clock menu more-arrow no yes

Filed under:

Thermhere's Weekly CFP Rankings Rant (Week 14)

New, 8 comments

Do you guys like Europe? Let's go to Monte Carlo.

Tim Heitman-USA TODAY Sports

"Not Necessarily a Bad Loss"

I'm not going to waste any time or words this week on the committee, the process or the relative strength of the teams (except for one note below).  I basically said my piece on that last week.  Nothing's really changed since then, so let's not waste our breath.  A bad process isn't going to get fixed midstream, Baylor doesn't have an advocate in the room, and the committee is finding ex post facto explanations for their rankings after they produce them, rather than using logic and reason to rank the teams.

I can't help myself on one thing: the Committee's bizarre defense of Ohio State's loss to Virginia Tech.

The following is a quote Jeff Long gave after the Committee released its first set of rankings:

Seems like a pretty unequivocal, unqualified statement, right?  At the time, Virginia Tech was 4-4 with 2 wins over Power 5 teams (including Ohio State).  Since then, Virginia Tech lost to 6-5 Boston College, beat 8-3 Duke, and lost to 3-8 Wake Forest (who has only 1 other win against FBS competition, and no other wins against Power 5 teams).

Somehow, that 1-2 record, including a 6-3 loss to one of the worst Power 5 teams in which Virginia Tech scored no points in regulation, has led Jeff Long to say the following about the Virginia Tech loss:

You might suggest that beating Duke should offset the horribleness of VT's loss to Wake.  That's not an entirely crazy position - but 8-3 Duke has only defeated ONE bowl eligible team (Georgia Tech).  That's going to increase by 2 as a matter of scheduling this weekend (either Virginia or Virginia Tech has to defeat the other) - but Duke's 8-3 record is pretty illusory in all practicality.

The loss, at home, by two touchdowns, to Virginia Tech by Ohio State is a terrible loss.  Jeff Long's defense of it, as soft as it may be, is patently ridiculous.

Crystal Ball

Last week, I made reference to a Monte Carlo simulation that I'd been running to try and assess Baylor's playoff odds.  This week, I've run it again in a couple of different scenarios.  I also ran 3 separate sets of probabilities this week - one based on my personal power ratings (which take into account only (1) who you played and (2) the score, and give equal weight to every game), one placed on Football Outsiders' F/+ team rating, and one based on Brian Fremeau's FEI.  I was able to incorporate simulations based on F/+ and FEI because Football Study Hall and Mr. Fremeau have each published their game projections for this week.  F/+ projections are available here at Football Study Hall, and FEI projections are available here on Brian's website.  A smart observer will note that neither F/+ nor FEI projections are available for conference title week.  I reached out to both Bill Connelly (the keeper of F/+) and Brian Fremeau to see if they could supply those for me, and didn't get a response, so I have done some guesstimating of their projections for conference title week based on the differences in perceived team strength between my ratings and each of their ratings.  If I get a response, I'll try and update the post.

I don't address the teams ranked 8 and higher below (other than noting the odds that one of those teams hops up into the top 4 in the rows labeled "Others") because there are too many variables.  Next week, I'll expand the simulation to include all possible playoff teams.

The outlook for Baylor ranges from dicey to pretty good, depending on how you feel like the Committee is going to sort through the Baylor/TCU head to head debate and the Mississippi State conference strength vs. Baylor/TCU/OSU conference title question.

Win Out Percentages

Today, ESPN tweeted out the odds that they believe each team will win out based on their FPI metric:

For reference purposes, here's a table showing the Win Out Percentages for the top 7 teams alongside those ESPN projections using my ratings, F/+ and FEI.

Therm F/+ FEI ESPN
Alabama 39.46% 49.46% 49.39% 46.00%
Oregon 66.50% 64.12% 58.18% 60.00%
Florida State 25.03% 54.69% 52.98% 51.00%
Mississippi State 34.85% 38.65% 39.78% 36.00%
TCU 70.74% 88.95% 88.52% 65.00%
Ohio State 45.31% 66.30% 59.40% 55.00%
Baylor 67.65% 67.77% 73.18% 71.00%

You can see from the table that my ratings track the others with a couple exceptions - my ratings are much higher on Florida (and therefore see an FSU loss as much more likely), my ratings and ESPN are a little lower on TCU winning out because we each see Texas as a little stronger than F/+ and FEI, and my ratings are slightly lower on Ohio State because they see Wisconsin as a pretty strong team.

The takeaway here, given that none of the above teams play each other, is that it's fairly unlikely (although not impossible) that everyone wins out.  My ratings give it only a 0.5% chance (1 in 200), F/+ only gives it a 2.5% chance (1 in 40), FEI only gives it a 2.3% chance (1 in 43), and ESPN only gives it a 1.3% chance (1 in 78).

Poll Inertia Case

I don't believe this to be true, but let's say that the Committee is stuck in complete "poll inertia."  By that, what I mean is that the assumption for the following runs of the Monte Carlo simulation is that absent a loss, a team is not moving down, absent a loss by a higher ranked team, a team is not moving up, and teams are going to move up in the same order in which they are currently ranked.  I'll discuss below why I think these are bad assumptions, but this is how some people feel, so I think providing the outcomes here will be useful for those people.

Therm Ratings

CFP Rank Wins Losses Playoff Rank
Alabama 1 11.16 1.69 39.4% 4
Oregon 2 11.65 1.35 66.4% 2
Florida State 3 12.03 0.97 26.5% 8
Mississippi State 4 10.50 1.65 34.8% 6
TCU 5 10.69 1.31 68.5% 1
Ohio State 6 11.40 1.60 38.9% 5
Baylor 7 10.65 1.35 47.9% 3
Others 77.5%

F/+ Ratings

CFP Rank Wins Losses Playoff Rank
Alabama 1 11.34 1.56 49.3% 4
Oregon 2 11.63 1.37 64.2% 2
Florida State 3 12.52 0.48 55.4% 3
Mississippi State 4 10.48 1.62 38.4% 6
TCU 5 10.88 1.12 82.1% 1
Ohio State 6 11.65 1.35 46.0% 5
Baylor 7 10.67 1.33 30.8% 7
Others 33.8%

FEI Ratings

CFP Rank Wins Losses Playoff Rank
Alabama 1 11.34 1.55 49.7% 4
Oregon 2 11.57 1.43 58.4% 2
Florida State 3 12.50 0.50 53.8% 3
Mississippi State 4 10.51 1.60 39.8% 6
TCU 5 10.89 1.11 82.5% 1
Ohio State 6 11.57 1.43 41.2% 5
Baylor 7 10.73 1.27 37.0% 7
Others 37.8%

Even using my power ratings (which think most highly of Baylor), we've only got just under a 50% shot at making the playoff if you assume rankings inertia has set in and won't change.  F/+ is particularly scary.

All of that said - the above reflects the possibility that we'll lose.  If you assume that we win out (and this discussion is pretty damn irrelevant if we don't), things do look somewhat better:

Therm Ratings (Win Out)

CFP Rank Wins Losses Playoff Rank
Alabama 1 11.15 1.69 38.9% 5
Oregon 2 11.65 1.35 66.3% 3
Florida State 3 12.02 0.98 25.7% 7
Mississippi State 4 10.50 1.65 34.8% 6
TCU 5 10.70 1.30 68.5% 2
Ohio State 6 11.40 1.60 39.1% 4
Baylor 7 11.00 1.00 72.3% 1
Others 54.5%

F/+ Ratings (Win Out)

CFP Rank Wins Losses Playoff Rank
Alabama 1 11.33 1.56 48.7% 4
Oregon 2 11.63 1.37 63.9% 2
Florida State 3 12.52 0.48 55.1% 3
Mississippi State 4 10.48 1.62 38.0% 7
TCU 5 10.88 1.12 81.9% 1
Ohio State 6 11.65 1.35 46.0% 6
Baylor 7 11.00 1.00 46.2% 5
Others 20.3%

FEI Ratings (Win Out)

CFP Rank Wins Losses Playoff Rank
Alabama 1 11.34 1.56 48.7% 5
Oregon 2 11.57 1.43 58.7% 2
Florida State 3 12.50 0.50 54.3% 3
Mississippi State 4 10.51 1.61 39.4% 7
TCU 5 10.89 1.11 82.9% 1
Ohio State 6 11.58 1.42 41.7% 6
Baylor 7 11.00 1.00 50.1% 4
Others 24.3%

Winning out makes the ranking inertia situation look a little less scary.  If the Committee has its head dug in the sand, even F/+ and FEI see Baylor as a 50/50 shot to make the playoffs (while my power rankings see it as 72% shot).

My Predicted Scenario

As opposed to the static scenario above, where winning guarantees a team doesn't drop, while I still don't think there's any way Alabama, Oregon or FSU drop below 3 if they win out, I think the most likely way that the committee sorts out the 4-7 mess is as follows:

  • If Ohio State wins out, it will be ranked ahead of all 3 other teams (unless Mississippi State makes and wins the SEC Championship Game)
  • If Baylor wins out, it will be ranked behind ahead of TCU and Mississippi State (unless Mississippi State makes and wins the SEC Championship Game)
  • If TCU wins out, it will be ranked behind Ohio State and Baylor, but ahead of Mississippi State (unless Mississippi State makes and wins the SEC Championship Game)
  • If Mississippi State beats Ole Miss but Alabama beats Auburn (which would be Alabama in the SEC Championship Game), Mississippi State will be ranked at the back of the 4 team group.  If Mississippi State beats Ole Miss, Bama stumbles against Auburn, and then MSU beats its SEC East opponent in the SEC Championship Game, it will (and frankly, probably should) jump to the front of this group

I've got a couple of reasons for my belief on how things will shake out.  I don' t think it's an accident that Jeff Long continues to bring up the fact that conference championships will probably influence the final order.  I also think he's been fairly careful (if not wrong) when talking about Baylor and TCU.  Specifically, he's continued to say that Baylor's resume doesn't match TCU's yet, and therefore the HTH doesn't come into play.  Baylor winning out would change that, and the committee would be forced to jump Baylor over TCU.  I agree with Aqua that the Committee's rankings (and Jeff Long's explanations) so far can't lead you to any other belief than that Ohio State is going to jump the other teams if it wins the Big 10.  I also think ultimately they're not going to put Mississippi State into the final 4 over a 1 loss conference champion from a Power 5 conference.  If Mississippi State had played anyone of substance in its non-conference (and don't come in here with your comments about Baylor's nonconference schedule - it is equally bad, yes), I think the Committee might be more sympathetic to them.  But 11-1 Mississippi State without the SEC title game would only have played 8 Power 5 opponents, compared to everyone else's 9 minimum (and 10 or 11 in several cases).  I just don't see that happening.

So with that in mind, below are the simulations assuming the Committee would order the teams as I've indicated above.

Therm Ratings

Pecking Order Wins Losses Playoff Rank
Alabama 1 11.16 1.69 39.2% 5
Oregon 2 11.65 1.35 66.1% 1
Florida State 3 12.02 0.98 26.1% 7
Ohio State 4 11.40 1.60 45.4% 4
Baylor 5 10.66 1.34 65.0% 2
TCU 6 10.70 1.30 58.2% 3
Mississippi State 7 10.51 1.64 23.0% 8
Others 77.0%

F/+ Ratings

Pecking Order Wins Losses Playoff Rank
Alabama 1 11.33 1.57 48.5% 6
Oregon 2 11.63 1.37 64.4% 2
Florida State 3 12.52 0.48 55.3% 5
Ohio State 4 11.66 1.34 66.4% 1
Baylor 5 10.67 1.33 59.4% 3
TCU 6 10.88 1.12 56.0% 4
Mississippi State 7 10.48 1.62 15.4% 7
Others 34.7%

FEI Ratings

Pecking Order Wins Losses Playoff Rank
Alabama 1 11.33 1.56 49.0% 6
Oregon 2 11.57 1.43 58.3% 3
Florida State 3 12.50 0.50 54.0% 5
Ohio State 4 11.57 1.43 58.2% 4
Baylor 5 10.73 1.27 66.4% 1
TCU 6 10.88 1.12 59.7% 2
Mississippi State 7 10.52 1.60 16.5% 7
Others 38.0%

This doesn't look nearly as bad as the rankings inertia scenario, right?  We're anywhere from a 3/5 to 2/3rds shot to make the playoff if you assume the committee applies at least some logic (as well as follows its stated guidelines to reward conference champions and value head to head matchups).  [/Someone in the comments is definitely going to come with the "This is all great, but the committee won't apply any logic response.  Your comment is noted in advance.]

As above, things look significantly better if you assume the Bears win out (and hoping to win out is the only reason we're having discussions about the Committee in the first place):

Therm Ratings (Win Out)

Pecking Order Wins Losses Playoff Rank
Alabama 1 11.16 1.69 39.1% 5
Oregon 2 11.65 1.35 66.3% 2
Florida State 3 12.02 0.98 26.1% 6
Ohio State 4 11.40 1.60 45.8% 4
Baylor 5 11.00 1.00 96.3% 1
TCU 6 10.70 1.30 53.4% 3
Mississippi State 7 10.51 1.64 19.4% 8
Others 53.5%

F/+ Ratings (Win Out)

Pecking Order Wins Losses Playoff Rank
Alabama 1 11.34 1.56 49.2% 5
Oregon 2 11.64 1.36 64.9% 2
Florida State 3 12.53 0.47 55.9% 4
Ohio State 4 11.65 1.35 66.0% 2
Baylor 5 11.00 1.00 86.7% 1
TCU 6 10.89 1.11 46.0% 6
Mississippi State 7 10.49 1.62 12.0% 7
Others 19.2%

FEI Ratings (Win Out)

Pecking Order Wins Losses Playoff Rank
Alabama 1 11.33 1.56 49.0% 6
Oregon 2 11.56 1.44 58.1% 2
Florida State 3 12.50 0.50 54.4% 4
Ohio State 4 11.57 1.43 58.1% 3
Baylor 5 11.00 1.00 90.1% 1
TCU 6 10.89 1.11 52.2% 5
Mississippi State 7 10.51 1.61 13.3% 7
Others 24.9%

Even FEI, which is most negative, only sees us as having a 10% shot to miss the playoffs if we win out and the committee uses their stated principles to rank the teams (meaning in 10% of simulations, Alabama, Oregon, Florida State and Ohio State all win out).

All of the above notwithstanding, like everybody else, I will sleep a lot easier next week if two or more of Auburn, Florida, Michigan, Texas or Ole Miss can pull off upsets this week.