Top 25 rankings

[Mr. Ypsi]

Ironic that you use just OWP and call is SOS.  That widely discredited metric is (essentially) the late, unlamented QOWI.  This year they have moved to a much more informative combination of OWP and OOWP - not just how what were your opponents' records, but what were THEIR opponents' records.

Record alone (while it may or may not be the SINGLE best indicator) does not tell the whole story.  At the extreme, an 0-10 team could be the 11th best in the country (if their opponents are 1 thru 10), while a 10-0 team could be the 11th worst!  Obviously, in real life, a 10-0 team is likely to be better than an 8-2 team, but by no means is that always true.

[redswarm81]

Ironic that you use just OWP and call is SOS.  That widely discredited metric is (essentially) the late, unlamented QOWI.  This year they have moved to a much more informative combination of OWP and OOWP - not just how what were your opponents' records, but what were THEIR opponents' records.

Which "that" widely discredited metric?  I didn't mention last year.  I was talking about OWP/SoS.

As I understand it, OWP is entirely different from the unrelated-to-anything, out-of-the-clear-blue-sky, Playoff Handbook-originated Quality of Wins Index.  Opponents' Winning Percentage is synonymous with Strength of Schedule.  There is no QoWI in 2007, is there?

It certainly is possible, that a 0 - 10 team is the 11th best team in the country.  If there are only 11 teams in the country, then comparison of records reaches that result more easily than by SoS analysis.

[Ralph Turner]

No QOWI in 2007.

[K-Mack]

As of today (end of week 8, right?), how many undefeated teams are there in D3?  Anyone?  Anyone?  Bueller?

16. And 17 winless.

The rest of my contribution to this discussion will be in ATN though. 

[Ralph Turner]

The Top 25 is showing considerable consensus among the top 6 schools.

1   Mount Union (25)   7-0  625 (Perfect #1 vote tally)

2   Mary Hardin-Baylor   7-0  596  (-4 from perfect #2)

3   UW-Whitewater           6-1  575 (Perfect #3)

4   St. John's                   8-0   549 (-1 vote from perfect #4)

5   Wheaton (Ill.)                7-0   521  (-4 votes from perfect #5)

6   Central                           8-0   494   (-6 votes from perfect #6)


The second tier of teams is the cluster around #7- #9.  A perfect #7 receives 475 votes; #8  -- 450; #9 -- 425.

7   St. John Fisher                    7-1      423

8   Washington and Jefferson       6-0      414

9   Salisbury                              8-0       411



Wesley comes in a solid #10.  A perfect #10 will receive 400 votes.  Wesley's 376 is closer to a perfect #11, but there is a drop-off to the next cluster of teams between #11 and #13.  Perfect #11 , #12 and #13 should receive 375 votes, 350 votes and 325 votes, respectively. 
10   Wesley                                     7-1       376

11   Alfred                                        7-0       343

12   Capital                                       6-1       337

13   Wabash                                    7-0        334


Occidental at #14 is clearly a step below in the minds of the voters.  A perfect #14 should get 300 votes, so the voters are telling us that Oxy "seems" like a #16 to them.
   
14   Occidental                             6-0        251 

Top 25 -- Week #8

[DutchHawk]

Awesome analysis Ralph!! I never knew that was how it worked with the perfect votes and all.

Half of the Top 6 in the West region...how does that play out if all are 10-0?

[redswarm81]

The Top 25 is showing considerable consensus among the top 6 schools.

1   Mount Union (25)   7-0  625 (Perfect #1 vote tally)

2   Mary Hardin-Baylor   7-0  596  (-4 from perfect #2)

3   UW-Whitewater           6-1  575 (Perfect #3)

4   St. John's                   8-0   549 (-1 vote from perfect #4)

5   Wheaton (Ill.)                7-0   521  (-4 votes from perfect #5)

6   Central                           8-0   494   (-6 votes from perfect #6)


The second tier of teams is the cluster around #7- #9.  A perfect #7 receives 475 votes; #8  -- 450; #9 -- 425.

7   St. John Fisher                    7-1      423

8   Washington and Jefferson       6-0      414

9   Salisbury                              8-0       411



Wesley comes in a solid #10.  A perfect #10 will receive 400 votes.  Wesley's 376 is closer to a perfect #11, but there is a drop-off to the next cluster of teams between #11 and #13.  Perfect #11 , #12 and #13 should receive 375 votes, 350 votes and 325 votes, respectively. 
10   Wesley                                     7-1       376

11   Alfred                                        7-0       343

12   Capital                                       6-1       337

13   Wabash                                    7-0        334


Occidental at #14 is clearly a step below in the minds of the voters.  A perfect #14 should get 300 votes, so the voters are telling us that Oxy "seems" like a #16 to them.
   
14   Occidental                             6-0        251 

Great stuff, Ralph.  As I'm sure you know, the potential for erroneous conclusion of your analysis increases as you move past the top few spots.  By the time you reach mid-poll, comparison to "perfect" scores is nearly impossible to gauge.¹  As a simple example, in a top 25 vote, if Voter A picked the Screaming Chickens #1, and Voter B picked the same team #25, then the Screaming Chickens would have a "perfect #13 vote tally," with very little consensus.

¹  The fact that the top teams in this week's poll do in fact have a pretty clear consensus makes my simple example impossible, but you get the idea.

[Ralph Turner]

Redswarm, your example is clearly borne out in the East Region Fan Poll and the South Region Fan Poll that are being conducted on these boards!  The fact that the poll conductors are sharing the vote tallies is very instructive.

There is much less consensus as we get farther down the poll.

I was particularly struck by how "tightly" the first 6 teams were voted this week.   

[redswarm81]

Redswarm, your example is clearly borne out in the East Region Fan Poll and the South Region Fan Poll that are being conducted on these boards!  The fact that the poll conductors are sharing the vote tallies is very instructive.

There is much less consensus as we get farther down the poll.

I was particularly struck by how "tightly" the first 6 teams were voted this week.   

Agreed on all points.  Great stuff.

[smedindy]

It's like the Top 25 is a parfait, with all of the good stuff at the top, though.

[Toby Taff]

So if UWW beats MHB by 1, lets say a 7-8 defensive struggle, UWW wins on a 2pt conversion in OT.  DO they swap spots or does someone else jump them?

[usee]

So if UWW beats MHB by 1, lets say a 7-8 defensive struggle, UWW wins on a 2pt conversion in OT.  DO they swap spots or does someone else jump them?

I would think they swap spots and/or they both stay in the top 5 unless there is a decent blowout. If UMHB wins by a large margin I would guess UWW would drop out of top 10. If UWW wins by a large margin I would guess UMHB wouldn't drop as far.

[Just Bill]

As a simple example, in a top 25 vote, if Voter A picked the Screaming Chickens #1, and Voter B picked the same team #25, then the Screaming Chickens would have a "perfect #13 vote tally," with very little consensus.

Wait!  You mean my beloved Screaming Chickens AREN'T #1?  What's with all the anti-poultry bias around here.

[Toby Taff]

As a simple example, in a top 25 vote, if Voter A picked the Screaming Chickens #1, and Voter B picked the same team #25, then the Screaming Chickens would have a "perfect #13 vote tally," with very little consensus.

Wait!  You mean my beloved Screaming Chickens AREN'T #1?  What's with all the anti-poultry bias around here.

My poultry bias is based on the fact that I'm a vegetarian.

[Pat Coleman]

If UMHB blows out UWW I would expect UMHB might pick up a first-place vote or two and UWW would slide a little. But frankly, other than St. John's, would we have confidence that teams are actually better than UWW? I'm not sure. It would just mean that UWW played two better teams than other teams have played.

Is it Capital, which would likely have two losses after playing Mount Union?
Is it Central, which keeps winning but dominates nobody?
Is it Wheaton, which is playing its third string quarterback and has three defensive ends out?
Is it St. John Fisher, which lost at Hartwick?
Is it Washington & Jefferson, whose opponents' winning percentage is in the low .400s?

That's the problem -- I would have a hard time believing all of those teams are better than UWW in this situation.