The Case for Ranking M over ND in the CFP

Submitted by Communist Football on November 13th, 2018 at 3:21 PM

The conventional wisdom is that if the top four teams win out, the final CFP ranking will be 1/Alabama, 2/Clemson, 3/ND, 4/Michigan. I will argue in this post that it is at least likely that Michigan will be ranked ahead of ND if each team wins out.

When teams are comparable, the CFP committee is supposed to consider four criteria:

  • Conference championships won
  • Strength of schedule
  • Head-to-head result (if it occurred)
  • Comparative outcomes of common opponents (without incenting margin of victory)

So let's go through these bullets, one by one.

Conference championships won: If each team wins out, the edge goes to Michigan.

Strength of schedule: Through 11 weeks, Michigan's SOS is ranked 32nd by Bill Connelly; Notre Dame's is ranked 59th. And that's before Michigan plays OSU and Northwestern for the second time. Edge: Michigan.

Head-to-head result: Notre Dame, albeit by one score in the first game of the season in a year in which everyone knows that Michigan has dramatically improved through the course of the season. Edge: Notre Dame.

Comparative outcomes of common opponents: This is the factor that people are underappreciating. Michigan beat Northwestern by one score early in the season; ND played a similarly close game (albeit with a slightly higher scoring margin) more recently. But Michigan gets to play Northwestern a second time. If that game's outcome is more lopsided, it could make a big statement with the committee regarding common opponents. Edge: Notre Dame for now, but possibly Michigan after the championship.

That's either 2-2 or 3-1 for Michigan, if Michigan wins out and beats Northwestern soundly in the B1G Championship.

This matters, of course, because the #4 seed is likely to play Alabama in the first round.  Should be interesting either way...

mgowill

November 13th, 2018 at 4:18 PM ^

I am paranoid about this.  I just keep thinking that if we win out, are crowned B1G Champs for the first time since 2004, and stand before the playoff committee against -

  • 13-0 Clemson
  • 12-1 Georgia
  • 12-0 Notre Dame
  • 12-1 Alabama

Michigan gets told to kindly go fuck themselves.  Alabama would essentially have the best loss at that point, as many wins, and the same amount of losses.  Sure we have B1G crown, but I can already hear the SEC slappies rolling out how shitty the B1G is this year.  They've already started that campaign by under ranking B1G teams from the first week they were released.  Look - they beat a 7-5 or 6-6 Northwestern team, the B1G is ass - I can hear them now.

I agree with you MGrowOld.  We need a two loss Georgia.

mGrowOld

November 13th, 2018 at 4:30 PM ^

Sadly no.  Each week the thing resets so when they evaluate the quality of the wins/losses they do so based on how the team is ranked at that time, not what they were ranked when they played.

This is why I've been screaming for a week now about what's coming.  By consistently over-ranking the SEC teams all wins are quality wins and all losses are quality losses.  By consistently under-ranking the B1G teams wins are not considered quality wins and all losses are considered bad losses.

It's a shell game where they cant lose.  

JPC

November 13th, 2018 at 3:32 PM ^

ND is good. They’re better than they were at the start of the season, just like we are. If we played again on a neutral field, I’m not sure we’d beat them.  

They're rightly ranked above us, regardless of their bullshit semi-ACC schedule. 

joeyb

November 13th, 2018 at 4:02 PM ^

There isn't a predictor that I'm aware of that doesn't have Michigan favored by 5+ on a neutral field. Even in the game we played, given the box score, S&P+ would have given us a 60% chance to win that game, i.e. if we played 10 games at their field, we'd be expected to win 6 of them. We just happened to catch one of the 4 where they beat us.

bluepalooza

November 13th, 2018 at 4:14 PM ^

Michigan would squash ND like a bug. Yes ND is a little better. Michigan has #1 D. Patterson is much better now then game 1. Oline is 100% better. You want to make an argument that Clemson and Bama would be M’s toughest opponents by far. Absolutely. ND deserves to be right where they are, but M would beat them by at least 17. 

kevbo1

November 13th, 2018 at 3:35 PM ^

ND SOS is pathetic outside of playing Michigan.  In a normal year it would have been tough, so I can't fault their scheduling.

EGD

November 13th, 2018 at 6:24 PM ^

This is how I feel about it.  Yeah, ND’s schedule turned out to be weak.  But they signed up to play Michigan, Virginia Tech, FSU, USC, Syracuse, Stanford, Pitt... That’s the kind of scheduling that should be rewarded, even if most of those teams are down this year.  

I don’t generally like the head-to-head comparison because ordinarily it’s used to compare teams with identical records—meaning the winner to the H2H matchup has another loss to somebody else.  So if ND loses a game then a 12-1 Michigan easily passes 11-1 ND despite the H2H because ND’s loss to Syracuse or USC is undoubtedly a worse loss than M’s loss to ND.  But right now ND is 12-0.  It’s not just that they beat M head-to-head, it’s that they are undefeated by virtue of that victory whereas M is a 1-loss team.  The rest of ND’s schedule would need to be deeply and indefensibly bad to even consider putting them behind Michigan in this scenario.  But ND’s schedule is easily defensible; they scheduled a bunch of traditionally strong teams and just happened to catch many of them in down years.  

NittanyFan

November 13th, 2018 at 3:35 PM ^

Playing Devil's Advocate --- aren't you incenting "margin of victory" in your paragraph on "comparative outcomes of common opponents?"  Which, per the Committee rules, you aren't supposed to do?

ND and Michigan will have one common opponent this year.  In the strictly binary sense (not incenting "margin of victory"): they both beat them.  Even if U-M beats Northwestern 100-0 in Indianapolis, they both will still have basically the same binary results (a neutral-site win theoretically means less that a road-win over Northwestern).

TrueBlue2003

November 13th, 2018 at 4:50 PM ^

I agree.  I read that as though the committee can compare outcomes against common opponents but not margins.  So if one team loses to a team that a comparable team beats, that comes would be considered. But since they both beat NW at NW, I'm not sure another neutral field win by M does anything.  They both beat NW and are better than NW.  That information is clear.

As others have mentioned, it would be really hard to see the committee put us ahead of ND if they're undefeated with a win over M.  Better to hope for them to lose.

I know we'd all lose our minds if the tables were turned and this whole board would scream favoritism and bias or whatever.

The Maizer

November 13th, 2018 at 3:42 PM ^

One thing overlooked in the head-to-head result thing is that even if you disregard the fact that they beat us and pretend that win is against a nameless team, it's still the best win any team has in the nation this year. Next best wins are LSU over Georgia and Texas over Oklahoma.

TrueBlue2003

November 13th, 2018 at 5:54 PM ^

My first response to this was that those are only the most impressive wins if you consider result against a given opponent without considering how the win occurred or where it occurred. 

I would have thought if you consider location and how a team achieved a given win, the most impressive win this year was definitely Alabama shutting out LSU at LSU.

FPI does a "game score" to determine how impressive a teams performance is in a given game and it turns out the three you listed are the highest ranked games I could find with game scores of 98 each.

Bama over LSU achieved a 97.

Michigan over PSU was a 96 and M over MSU was a 95 which surprised me for a 21-7 game.  Clearly takes into account yardage differential and everything.

 

S.D. Jones

November 13th, 2018 at 3:55 PM ^

While I think the committee isn't in love with ND and might like to drop 'em a spot, it won't do it. There's no plausible way to justify the move this week. They'd have some cover if they wait 'til we win the Big Ten championship, but right now it's inexplicable.

Personally, as long as ND's undefeated, I think they deserve to be ranked ahead of us. Head-to-head should mean something, and I'm pretty sure we lost. Also, imagine if the situation were reversed and we got bumped...

J.

November 13th, 2018 at 4:07 PM ^

Head to head means something, but it doesn't mean everything.

Otherwise, I guess Akron is better than OSU -- Akron beat NW who beat Purdue who beat OSU.

If you throw out the records against the bottom-dwellers and just look at quality wins, ND has one or two and Michigan has somewhere between two and four.  I mean, honestly, Northwestern is ND's second-best win.

S.D. Jones

November 13th, 2018 at 5:03 PM ^

I agree it doesn't mean everything. But it matters when there's a slim margin between two teams, as there is now.

The Akron-is-better-than-OSU argument is a strawman since it's based on the Kevin Bacon transitive property; even if it were an actual head-to-head comparison, it still wouldn't make Akron the better team because the Zips are 4-5 and the Bucks are 9-1.

Between ND and UM, the scales are far more balanced, however.

--ND has no losses, we have one. That would put us at a disadvantage even if they weren't the folks who gave us the L. We have a demerit on our record, and their slate is clean, simple as that.

--While us two may think UM has many more good wins, that's not how the committee will see it. As far as that panel's concerned, a "good win" is defined a victory over a squad currently in the top 25. Based on tonight's likely rankings, they'll have one (us) and we'll have one (PSU). If Sparty sticks, we get a slight boost, and if NW makes the list, it it's a wash.

--As for how we fared against common opponents, our respective NW games is at best a push for us.

--Strength of schedule is definitely in our favor, but it's not so large a contrast as to overwhelm the other evidence, especially the L-shaped elephant in the room.

All in all, we're not so far superior to ND that the head-to-head doesn't matter. In fact, it's probably the most important factor at this time.

J.

November 13th, 2018 at 5:22 PM ^

If Michigan is punished for playing a game against Notre Dame, instead of keeping their game against Arkansas, you may as well blow the whole system up.  It's already rare to get a quality out-of-conference matchup; teams -- including Michigan -- crow about their "one Power Five opponent," as if that's somehow enough.

If all they're going to do is rank teams by the number of losses -- Group of 5 excluded -- expect to see more and more cupcakes on the schedule going forward.  This hyper-fixation on the number of losses, without ever considering the opponents, is slowly ruining college football.  

J.

November 13th, 2018 at 7:35 PM ^

According to the wisdom of the board, which says that they can never be ranked higher than Notre Dame as long as Notre Dame hasn't lost, they're absolutely punished for playing the game.  If they hadn't played it, Notre Dame has no top wins, Michigan has a victory over an SEC doormat, and undefeated, #3 Michigan doesn't have to worry about Alabama / Georgia.

MgoHillbilly

November 13th, 2018 at 3:58 PM ^

I think it's possible if we dominate here on out and they simply take care of business.  The committee will put a one loss team in front of them as they've shown when lsu was first ranked. 

footballguy

November 13th, 2018 at 3:59 PM ^

Can we please stop with this? 12-0 ND will be ahead of 12-1 Michigan.

If ND jumped Michigan while Michigan was undefeated and Notre Dame has a loss, this place would burn down.

It ain’t happenin. 

mgobleu

November 13th, 2018 at 4:27 PM ^

This. ^^^^

If the tables were turned and any 12-1 team we beat jumped an undefeated Michigan, Ann Arbor torch & pitchfork would be sold out in seconds and the committee would be nervous as a donut in Charlie Weis' lunchbox. 

Beating teams should matter even if it was 12 weeks ago. 

S.D. Jones

November 13th, 2018 at 5:11 PM ^

I agree with that. An undefeated ND squad (or any non-UCF team) is hard to snub, but with one loss, the committee could bring in mitigating factors in our favor, saying their loss is worse because it's more recent, etc. Also, I think the committee could argue it wasn't so much judging a ND-UM horse race, but comparing all the one-loss teams as a group, so the head-to-head tiebreaker didn't come into play.

L'Carpetron Do…

November 13th, 2018 at 4:05 PM ^

None of this matters of course, if Michigan doesn't beat Ind, OSU and NW. I listened to Tirico bitch about this the other day on the ND broadcast and I hate to agree with him but he's right: a 12-1 Michigan should not jump a 12-0 ND, otherwise the head-to-head matchup doesn't mean anything.  Michigan is an improved team and everyone knows that but ND is playing well now too, after getting through rock fights vs the likes of Vandy and Pitt. A few weeks ago I thought  M would destroy ND in a rematch but now I think its more of a toss-up (although I would give M the edge and the game would be much different). But ultimately that head-to-head is ND's trump card (but a 11-1 ND team should be behind M for a number of factors).

butuka21

November 13th, 2018 at 4:10 PM ^

They win out they are in, they do not win out they are out.  Just beat Indiana first.  

I personally believe that if ND and Michigan were to play again it would be a 10+ point victory for Blue, and Michigan is a much better football team now then they were the 1st game.  I also taking my fandom out of it can't place Michigan in over an undefeated ND who did win head to head.  IF ND loses then I can see Michigan jumping them with a win over Ohio state and a Big ten Championship win over NW, but none of it matters if Michigan does not win out.  We need Alabama to win out.