2007 Week 4 Beatpaths Rankings

There were a bunch of upsets this week, which means that the graph is changing a lot.

2007-4-Nfl-Clean

There are some three-team beatloops this week:

  • CAR=>ATL=>HOU=>CAR
  • ARI=>SEA=>SF=>ARI
  • ARI=>PIT=>SF=>ARI
  • CHI=>KC=>SD=>CHI

Since ARI=>SEA is reflected above, that means that the sole beatloop from last week, ARI=>SEA=>CIN=>BAL=>ARI is overridden, and SEA=>CIN=>BAL=>ARI re-emerges into the graph.

Rank Team Notes Last Week BeatPower

1

(Beat DEN) Indianapolis, Green Bay, and Dallas are still all roughly tied with each other in terms of beatwins and strength of beatpath tiebreaker, so Indianapolis remains on top due to the prior week’s ranking.

1

100.0

(16/16 – 0/16)

2

(Beat MIN) I liked how classy Favre was about breaking Marino’s record. I think it’s pretty wild how Green Bay is doing so well this year, mostly because I don’t have a very clear understanding of how their team has improved personnel-wise. Doesn’t this team still have some major holes?

2

100.0

(15/15 – 0/15)

3

(Beat STL) I’m currently of the belief that Parcells whipped these guys into shape, and they’re celebrating with Wade Phillips. I wonder at what point they go soft.

3

100.0

(15/15 – 0/15)

4

(Beat PHI) The Giants reinforce an existing beatpath, retaining their ranking. I didn’t see any highlights of this game, but all of a sudden their defense sounds scary good. The NYG=>PHI=>DET beatpath seems like it gulfs a long distance in quality if these teams have been playing according to their ability. Very strange.

4

86.7

(13/15 – 2/15)

5

(Bye) Tennessee retains their ranking on the bye.

5

92.3

(12/13 – 1/13)

6

(Bye) Washington rises a slot due to New England’s opponents looking worse and worse.

7

80.0

(12/15 – 3/15)

7

(Bye) Jacksonville had a bye, but leapfrogged New England, Pittsburgh, and Seattle due to quality of opponents.

10

83.3

(10/12 – 2/12)

8

(Lost to NYG) Philadelphia lost to a highly ranked team, but rises because their quality of beatwins is looking better.

12

73.3

(11/15 – 4/15)

9

(Beat SF) Seattle regains a beatwin over Cincinnati, which doesn’t help them much.

9

95.0

(9/10 – 0/10)

10

(Lost to ATL) Houston loses their beatwin over Carolina, but their previous win over Kansas City is starting to look better.

11

70.0

(3/5 – 1/5)

11

(Beat CAR) The beatpaths here aren’t very long so there is a lot of play in these rankings, but Tampa Bay continues to climb with some nice wins.

17

75.0

(3/4 – 1/4)

12

(Beat CIN) All of their opponents are looking worse and worse, the victories less impressive. There is still a lot of room for this team to be tested.

6

100.0

(7/7 – 0/7)

13

(Lost to ARI) Pittsburgh’s loss immediately gets beatlooped away. If we were using beatflukes, Pitt’s victory over SF would be retained. But either way, their quality of opponents isn’t looking real great, with the possible exception of Cleveland.

8

93.8

(7/8 – 0/8)

14

(Lost to IND) Denver doesn’t get hurt badly by losing to the top ranked team, and their two victories are looking slightly more impressive now. I watched this game this week. Even if I accounted for Cutler’s interception (imagining a Denver touchdown rather than an Indy one), and better Denver special teams (perhaps resulting in one more stop), it still seems Denver would have lost. I’m not even sure a better running defense would have helped. Consistently giving up eight-yard passes on first and ten is just impossible to get past. I wish I knew more about what the defense’s problem is and if it’s just a scheme that needs to gel more, or if it’s Bates messing everything up by putting scheme ahead of personnel. Also, just to extend a long paragraph, I still think Denver’s offensive playcalling sucks. Denver gets into the red zone on one drive, and Cutler doesn’t even take one shot to the end zone. That’s not a good way to score a touchdown.

16

72.7

(8/11 – 3/11)

15

(Beat CHI) Detroit’s quality of beatwins is looking better too, despite Chicago looking worse every week.

18

66.7

(10/15 – 5/15)

16

(Lost to SEA) Beatloops lead to SF having an ambiguous ranking, but they’re looking a lot worse now… I suppose it’s not worse than their history, but they had some serious expectations building up on them which they’re having trouble meeting.

15

62.5

(1/4 – 0/4)

17

(Beat SD) Kansas City sheds its beatloss to Chicago by defeating San Diego. Good thing, since Chicago was looking so much worse (which made the emergence of a beatloop more likely anyway).

24

50.0

(2/6 – 2/6)

18

(Lost to TB) Another ambiguous ranking – Carolina could be quite a bit higher or lower than this.

19

40.0

(1/5 – 2/5)

19

(Lost to KC) It’s crazy how different this team is. Everyone’s dogpiling on Norv Turner, but how does coaching matter this much? I’d like to see an examination of what kind of playcalling is making this happen. Seems San Diego should be doing better than this even if the coaching staff just takes the day off. Are they calling plays that are actually getting in the way? On the other hand, San Diego has a history of being a pretty undisciplined team. They were completely streaky – flashes of brilliance with many other phoned-in losses – up until last year. So this might just be the re-emergence of their true nature.

13

25.0

(0/4 – 2/4)

20

(Bye) Beaten by three very good teams that are looking better, New Orleans rises slightly.

23

0.0

(0/4 – 4/4)

21

(Beat NYJ) Buffalo rises slightly after defeating the Jets.

25

25.0

(2/8 – 6/8)

22

(Beat MIA) Oakland is a real nexus team in the beatpath graph right now. They’ve got a beatpath of significant length underneath them, all of teams that are looking rather crappy right now, but could all start looking better in the future (particularly Cleveland). And on the other side, Oakland’s placement is important to both the NFC East and the AFC South. Oakland’s future performance affects quite a lot of teams right now, and OAK=>CLE is especially interesting. And doesn’t it seem like Culpepper finding a home in Oakland is a particularly Raidersesque storyline? It should scare Denver given how the Jaguars dominated Denver with a similar team.

28

37.5

(6/16 – 10/16)

23

(Lost to DAL) No big penalty losing to Dallas at this point.

27

0.0

(0/5 – 5/5)

24

(Lost to DET) In addition to losing to Detroit, Chicago loses credit for their win over Kansas City.

20

12.5

(0/8 – 6/8)

25

(Beat BAL) Cleveland reinforces a beatpath to Baltimore.

29

29.4

(5/17 – 12/17)

26

(Lost to GB) Minnesota suffers no big penalty (at this point) in losing to Green Bay.

30

10.0

(1/10 – 9/10)

27

(Lost to NE) They actually held the 2007 Patriots to their lowest point total ever. Who says their defense sucks?

31

21.1

(4/19 – 15/19)

28

(Beat HOU) Atlanta sheds a beatloss to Carolina, and the teams above them still might end up being pretty good teams.

32

3.8

(0/13 – 12/13)

29

(Lost to CLE) A rather spectacular collapse by Baltimore, mostly because CIN=>BAL re-emerged on top of their loss to Cleveland.

14

15.8

(3/19 – 16/19)

30

(Lost to BUF) Their beatwin over Miami doesn’t really give them any protection to falling further down the rankings.

21

5.3

(1/19 – 18/19)

31

(Beat PIT) Arizona gets a little beaten down this week because they enter into two beatloops from their San Francisco loss, while BAL=>ARI re-emerges as well. While they shed their SF loss, it doesn’t make a difference because of an alternate beatpath. (ARI=>PIT and ARI=>SEA would both be seen as a beatflukes if we were using that variant.)

22

2.8

(0/18 – 17/18)

32

(Lost to OAK) Seems to be the most appropriate ranking for Miami at this point.

26

0.0

(0/19 – 19/19)

28 Responses to 2007 Week 4 Beatpaths Rankings

  1. Lebkin says:

    These rankings perfectly show how flawed mainstream analysis is. Everyone is talking about the Patroits going 16-0 based on their 4-0 record and wide margin of victory. But your rankings how poor the quality of their opponents has been, and how we don’t really know much about them yet. A victory next week against Cleveland won’t much either. The real test will be against Dallas on the 14th.

  2. Kenneth says:

    Love the website, hate the results (good job Griese). A few answers to your questions from a follower of the NFC:

    Re: Green Bay, the team was mostly young last year, but had a talented defense and some talent at wide receiver and a decent O-Line (plus, of course, Favre). This year, the defense has come into it’s own, and Favre has been good enough to move the offense. Their problem is that they have no run game to speak of, so they’re basically the Eagles circa 2002 or so. Dangerous and strong, but flawed. Which is good enough to look really strong in the NFC right now.

    As for Dallas, I think they’re extremely talented, especially on offense, but they haven’t really been challenged by a good defense yet (I was hoping my Bears would be that, but no chance…*sniff*). We’ll see how the team reacts when they play a good defense (read: Patriots).

    As for the Giants, it’s less that their defense is really that good, as that they have a very dangerous pass rush. When that works (as in the Philly game), it makes them very good (it helps that Philly can’t run). Against a team that can slow down that rush, the flaws in their back seven are exposed.

  3. Jon Coit says:

    RE: Chargers.

    There have been problems with execution, esp. offensive line blocking. That said, the playcalling vs. GB and KC was nutty. At GB, a tough road game, the Chargers had been passing well most of the day, and getting stuffed run after run. With a 4 point lead, Turner elects to continue getting stuffed rather than try to put the game out of reach with the pass. Vs. KC, the Chargers had been running well in the first half, but passing poorly, and had a 16-6 lead. Turner said after the game he elected to go away from the run in the second half because he thought they could put the game away with the pass–which wasn’t working due to KC’s pass rush. Defensively, the Pats showed how to attack them; a spread formation and quick throws neutralize the pass rush. Huard said after the game that their 50-yard TD pass (to go up 23-16) was the same play as Favre’s GW the previous week. I wish I could believe (like Turner does) that the offense is just on the edge of hitting on all cylinders and making him look smart.

  4. Kirk says:

    “These rankings perfectly show how flawed mainstream analysis is. Everyone is talking about the Patroits going 16-0 based on their 4-0 record and wide margin of victory. But your rankings how poor the quality of their opponents has been, and how we don’t really know much about them yet. A victory next week against Cleveland won’t much either. The real test will be against Dallas on the 14th.”

    You should watch the games. I’ll be the first to say the Pats haven’t had a difficult schedule, but they are completely destroying the competition. That’s the flaw in this analysis. They look solely at wins and losses, and not the quality of play. The Pats are clearly the best team in the NFL at this point.

  5. ThunderThumbs says:

    Ha, I’m with you up until the last sentence. At this point, the only thing that is clear is that they are comfortably better than the teams they have played. And going by who beats who, we can say that they are almost definitely better than BUF, CIN, SD, BAL, NYJ, ARI, and MIA.

    Don’t forget that the NFL playoffs is another system that looks solely at wins and losses and not quality of play (with the exception of some very distant tiebreakers). I don’t make many claims about this system regarding picks accuracy and the like, but I do believe that by the end of the season, this is a better way to pick what teams go to the playoffs.

  6. Kirk says:

    Well yeah, except the NFL playoffs are determined after 16 weeks of play, whereas right now you’re trying to make “objective” analysis after 4 weeks of play.

    Basically, there’s not a damn thing the Pats could do right now, in your system, to be ranked where they actually belong. I’m not complaining; I find the system very intriguing. But it’s ignorant to point out that they haven’t beaten anyone and ignore the fact that they’ve outclassed their opponents.

    For example. The Pats are averaging 37 pts/game and giving up 12 pts/game.

    Their opponents, ignoring games played against the Pats, are averaging 19.92 pts/game and giving up 23.25 pts/game.

    So basically, the Patriots are playing better against their opponents, than the other teams playing against their opponents. Much better in fact.

    There are two marks of a great team. The first is by defeating other good teams. The second is by outclassing weaker teams. The Pats haven’t had a chance to do the first, but anybody who’s paid any attention to the NFL this year knows they are not only capable of beating the Colts/Cowboys/Packers/Steelers, but would do so more often than not.

  7. Jon Coit says:

    The Pats will play those teams once, in the case of the Colts and Steelers possibly twice. Unless Allah has written down the victor in the Book of Time we don’t know who will win, and since they will only play at most twice it is impossible to actually *know* who will win “more often than not.” We might guess based on past performance or using some other method, but the accuracy of those guesses would also vary based on the criteria we choose. The Pats win over my Chargers sure doesn’t look as impressive as it did in week 2, eh?

  8. Justin says:

    First of all, great site, I just discovered it a week or two ago.

    I was curious about the application of “beatflukes”, as I was thinking they would help Arizona, not hurt them. I know they’re not being used yet, but follow me here for a minute:

    AZ has two unique beatpaths to SF: AZ –> SEA –> SF, and AZ –> PIT –> SF; so could one say that this shows SF–>AZ to be a fluke and remove it? That would break both of those 3-way beatloops, and after removal of the loop AZ –> SEA –> CIN –> BAL –> AZ, Arizona would be left with only 1 connection AZ –> PIT, which would put them quite a bit higher in the rankings.

    Just curious if there’s something wrong with that take.

  9. ThunderThumbs says:

    those are beatpaths that are within the beatloops being considered. It has to be a separate beatpath, separate from the beatloops that would otherwise be removed.

    If beatflukes were applied, then we would find that SEA and PIT both have alternate beatpaths to Arizona, so ARI=>SEA and ARI=>PIT would be considered beatflukes. (So would ATL=>HOU.) The main effect would be that SEA and HOU would be ranked three slots higher in the power rankings… PIT two slots higher… STL would be a couple slots lower. A couple of others adjusted by one slot, low in the rankings.

  10. Justin says:

    Ok, so beatloops are still taken out first (starting with the smallest), and then beatflukes are looked for, considering only the remaining paths? I think that makes sense.

  11. Kirk says:

    ” The Pats win over my Chargers sure doesn’t look as impressive as it did in week 2, eh?”

    Sure it does. A 38-14 pounding of any team, even the Chargers, is impressive.

  12. ThunderThumbs says:

    Jon, I guess that’s a compliment? 🙂

  13. Kirk says:

    Also, I don’t want anyone to think I’m hating on the rankings. I like the rankings, and I’m fine with what they say right now. There’s going to be a lot of variance with only 4 games (3 for some) to take info from. My only point is some of you seem to be living and dying by these rankings and refusing to take into account other, and in my opinion, more relevant factors because beatpaths doesn’t reflect them.

    So yeah, TT, I love your site. I get bored at work and attempt to come up with the beatpaths myself before I see how they compare to what you have. All love from my end.

  14. ThunderThumbs says:

    Thanks Kirk – yeah, it’s just a tool, it’s just fascinating how accurate they can end up for a system that is based off of just wins and losses. Especially when it exposes some media biases by making some teams out to be better or worse than they’ve actually performed given their competition.

  15. […] Anyway, you won’t be surprised to see that the Colts, Cowboys and Packers — all undefeated teams — are at the top, and the winless Dolphins are at the very bottom. But the surprising thing is that the Patriots — the consensus choice as the best team in the league — are right in the middle. Why? The official explanation: All of their opponents are looking worse and worse, the victories less impressive. There is still a lot of room for this team to be tested. […]

  16. […] Anyway, you won’t be surprised to see that the Colts, Cowboys and Packers — all undefeated teams — are at the top, and the winless Dolphins are at the very bottom. But the surprising thing is that the Patriots — the consensus choice as the best team in the league — are right in the middle. Why? The official explanation: All of their opponents are looking worse and worse, the victories less impressive. There is still a lot of room for this team to be tested. […]

  17. John says:

    I see great potential for a java app that allows a user to “play with” different variables. Maybe a weight factor, one that looks at point spread, yards gained/allowed, etc. Calculations done where top teams or bottom teams are removed from the equation. The community potential is there too.

    Oh how you could roll with this idea. Maybe next year.

  18. alan76 says:

    What your diagnosis looks like is a Ponzi scheme that arrives at predetermined goals. Of course we will know how good or bad your rankings are after the year is played, but you’re ranking of teams is more like a whirlpool that swirls up the teams until the worst ones go down the vortex first and the best are left at the top. Ponzi depended on continually finding suckers to pay previous suckers. Your charting depends on keeping the rankings moving to retain interest and obfuscate the end results.

  19. Jon Coit says:

    TT–if only the compliments could scrub the image of Rivers sitting on his behind watching 96 take the INT to the house… 🙂

    Re: relevant factors
    FO did an article a while back, here:

    http://www.footballoutsiders.com/index.php?p=3303

    To paraphrase the results, in playoff matchups, teams that had more blowout wins against weak opponents than their opponents were 21-7, while teams that had more close wins vs. strong opponents than the were 9-12. Point being, if there is any statisical corroboration here, it’s that the Pats dominating wins over crappy teams (sigh) so far actually make them a *better* playoff candidate on average than the other 4-0 teams.

    Ok, but to belabor this: I’d be a happy M-Fer (to quote O’Reilly) if I were a Pats fan. VERY happy. But whether or not the Pats are the “best” team–isn’t that beside the point? Presumably you’d be happier with a 4th SB win in Feb. than with being the “best” team at any point in the season. Moreover, as the above article points out, the 2003 and 2004 Pats are unique in that by those criteria they weren’t the “best” team in the league. I’m not quite as happy as I used to be about the Chargers being 12-4 in 2004 and 14-2 last year…

  20. JT says:

    This is why the season is 16 games long, and not 4 weeks long. At this point, teams have only played 3 or 4 games, so it’s very easy for a “possibly very good team” like the Patriots to start the season playing seemingly weaker teams. Having followed the Beatpaths rankings for the past two seasons, things doing really start firming up until week 8 or so. Between now and then, the following things could happen to a team like the Patriots:

    1. They beat teams who are good (ie, have beaten other quality teams), causing them to have more impressive victories, building longer beatpaths, and rising in the rankings.
    2. The teams they have previously beaten could win, making NE’s victory over those teams look better and pushing them up in the rankings.
    3. The Pats could struggle against tougher opponents, and end up staying where they are at.

  21. Kenneth says:

    Kirk–switch “Bears” for “Patriots” and turn the clock back a year or so, and you’d get a lot of my posts. Hello!

    Also, I can’t decide if alan76 is genius or madness.

    Basically, what you have understand/accept about these ratings/rankings is that they are an attempt to quantify something very specific–which teams have shown, through wins, that they are better than other teams–and not something more nebulous, like “which team is better” or even “which team is likely to win in a contest between two”. All this is doing is showing what teams HAVE done, and if you try to take them to be something else, you’ll probably be upset with it. But that doesn’t mean that the methodology has to be changed; it is what it is (though I’d be interested in tweaks, too).

    It’s like if you were trying to determine the relative value of two sub sandwiches, and you decided to measure their lengths. Does this really mean the longer one is the better value? No, there are other things like toppings to consider. But there’s no reason to disparage the measurement because of the things it’s not measuring. Take it at face value.

  22. […] Anyway, you won’t be surprised to see that the Colts, Cowboys and Packers — all undefeated teams — are at the top, and the winless Dolphins are at the very bottom. But the surprising thing is that the Patriots — the consensus choice as the best team in the league — are right in the middle. Why? The official explanation: All of their opponents are looking worse and worse, the victories less impressive. There is still a lot of room for this team to be tested. […]

  23. Kirk says:

    Kenneth, I agree 100%.

    I want to make sure I understand something about reinforcing beatwins. With the Bears win over the Packers, the Bears have another path to the Chargers now, so their should be a path of CHI-GB-SD, but not CHI-KC-SD, am I correct?

  24. Kirk says:

    And if beatflukes were in effect, Chargers > Bears would be considered a beatfluke and the CHI-KC-SD path would be reborn?

  25. Kirk says:

    Except PHI-DET-CHI-GB is now a loop. So the Bears have no beatwins over anyone.

    I think.

    This is too complicated for a guy with such little brainpower.

  26. Kirk says:

    My constant posts are great for advertising, you know. :p

    So this is what I have.

    GB=>SD

    GB=>NYG=>Wash=>Phi

    Phi=>Det=>Chi=>GB disappears due to a beatloop, but due to alternate paths, GB keeps its paths to SD and Phi.

  27. ThunderThumbs says:

    ehhehe keep it up Kirk. By the way, people occasionally try to spell my name Kirk and I tell them, “Well, you got ONE letter right…”

    There’s actually only one additional beatloop right now, CHI=>GB=>SD=>CHI. So really, the only effect of that upset is that it makes GB lose credit for their SD win. No real changes in the rankings from it, but GB’s place atop the graph is a little more rickety. In the fluke system, CHI=>GB would currently be seen as a fluke.

    However, SD now looks stronger. Which helps NE.

  28. otbricki says:

    So Buffalo is leading Dallas right now. Maybe they aren’t really that bad … It will be interesting to see how this works out, especially for NE.

Leave a Reply

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