2009 NFL Week 5 Beatpath Power Rankings

Kenneth here again folks! And yay, we’ve got a beatloop this week! That’s fortunate, because it’s kind of the reason for this site. I mean, if there weren’t any beatloops we’d call it…uh…well, something else. Whatever, let’s just go to the rankings.

Rank Team Notes Last Week BeatPower

1

(Beat NE) There’s actually a slightly different tiebreaker method in place this week–ThunderThumbs can explain it fully–but I don’t think it would matter in this case. Denver beating New England gave them a lot of support. Meanwhile, I think the Broncos are definitely for real. Not that beating the Patriots matters so much more than any of the others, but this is just one too many quality wins to deny.

2

100.0

(22/22 – 0/22)

2

(Beat STL) Meanwhile, as you might expect, beating the Rams doesn’t really help any. It’s too bad–I would have loved to see the chaos if that game had gone the other way.

1

100.0

(19/19 – 0/19)

3

(Beat BAL) Beating the Ravens seems impressive, but it doesn’t add a lot in terms of teams underneath for the Bengals; they already had paths to most of those teams.

4

94.7

(18/19 – 1/19)

4

(Bye) The Packers rise a bit on the bye, because they’re not hurt as much as other teams by the new beatloop (FORESHADOWING!).

7

83.3

(15/18 – 3/18)

5

(Bye) Okay, here’s a team that got hurt. The Jets were propping up the Saints somewhat, and the beatloop has hurt the Jets’ standing. So the Saints tumble somewhat too.

3

100.0

(12/12 – 0/12)

6

(Beat SF) The LeapOfTheWeek! The Dolphins beating the Jets let the Falcons get rid of their beatloss to the Patriots. If you didn’t follow that, it’s because Miami beat New York, who beat New England, who beat Atlanta, who beat Miami. Meaning the beatpaths system views all of that data as ambiguous. Now, who had week 5 as the first time I’d have to explain THAT?

16

90.0

(12/15 – 0/15)

7

(Beat TEN) Colts gain little from beating the Titans. Meanwhile, I’m beating my brains out during that game because I’ve got Peyton Manning on my fantasy team but my opponent has Austin Collie.

6

100.0

(11/11 – 0/11)

8

(Bye) Part of the reason that I have had trouble buying the Broncos and Bengals is that I saw Orton and Benson when they were in Chicago, and I have trouble believing they can keep this up. It seems like an indictment of Lovie Smith (or maybe OC Ron Turner) if they do.

9

77.8

(14/18 – 4/18)

9

(Beat OAK) Part of the annoying thing about no beatloops is that I have nothing to say about games like this. I mean, the Giants even gained a beatwin here. But who cares?

11

100.0

(7/7 – 0/7)

10

(Lost to DEN) Pats lose to the Broncos, and the new beatloop removes their beatwin over Atlanta. Overall, there have been better weeks in New England.

8

81.8

(8/11 – 1/11)

11

(Lost to ATL) Man, where did that come from? San Fran doesn’t lose much because the Falcons are ranked highly, but still–that’s the kind of score you would expect the 49ers Bay compatriots to be on the other end of. Which, I guess they were.

10

80.0

(8/10 – 2/10)

12

(Lost to CIN) When it comes right down to it, Baltimore hasn’t really beaten anyone of significance this year (unless San Diego is dogging it again). They’ve looked impressive, yes. But they haven’t beaten good teams.

12

70.0

(7/10 – 3/10)

13

(Beat DET) Beating the Lions gives the Steelers some actual help in the form of beatpaths. So would a win against the Browns next week. How the mighty have fallen.

14

66.7

(10/15 – 5/15)

14

(Beat TB) Okay, what is the deal beating clearly inferior teams this week? Thank god for the Jets and Dolphins, or this might have been a completely boring week.

15

83.3

(5/6 – 1/6)

15

(Lost to MIA) Hey, if some rise, others must fall. No beatpath to the Patriots leaves the Jets in an ugly position. On the other hand, they’re only under the Saints, so that’s not bad for the future.

5

68.8

(4/8 – 1/8)

16

(Beat KC) Before the first Austin TD, I would have bet money that the Cowboys were going to blow it again. So kudos to them for pulling it out. Still, no big bonus for beating a winless team.

17

71.4

(5/7 – 2/7)

17

(Beat HOU) Cardinals already had a beatpath to Houston through Jacksonville, so no help there; and the 49ers looking slightly worse hurts the Cardinals a bit.

13

55.6

(5/9 – 4/9)

18

(Beat JAC) A win over Jacksonville actually helps out the Seahawks a good deal; the Jaguars are propping some people up with their win over the Texans.

23

42.9

(6/14 – 8/14)

19

(Beat WAS) Panthers get on the board with a win over Washington, and thus move up. That game looked like it hinged on a few small but key plays/bounces. That kind of thing is fascinating when you have two great teams going at it, but between two teams that suck it just seems to magnify their impotence.

25

33.3

(3/9 – 6/9)

20

(Bye) I hope during the bye someone sent the Chargers a note telling them that the Broncos are good this year and it’s time to start playing for real.

21

38.5

(5/13 – 8/13)

21

(Beat BUF) A win–well, it’s hard to call that a win, but it counts as one nonetheless–boosts the Browns a bit. Actually, it’s funny–depending on what you call Baltimore, the Browns have only lost to good teams so far. If they start winning some, they look like they have room to move up in the graph.

24

28.6

(2/7 – 5/7)

22

(Lost to PIT) I have the feeling that Matthew Stafford would have lost that game, too, but what do I know. Maybe he has special Calvin-Johnson-leg-healing powers.

19

30.0

(3/10 – 7/10)

23

(Lost to SEA) There was somebody trying to stop me from commenting on the Jaguars here, so I was afraid it wouldn’t get up. Then I realized the guy in my way was Rashean Mathis. Zing!

18

28.6

(4/14 – 10/14)

24

(Beat NYJ) Beating the Jets looks nice, but because of the loop the path doesn’t go into the graph for the Dolphins. Instead, they get to shed their loss to the Falcons, who weren’t really holding them down much anyway. Meanwhile, they gain no support, so…here they are. It will take some time for the Dolphins to climb out from the pile they’re under.

26

21.4

(2/14 – 10/14)

25

(Lost to ARI) How’d that power formation work out for you Texans? Houston drops a bit because, ironically, the Seahawks go over the Jaguars, which piles a lot more teams on top of them both.

20

18.8

(3/16 – 13/16)

26

(Lost to CAR) Time will tell whether being the first team beat this year by the Lions or by the Panthers is the bigger scar. But in either case, that’s a nice pair of black eyes for you. I’m going to say this: Carlos Rogers is right, and if Daniel Snyder doesn’t back him up he’s a tool.

22

13.3

(2/15 – 13/15)

27

(Lost to CLE) Bills lose a game in which both teams agree not to play offense. I thought Buffalo had great special teams?

28

7.1

(1/14 – 13/14)

28

(Lost to IND) At this point, it’s just scheduling quirks that determine which teams are worse. The Titans are helped by the fact that most of their opponents are stacked on top of each other.

29

0.0

(0/15 – 15/15)

29

(Lost to NYG) How these jokers beat the Chiefs is beyond me.

30

5.0

(1/20 – 19/20)

30

(Lost to MIN) I watched some of this game, and I’m going to make a bold prediction–the Rams will win a game this year. When they weren’t making huge crippling mistakes, they were hanging well with the Vikings in the first half. It was just 3 or 4 awful plays that killed them. It seems likely that they can cut down on that, at least for one game.

27

0.0

(0/17 – 17/17)

31

(Lost to DAL) Same here. I don’t think much of Dallas but they are somewhat quality, and the Chiefs played them well. Side note: How cool would it have been to see the Chiefs beat the Cowboys as the Dallas Texans?

32

0.0

(0/22 – 22/22)

32

(Lost to PHI) It seems like there’s a lot of different takes on this Josh Johnson kid. Some people think he stinks and should be benched; other people think the team stunk and Johnson did better than could be expected. I didn’t watch the game but what I saw gave me the latter impression.

31

0.0

(0/21 – 21/21)

13 Responses to 2009 NFL Week 5 Beatpath Power Rankings

  1. ThunderThumbs says:

    I think Atlanta was helped as much by their win over San Francisco as they were by the disappearing beatloop – either way, a very good week for them. The other thing that is interesting to note is that there will be mass carnage this coming week if San Diego beats Denver. Denver has multiple routes of the same length to San Diego – it’s the system’s way of saying that Denver Really Should Beat San Diego – and if they don’t, all sorts of assumptions about this season get thrown out the window – Denver will completely crater.

  2. The MOOSE says:

    Well if SD does beat DEN, we’d get our first difference in the Iterative graph and SD->DEN would immediately disappear, so the game is pretty irrelevant from that perspective. From the Standard perspective DEN would fall to about where NYG is. And they’d still have a chance to regain the paths they lose by beating SD in the rematch.

  3. Thurhame says:

    Denver’s win over New England really wasn’t that big. It only gave them two new beatwins (NE, BAL), and CIN->BAL would have given them a beatwin over Baltimore anyways.

  4. Kenneth says:

    That’s true, I suppose. It didn’t occur to me at the time that the Baltimore tree came in from Cincy anyway.

  5. ThunderThumbs says:

    Yeah, and it underscores how precarious Denver’s position is too, I think – so much is propped up by that Cincinnati win. Of course, Denver could hardly do better than it has so far this season.

  6. Thurhame says:

    I hadn’t noticed that about Denver’s position, but you’re right; without Cincinnati they would lose nine of their teams underneath (over one third).

    Speaking of Cincinnati, they are propped up to a lesser extent by Green Bay. Without CIN->GB they would lose over 25% of their teams underneath.

    In turn, Green Bay relies almost entirely on GB->CHI. Everything they have except St. Louis comes from that beatwin.

    A slightly precarious beatpath.

  7. Tom says:

    Are there really any positions this early in the season (based on so little data) that aren’t precarious?

  8. Thurhame says:

    Kenneth mentioned a new tiebreaker…what is it?

  9. ThunderThumbs says:

    The new tiebreakers is Edges. I haven’t managed to come up with a better name. Sometimes we call it EdgePower.

    Take all the arrows below a team (in that team’s beatpaths), and subtract all of the arrows above a team. This includes hidden “redundant” arrows that might not show in the graph I post.

    It’s proven to be more accurate than the nodes approach (last week’s tiebreaker). I felt all right about switching because it hasn’t let to any different picks so far this season.

  10. […] The Winning Ways of Winners « 2009 NFL Week 5 Beatpath Rankings October 15th, 2009 2008 NFL Picks, […]

  11. doktarr says:

    It’s interesting that some other more raw-stats based ranking algorithms massively devalue teams like Cincinnati (18th in footballoutsiders.com rankings) for their close wins. Advancednflstats.com goes even further, and has Tennessee ranked above Minnesota (really), based on vastly different schedule strengths.

    I’m pretty comfortable with this – different data sets should produce different results, especially at this stage of the season. As I’ve said before, I’d be very interested to see the beatpath algorithm working on a different set of data. Obviously Moose has done the points stats, but something like game-by-game VOA numbers from footballoutsiders would be even more interesting to me. I’m more broadly interested in beatpaths as a SoS correction method, moreso than I am interested in simply looking at wins and losses in a new way.

    Of course, there are some teams every algorithm agrees on, like the Colts and Saints (good) and the Rams and Raiders (bad).

  12. ThunderThumbs says:

    How do you mean a different set of data? Curious how this could be applied to VOA numbers. Do you mean, VOA numbers being applied as tiebreakers to a beatpath-graph-respecting set of rankings? I think as soon as you create a method that involves directly contradicting a beatpath (like having Tennessee ranked ahead of Minnesota, who has beatpaths to them), then it kind of throws everything out.

  13. doktarr says:

    No, I mean something completely different – using the VOA differential of a game as the weight of the edge. Much like Moose’s “weighted” algorithm uses the score differential as the weight of an edge, I would suggest VOA differential as the weight of an edge.

    Unfortunately, footballoutsiders doesn’t release the VOA game-by-game data currently, even on the premium side. They do have week-by-week DVOA in the premium database, but that already has the opponent adjustments in it, which defeats the whole point of applying a graph-based algorithm to it.

    If they ever do start releasing that data, I’m definitely going to want to start running that graph, though, just to see how the graph-based strength-of-schedule adjustments compare to footballoutsider’s more involved opponent adjustment procedures.

Leave a Reply

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