NCAA Week 13 Top Forty

Lots of beatpaths news today. We found the new college football data, so we have the new Top Forty for Division 1-A.

2005-13-Ncaa40-Clean

Again, the teams that are floating by themselves are the ones that have no beatpaths to anyone else in the Top Forty. They all have plenty of beatpaths to the other Division 1-A teams. Since this is only a partial graph, a top row team isn’t necessarily better than teams in lower rows. The Beatpaths power rankings still have USC and Texas at the top.

Read on for the power rankings for the Top 25 teams, along with where they were ranked last week under this algorithm, and where the BCS has them now:

Rank Team Notes Last Week BCS BeatPower

1

USC (Beat Fresno St)

1

1

100.0

(375/375 – 0/375)

2

Texas (Bye)

2

2

100.0

(387/387 – 0/387)

3

LSU (Beat Mississippi) LSU now has alternate beatpaths to every team it has played.

5

4

100.0

(382/382 – 0/382)

4

Oregon (Beat Oregon St) The only team they’ve played that they don’t have a beatpath to is USC. However, they did get stomped by USC. I’m trying to think of if there’s a way to better account for that other than actually looking at point differential. There might still be reason to look at average beatpower of a team’s beatwins.

4

7

99.7

(364/365 – 1/365)

5

Virginia Tech (Beat Virginia)

12

5

99.7

(365/367 – 0/367)

6

West Virginia (Bye)West Virginia must have experienced some good luck from teams entering their beatpaths.

14

11

99.7

(344/345 – 1/345)

7

Fresno St (Lost to USC) Some impressive wins, and two competitive losses to teams in the top five.

9

19

99.4

(312/314 – 2/314)

8

Ohio St (Beat Michigan) These guys screwed up Penn State.

7

6

99.4

(356/359 – 1/359)

9

TCU (Bye)

17

13

99.7

(318/320 – 0/320)

10

Miami FL (Lost to Georgia Tech)

11

9

99.4

(357/361 – 0/361)

11

Penn St (Beat Michigan St) Penn State lost to Michigan earlier this year. Ohio State beat Michigan this week, which means Penn State lost credit for beating Ohio State.

6

3

99.4

(336/340 – 0/340)

12

Louisville (Bye)

15

20

99.3

(339/342 – 2/342)

13

UCLA (Bye)

8

12

99.7

(384/386 – 0/386)

14

Colorado (Bye)

13

99.3

(343/346 – 2/346)

15

Texas Tech (Beat Oklahoma)

16

16

99.6

(371/373 – 1/373)

16

Georgia Tech (Beat Miami FL)

24

21

99.2

(354/360 – 0/360)

17

Auburn (Beat Alabama)

25

10

99.4

(353/355 – 2/355)

18

Alabama (Lost to Auburn) Yikes!

3

14

99.1

(327/330 – 3/330)

19

Oklahoma (Lost to Texas Tech)

19

98.9

(370/375 – 3/375)

20

Georgia (Beat Kentucky)

26

15

98.8

(334/339 – 3/339)

21

Wisconsin (Bye)

20

17

99.4

(349/353 – 0/353)

22

Clemson (Beat South Carolina)

42

99.1

(319/325 – 0/325)

23

Minnesota (Lost to Iowa)

21

98.7

(345/351 – 3/351)

24

Michigan (Lost to Ohio St)

29

22

98.8

(315/323 – 0/323)

25

Notre Dame (Beat Syracuse)

31

8

99.2

(317/321 – 1/321)

26

South Carolina (Lost to Clemson)

28

98.1

(303/310 – 5/310)

27

Boston College (Beat Maryland)

32

18

98.7

(301/308 – 1/308)

28

Florida (Bye)

18

23

98.2

(303/310 – 4/310)

29

Virginia (Lost to Virginia Tech)

30

98.7

(256/262 – 1/262)

30

NC State (Beat Middle Tenn St)

49

98.5

(266/273 – 1/273)

31

North Carolina (Beat Duke)

27

97.3

(263/273 – 5/273)

32

Nebraska (Bye)

36

98.4

(339/346 – 4/346)

33

Kansas (Bye)

22

97.8

(314/322 – 6/322)

34

Texas A&M (Bye)

43

97.5

(272/282 – 4/282)

35

Florida St (Bye)

45

24

97.6

(137/144 – 0/144)

36

Iowa St (Bye)

39

97.8

(331/341 – 5/341)

37

Kansas St (Beat Missouri)

46

96.3

(257/269 – 8/269)

38

Texas-El Paso (Lost to UAB) UTEP is finally exposed. Actually, this was probably an upset.

10

99.7

(314/316 – 0/316)

39

Tennessee (Lost to Vanderbilt)

33

96.2

(255/266 – 9/266)

40

Maryland (Lost to Boston College)

23

94.4

(81/89 – 2/89)

11 Responses to NCAA Week 13 Top Forty

  1. Wes M says:

    While the final score of the USC-Oregon game shows a blowout, the game was fairly close for a long time, and Oregon had the momentum pulled out from under them by TWO touchdowns taken away by the officials. (And one of those was given an “our bad” by the PAC 10 later.)

    Note: I’m an OSU person, defending the Ucks is almost painful.

  2. Pat says:

    Wait, what impressive wins does Fresno State have? They haven’t beaten anyone in the Top 40 (nor in anyone else’s Top 40 – their best win is over Boise St.).

    I’m … really not sure about the algorithm for handling Division I-AA teams. Is the fraction of Div IAA teams in the top 40’s beatpaths pretty much constant for all of them?

    It just seems like there are far too many teams with beatlosses way high up in the rankings. I think the huge number of teams is diluting the ability of beatlosses to push down a ranking. Losing 1/300 isn’t exactly a big penalty.

  3. Pat says:

    Oh, and the conference graphs are still useful even if it neglects out-of-conference loops, in my opinion. The conferences themselves are so ridiculously better connected than out-of-conference games that I think the in-conference graphs might be more representative in certain cases, because it gives teams a chance to prove a beatloss/beatloop wrong.

    Plus they’re neat looking.

    One thing I’m thinking about doing is writing a program to convert beatloops into animated GIFs, and doing the season with them. Granted, you’d prefer to animate it with the teams moving around, but, uh, that would require a change in graphviz, I think.

  4. ThunderThumbs says:

    I don’t really know much about Fresno State – I guess I just assumed. All I know is that if anyone else has beaten them other than USC/Oregon, they’ve beaten someone else impressive enough to get alternative beatpaths to them and a whole bunch of other teams.

    I’m not sure what you’re asking about Division I-AA ? If any of these teams do suffer a beatloss to a Div IAA team, it does get counted in their stats. That’s why Stanford dropped out of the Top 20 when I made that adjustment.

    The main way the beatlosses control things is that it’s still impossible for a team to be ranked ahead of another team that has a beatwin to it. I actually kind of like that because it seems kind of silly that a number #3 team should tumble way down the rankings if they lose to #1. The only way a team’s beatloss number would be low is if they are beaten by a very excellent team. Otherwise, it really is possible for a team to get beaten once by a mediocre team, and then have a beatloss number up in the 100’s.

  5. ThunderThumbs says:

    I should be pretty well set up to generate the conference graphs now – I might try and go through them tonight.

    I know, I’ve been wanting to animate, too – maybe there’s another package out there that can actually read a series of digraph files and do something nifty with them.

  6. ThunderThumbs says:

    I’ve been looking into it – it looks like the best way would be to create a graph with the nodes absolutely positioned, and then make the edges visible and invisible.

    I should be able to do it with the NFL, for a weekly rundown showing all the game results. Then at the end you just show the new layout.

  7. Pat says:

    I’m not sure what you’re asking about Division I-AA ?

    Well, the main difference between this week and last week is the inclusion of all teams, which means instead of being connected to ~70 teams, you’re connected to ~300. That’s what I’m not sure about, because it seems to heavily dilute the contributions of beatlosses.

  8. jstnrgrs says:

    I just discovered your website, and I am impressed by what you have done. Is there any reason that your system could not rank all 702 college football teams that play connecting schedules? That would be a great graph to see.

  9. jstnrgrs says:

    nevermind, I’ve looked through the site a little more now.

  10. jstnrgrs says:

    Any chance we’ll see an updated top 40 now that all CFB games are complete?

  11. ThunderThumbs says:

    Sure thing – the new final top 40 is here.

Leave a Reply

Your email address will not be published. Required fields are marked *