outperforms State in MOV Diff, Michigan beats State 69.2% of the time.  When Michigan does NOT outperform state in MOV Diff, Michigan beats State 57.1% of the time.  That's a slight correlation.
  • When Notre Dame splits its series against Michigan and State, the team that beat Notre Dame is the team that wins the UM/MSU game exactly 50% of the time.  That's zero correlation. 
  • When Michigan comes out on top of the split it goes on to beat State 60% of the time.  When State comes out on top it beats Michigan 33.3% of the time.  Overall Michigan has beaten State 62.% of the time.
  • When Notre Dame beats both Michigan schools, Michigan beats State 77.8% of the time, and when Notre Dame loses to both, Michigan beats State 60% of the time. 
  • MOV Diff and MOV are within two touchdowns of each other 9 times out of 20.  The other 11 times they are further apart.  The most extreme disagreement was in 1998 when Notre Dame beat Michigan by 16, then fell to MSU by 22... but then Michigan beat State 29-17.

BULLET POINTS THAT ARE SORT OF FUN BUT NOT DIRECTLY RELATED TO THE ORIGIONAL QUESTION

  • Michigan's Head-to-Head record against Notre Dame since 1988: 10-9-1
  • Michigan State's Head-to-Head record against Notre Dame since 1988: 10-12.  State's actually 10-5 against ND since their 1995-1996 break; before that MSU hasn't beaten Notre Dame since 1986.
  • Michigan's Head-to-Head record against Michigan State since 1988: 15-9**
  • State playing ND before Michigan does is indeed highly unusual: it's only happened once in the last 24 years - in 1997.  That's right:  Every time MSU plays Notre Dame before Michigan does, Michigan wins a National Championship.
  • Since Overtime began in 1996, the UM/MSU game has gone to overtime three times, the MSU/ND game has gone to overtime twice, and the UM/ND game never has.

**but one of those losses was BULLSHIT.

YOUR STATISTICS ARE BAD AND YOU SHOULD FEEL BAD

Pretty much, but I spent all this time making the graph and there hasn't been a new diary in two days so I'm posting it anyhow. 

SUPPOSING FOR A MOMENT THAT I TOOK THESE STATS SERIOUSLY WHAT CAN I EXPECT?

Well, first of all since we're playing Notre Dame after State does: National Championship.  I mean, it's right there in the data.  Apart from that, if State beats Notre Dame and then Notre Dame beats us don't panic - of the 3 times that's happened in the past 20 years we've gone on to beat State twice anyhow.  Likewise don't get too confident if Notre Dame beats State but loses to us.  Don't put a lot of stock into the Michigan schools' respective margins of victories in the two games against Notre Dame - more often than not, it turns out to be meaningless when Michigan and State take the field against each other.

Comments

Everyone Murders

September 10th, 2012 at 5:44 PM ^

I don't know how predictive this analysis is, given how rosters change from year to year and all the relatively small sample size.*  But it's really interesting and fun to read - thanks for posting this.

Oh, and with regard to your observation that

* This [2001 Spartan Bob] game gets an asterisk because it was BULLSHIT.

Word.

*Brhino acknowledges that the results may not be very predictive in the diary. 

WolverineHistorian

September 10th, 2012 at 5:55 PM ^

Before RichRod took over, none of these stats meant anything the year Michigan and MSU played each other in Ann Arbor.  During that same time frame you used up until 2008, MSU was 1-9 at the big house with the lone victory coming in 1990 when Eddie Brown tripped Desmond Howard on the 2 point conversion and didn't get flagged.  We need to get things back in that direction.  From 1968 to 2008, MSU had a .166 winning percentage at Michigan Stadium.  So to see them win the last 2 on our field (and led by that jackass of all coaches) is one of the worst feelings in the world.  Our boys need to get the earth back on it's axis this year. 

Stats aside, Michigan/Notre Dame is going to be no different than it's ever been.  8 times out of 10, it's going to be a game that's decided late in the 4th quarter with a victory of around 4 points.  Win or lose, it's going to be a nail biter and you will feel like you aged 10 years in the 4th quarter alone.

Ever since Notre Dame started their demise in 1994, Sparty has owned them.  And the games have had so many insane plays and outcomes in MSU's favor that it's almost mind boggling (2006 epic choke job aside.)

 

oHOWiHATEohioSTATE

September 10th, 2012 at 6:08 PM ^

The 1990 game was equally bullshit! Desmond was tripped in the endzone on the 2 point conversion. Some say he caught the damn ball anyways. Yes, Im still very bitter as that is my first recollection as a child learning that life isn't always fair.

snarling wolverine

September 10th, 2012 at 7:23 PM ^

That's right: Every time MSU plays Notre Dame before Michigan does, Michigan wins a National Championship.

That may be unlikely, but we've also won the Big Ten title and Rose Bowl (and had a player win the Heisman) every year that's happened, which would be pretty sweet.

SchembechlerDisciple

September 11th, 2012 at 7:55 AM ^

...I read this... Was disgusted by the "overall 62%" stat... Then felt better about the "15-9 overall record"... Then scratched my head... Then took out my phone and entered 15/24, found out it was .625... Then felt disgusted again... And SUPER stupid for going through that entire pathetic exercise.

Johnny Blood

September 12th, 2012 at 9:01 AM ^

I've always thought MSU benefited significantly from playing ND after us -- I guess we'll see this season if we get a similar benefit from it.

And, to be fair, that fake FG that MSU ran on ND to end the game in overtime was awesome.  I'm no fan of Dantonio, but that took a lot of guts.

snoopblue

September 13th, 2012 at 7:16 PM ^

With ND going to the ACC, it will suck if they rotate Michigan and MSU. I am not only looking forward to watching us play UMASS, but also watching MSU-ND just to see how they both look in their first "real" games.

Also, I think because both Kelly and Dantonio came to their current positions from Cincinnatti, I think this gives the game another dynamic. Who's the better coach? It should be an interesting game. I can honestly say I really don't care who wins, but an ND win would be nice for us. I'm sure we can use the film to find a way to use Denard effectively with their QB being somewhat similar - with a much better arm. A loss would increase the intensity of our night game @ ND the following week.