"Spend a few minutes reading College Football Resource" - Whit Watson, Sun Sports

"Maybe you should start your own blog" - Bruce Feldman, ESPN

"[An] Excellent resource for all things college football. It’s blog index is the definitive listing of the CFB blogosphere ... [A] must-read for fans." - Sports Illustrated (On Campus)

"The big daddy of them all, the nerve center of this twisted college football blogsphere" - The House Rock Built

"Unsurprisingly, College Football Resource has generated some discussion" -Dawg Sports

Top Teams 2008

After Week Seven

  1. Alabama
  2. Penn State
  3. Texas
  4. Oklahoma
  5. Florida
  6. USC
  7. Georgia
  8. LSU
  9. BYU
  10. Missouri
  11. Ohio State
  12. Oklahoma State
  13. Texas Tech
  14. Utah
  15. Kansas
  16. USF
  17. North Carolina
  18. Miami
  19. Boise State
  20. Georgia Tech
Display
RSS
Search CFR
Submission Corner
« Some teams evolve, others devolve | Main | Unpredictable teams, and risk/reward »
Wednesday
Nov162005

Justifying the rankings

I see this week SI's Stewart Mandel is taking some of the same rounds of questioning about his rankings that I've faced doing my rankings.

Mandel's replies are pretty interesting and instructive.  Be sure and read the entire thing.

Some of the best back-and-forths:

 

Help me understand something. Notre Dame continues to be ranked ahead of such one-loss teams as Alabama, Virginia Tech and Oregon. Why?

 


--Jonathan, Baton Rouge, La.

Probably because they're playing better than those teams. The Tide's offense has gone in the toilet, the Ducks have barely survived their two games since losing starting quarterback Kellen Clemens and the last time we saw the Hokies, they were on the wrong end of a 27-7 score. If you'd prefer, I could start ranking teams in strict order of their records, in which case I'll start letting our interns fill out my ballot each week.

 

 

Classic. Who... Is... Best...? You go first. Who... Is... Next...? You go after them.

 

Notre Dame lost to Michigan State (5-5, 2-5 in the Big Ten). How on earth can the Irish be ranked as high as they are, and why doesn't anyone ever bring that up? It's like people forgot that game.

 


--Kyle, Houston

You know what's interesting? Despite Miami moving up to No. 3 in the BCS standings this week -- a development that has far more potential ramifications than whether Notre Dame is sixth or ninth -- I haven't heard a single person say, "Hey, have you people forgotten that the 'Canes lost to a Florida State team that got blown out by Clemson?" I'd imagine that's because most people generally accept the notion that Miami has gotten significantly better since that game, while the 'Noles have regressed. That's pretty much how a season works -- some teams evolve from the first game to the last game, others devolve.

But apparently, judging by my e-mail, that's not the case with Notre Dame and Michigan State. Apparently the Irish are exactly the same team today that they were two months ago, and apparently the Spartans were just as bad when they started 4-0 and were scoring 49 points per game on people than they are now that they've lost five of six and couldn't crack 20 on Minnesota. I suppose it's not remotely possible that, say, the Irish defense is substantially more comfortable after nine games under a new coaching staff than it was after two, or that the more games the Spartans played, the more scouting opportunities opposing defensive coordinators had to figure out their spread offense. I'm not trying to make excuses for what was clearly a bad loss; I just feel that a game played on Sept. 17 has little-to-no-bearing on how a team is playing on Nov. 17.

Thank you! Remember the whole twisty Michigan/Michigan State/Notre Dame pretzel earlier this year? Things have resolved themselves nicely, now haven't they?

So much confusion is created because fans have this idea that teams should be ranked by order of loss and schedule strength and conference strength. Problem is, those inputs don't give much weight to determining just how good a team is, and what its real world value is. We're obsessed with trying as hard as possible to keep human factors out of our determinations, when the greatest tool available in ranking teams is not in their numbers or records, but in what our eyes see and what our brains process. Let people rank teams, let them get it right, even if mistakes are made. The current agnostic approach is causing such ridiculous confusion that when someone like Mandel comes along with half a mind to put some observation and thought into his rankings, people flip out.

You know, he may end up being wrong, but at least he made better use of the available tools in making his rankings. 

PrintView Printer Friendly Version

EmailEmail Article to Friend

Reader Comments (4)

Sorry to seem like I'm beating a dead horse here, but I think Mandel's argument against Miami is a terrible one, for a simple reason:

UM has a 'signature' victory. Notre Dame has a signature LOSS.

In his point about Virginia Tech: "the last time we saw the Hokies, they were on the wrong end of a 27-7 score." You can't say that and then not mention that UM was on the RIGHT side of said score.

True, the loss to the Noles wasn't a very good one. But Miami has since not only not lost again (as opposed to having 2 losses, which is a difference in and of itself), they beat the team some pundits were claiming should have been ranked ahead of Texas... up until they got shelled in Blacksburg. The only way I will be convinced that ND has anything close to a win anywhere near the caliber of Miami's blowout over VT is if Michigan can beat Ohio State.

If they don't, ND will have one close victory over a 7-4 team as their best win, and have their best game be a close loss at home against a team that has allowed 'close' losses to such illuminaries as Stanford or UCLA last season.

It's fine if you want to say that no game played in early September has a bearing on judging a team in November... I STILL can't see the rationalization for placing ND so high with blowout victories over BYU, a tailspinning Tennessee team, or the MIDSHIPMEN! Whoooooo!

Plus, on his point about MSU starting so well: aside from the ND win, MSU beat teams that are currently: 4-6 (Hawaii in the WAC), 1-9 (Kent State in the MAC), and 2-8 (Illinois in the Big 10). Come. On. That's 3 absolutely horrific teams, and the offenses in the SEC would have been running up 40 points a game on them. At least.

Michigan State benefitted from the media glorification of Notre Dame - "we love ND, they must be good, so if this team beat ND they must have been good at some point. No way could we revise our analysis concerning the fact that the Irish lost to the Spartans at home. Nope. Must be the Spartans just are getting outcoached now - despite the fact that Charlie Weis is a supposed miracle worker and should have been able to figure out the spread offense with merely a glance."

Really, come on. One loss to rival Florida State (who, even with the tailspin, are still 7-3 and will almost certainly win the ACC Atlantic division) in the first game of the season, AT FSU, is in no way comparable to dropping a home game several games into the season to an MSU team that feasted on horrible competition early in the year and has been exposed by the better Big-10 teams since.
November 16, 2005 | Unregistered CommenterUnderbruin
But he wasn't comparing Miami to ND. I think he has Miami ahead of ND (correct me if I'm wrong). He was using Miami as an example of his logic on why people's picking on ND's loss to Mich St. doesn't matter now. In other words, he's not talking about the teams' wins, he's talking about their losses. And I do think, considering that it was Weis's second/third game with his new team, that the loss to Mich St. is an anomaly to this current version of ND and has no value in determing how good ND is now, just as the lost to FL St. has not value in determining how good the Canes are now.

By the way, I'm not a ND fan.
November 16, 2005 | Unregistered CommenterTom
This brings up a really good point. Either the rankings are about who has the best record (ie, rank teams based on win/loss) or they are an independant mark of who are the best teams. If the latter is the case, then they should function as a better predictor of future performance than simple w/l + strength of schedule. In other words, either the rankings are a better predictor of future performance or the w/l records are. That's testable, for someone who has lots of free time. Are humans really better at predicting future performance, or can simple statistics do a better job?
November 16, 2005 | Unregistered CommenterAdam
LD hasn't crunched the numbers for Pollsters v. Records in predictions as Adam requests, but he does list a ranking of teams by W/L + head-to-head + SOS (using Colley, one of the BCS computers).

http://gunslingers.blogspot.com/2005/11/lebowski-rankings-week-6.html

Maybe we should take a page from golf and call it a leaderboard rather than a ranking. It's based more on what you've done than on how you look doing it. College football is about performance not popularity, but the latter is the way the championship game was chosen last year. It doesn't look like there will be a controversy this year, but I definitely believe that a win should count more than a "good" loss. Therefore ND is behind all 1-loss teams [and at least a few 2 loss-teams] on the leaderboard.
November 17, 2005 | Unregistered CommenterThe General

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
Post:
 
All HTML will be escaped. Hyperlinks will be created for URLs automatically.