Only two upsets this week, although one of them affects the graph considerably. Jacksonville was ranked only three spots behind Tampa Bay, but by busting Tampa Bay’s beatpath to them, they untie the NFC South from the AFC South, which flattens the graph considerably.
And, the rankings:
| Rank | Team | Notes | Last Week | BeatPower |
1 |
 |
(Beat WAS) New England is finally ranked #1 after demolishing Washington. I’m also officially on the Pats bandwagon now, we’ve finally got a team that has a mystique that I haven’t remembered since the ’86 Bears or a couple of those ’49ers teams. |
3 |
100.0(25/25 – 0/25) |
2 |
 |
(Beat DEN) All right, I finally get it. Brett Favre is like a mythic hero or something. The announcers were building him up as he came out for the opening overtime snap, talking about how this kind of moment was made for him, and then he throws an 82-yard bomb to end the game. That was some good television right there. |
1 |
97.9(23/24 – 0/24) |
3 |
 |
(Beat CAR) Denver is a tougher opponent than Carolina, so Indianapolis isn’t able to gain ground on Green Bay this week. Up until this week, Indianapolis would have been favored over New England, but New England’s schedule has finally ended up looking a little bit tougher than Indianapolis’. It’s still pretty close overall though. Definitely the biggest game of the year so far, I believe. I think there were a couple of others this season that were pretty highly ranked, but there hasn’t been a #1 vs #2 or a #1 vs #3. |
2 |
100.0(21/21 – 0/21) |
4 |
 |
(Beat TB) Jacksonville is able to shed its beatloss to Tennessee by defeating Tampa Bay. |
11 |
89.5(16/19 – 1/19) |
5 |
 |
(Bye) The top three teams of the NFC East are basically leapfrogged by Jacksonville. |
4 |
95.2(20/21 – 1/21) |
6 |
 |
(Beat MIA) No great reward for defeating Miami, but their beatpath to Washington protects them from falling. |
5 |
85.7(18/21 – 3/21) |
7 |
 |
(Lost to NE) No big penalty for losing to New England. The system here doesn’t pay attention to victory margin or anything like that. But wow was that an unbalanced game. |
6 |
81.0(17/21 – 4/21) |
8 |
 |
(Beat HOU) San Diego rises a notch due to Denver rising. |
9 |
83.3(13/18 – 1/18) |
9 |
 |
(Lost to GB) The loss to Green Bay doesn’t impact Denver’s ranking too much, but Jacksonville’s victory over Tampa Bay does. Denver’s four losses are all to the top eight teams in the league. I was at rehearsal during the game, watching the box score, and got depressed. But when I came home to watch the second half, I actually got pretty excited – Denver’s got some great pieces that just aren’t all the way glued together yet. Their special teams are much improved, which means their field position is a lot better – I think they actually won the field position battle for the first time all season. |
12 |
70.6(12/17 – 5/17) |
10 |
 |
(Beat CIN) Denver’s victory also enables Pittsburgh to rise a few notches after they defeat Cincinnati. |
14 |
62.5(10/16 – 6/16) |
11 |
 |
(Beat CHI) Detroit beats Chicago, but falls in the rankings due to Tampa Bay’s loss. |
7 |
73.8(15/21 – 5/21) |
12 |
 |
(Beat OAK) Tennessee loses their beatpath over Jacksonville and falls slightly despite beating Oakland. |
10 |
86.7(12/15 – 1/15) |
13 |
 |
(Bye) Kansas City sheds their beatloss to Houston, thanks to San Diego beating Houston, which helps them rise a couple of slots. |
15 |
73.3(9/15 – 2/15) |
14 |
 |
(Bye) Seattle is also able to rise thanks to Pittsburgh rising. |
19 |
50.0(7/16 – 7/16) |
15 |
 |
(Lost to JAC) Tampa Bay has the biggest drop this week – their beatpath over Tennessee gave them a lot of power, but they lost it with their loss to Jacksonville. Still the best in the NFC South, though. |
8 |
46.9(6/16 – 7/16) |
16 |
 |
(Lost to SD) Houston loses their beatpath to Kansas City since Kansas City had earlier defeated San Diego. It might re-emerge though if San Diego defeats Kansas City in the future. |
13 |
37.5(1/8 – 3/8) |
17 |
 |
(Beat NYJ) Buffalo is able to rise thanks to Pittsburgh rising. Buffalo’s losses have all been to good teams. |
24 |
42.9(6/14 – 8/14) |
18 |
 |
(Lost to IND) Carolina roughly holds steady after the loss to Indianapolis. |
17 |
39.3(5/14 – 8/14) |
19 |
 |
(Beat MIN) Philadelphia rises a couple of notches after defeating Minnesota. I still feel I detect a bit of pro-Philadelphia (or pro-NFC East) bias in the media, given where Philadelphia is right now in the rankings, but it’s not as bad as it used to be. |
23 |
35.0(2/10 – 5/10) |
20 |
 |
(Lost to TEN) An interesting question to ponder… which is flukier, Oakland’s loss to Tennessee, or Oakland’s victory over Cleveland? It seems that Cleveland is on its way up, and Tennessee is on its way down. Whose coattails will Oakland be riding on? |
18 |
40.0(8/20 – 12/20) |
21 |
 |
(Lost to PHI) Minnesota loses again. Their victory over Chicago is currently keeping them from three beatlosses – KC, GB, and PHI. That’ll probably catch up with them. |
22 |
25.0(1/10 – 6/10) |
22 |
 |
(Lost to DET) Chicago seems to be one of the least consistent teams this year. They’re in half of the beatloops so far, for one thing – the system just doesn’t quite know what to do with them. (Neither does the city, probably.) |
16 |
20.0(0/10 – 6/10) |
23 |
 |
(Beat STL) Did I say they seemed to be on the way up? I guess I meant as a broader trend. There’s no real reward to beating St. Louis, and a few other teams must have leapfrogged them this week. |
20 |
33.3(7/21 – 14/21) |
24 |
 |
(Lost to PIT) Cincinnati about holds steady near the bottom of the rankings. |
25 |
27.3(6/22 – 16/22) |
25 |
 |
(Beat SF) People are starting to buzz about the Saints again, but look at the quality of the teams they’ve defeated. |
21 |
25.0(3/14 – 10/14) |
26 |
 |
(Bye) Atlanta holds steady on the bye. |
26 |
3.6(0/14 – 13/14) |
27 |
 |
(Bye) Baltimore holds steady on the bye. They’ve finally got some opportunity to play higher quality opponents in the coming weeks. |
27 |
21.7(5/23 – 18/23) |
28 |
 |
(Lost to BUF) The Jets hold steady on the expected loss to Buffalo. |
28 |
4.8(1/21 – 20/21) |
29 |
 |
(Bye) Arizona gets to start playing again. I also agree that they appear to be a better team than their ranking – I wonder when they’ll start busting through the teams above them. |
30 |
6.5(1/23 – 21/23) |
30 |
 |
(Lost to NO) Gore and Smith were apparently both back last week, and the team apparently still looked awful. What’s the story with this team’s collapse? They were looking good at the end of last season. Can it really all be explained by Norv’s departure? |
29 |
6.2(1/24 – 22/24) |
31 |
 |
(Lost to NYG) It appears the Dolphins played not to lose overseas, and failed. |
31 |
0.0(0/22 – 22/22) |
32 |
 |
(Lost to CLE) I wonder how many weeks in a row the Rams will be down here. |
32 |
0.0(0/24 – 24/24) |
Indy has a beatpath to everyone except the NFCE, NFCN, New England, and San Diego.
New England has a beatpath to everyone except the AFCS, Green Bay, and KC.
Barring some new beatloops, the winner of the game will have a beatpath to everyone except Green Bay.
I think there is a pretty glaring flaw in the beatpath calculation. The NFC East, in general, has trounced Detroit. Philly’s win was convincing. On the other hand, Philly’s loss to Chicago causes it to “lose” the benefit of their win over Detroit. Yet Detroit has 2 wins over Chicago, now. Seems to me, after reviewing the Philly/Chi game, the Philly loss is a fluke. That is, a fumble isn’t a fumble because of a bizarre ruling by the ref, and an improbable 98 yard drive at the end of the game to win.
In all common sense models, Philly would still get credit due to the 2 Detroit wins over Chicago counting as “more” than the single Philly loss to Chicago. In addition, the trouncing Philly gave to Detroit is more indicative of their potential than the 2 point loss to Chicago.
I use Philly as an example, because it’s glaring. But it could apply in a few other cases, too.
Okay, let me make sure I’m following. Detroit beat Chicago twice. Philly beat Detroit. Chicago beat Philly.
I think the only argument you can make there is that Detroit is definitely better than Chicago if you go by wins and losses.
It’s ambiguous whether that means DET->CHI->PHI or PHI->DET->CHI though.
You’re saying that PHI’s big win over DET should count more than PHI’s close loss to CHI. I suppose I can understand your point of view there, but the point of the system is to pay attention to wins and losses only, and then rely on the rest of the interrelations to settle the rest.
Beatpaths chooses to therefore punt on those two questions. And Philadelphia does happen to be currently ranked ahead of Chicago, so it does see Philadelphia as better than Chicago, despite the loss. Detroit has had several other quality wins, though, so they are seen as better, all things considered.
For those who are interested, these are the BeatLoops I recorded this week.
NO => SEA => TB => NO
CAR => ATL => HOU => CAR
MIN => CHI => KC => MIN
MIN => CHI => GB => MIN
MIN => CHI => PHI => MIN
HOU => KC => SD => HOU
KC => SD => CHI => KC
GB => SD => CHI => GB
PHI => DET => CHI => PHI
TEN => JAX => TB => TEN
PIT => SF => ARI => PIT
SEA => SF => ARI => SEA
Yes, I understand the concept of concentrating on wins. But there HAS to be meaning to the wins. Fluke wins, which Chicago’s over Philly clearly is, should count for “less”. A dominating win (Philly over Detroit) should count for “more”. In addition, if the loss to a lesser team in a beatloop is close, but there are 2 bigger wins by another team over someone in that loop (Det over Chi), the 2 wins should count for “more”, making the win over THAT team more valuable than the close loss.
Over at Football Outsiders, it is pretty well documented that a win of 2 TDs or more is likely to be repeated in a rematch – the losing team is unlikely to win again. I’ve done my own cursory research on this and it does stand up. So a close loss isn’t saying 2 teams are NECESSARILY equal – it just says that they COULD be equal OR one team had an off day.
In the example I’ve cited, Philly has 2 wins over teams in beatloops (Det/Minn) and those teams have 3 wins over the other team in the beatloop (Chi). There has to be more value in the wins Philly has over Det/Minn than there is in the loss to Chi (which, after seeing some of the flukiness in that game, you’d have to agree is the case).
I’m not making a case for Philly. I’m making a case for alteration in managing Beatloops. Point totals in victory/defeat are not arbitrary, they do tell stories, just as wins and losses do.
BTW, it’s worth noting that if you review the scores of dominant teams, they rarely get blown out. Case in point, Indianapolis, last year, lost only one game by more than 7 points – to Jacksonville at home. That’s a pretty telling statistic. They had 4 dominating wins during the regular season, versus 1 game in which they were dominated.
Similarly, New England had 6 dominating wins versus one game in which they were badly defeated (at Miami of all places).
In some of those closer games, you would likely review and find that both these teams let up at the end and allowed the opponent to score late.
Do dominating wins tell a story? Yes. Philly had 7 dominating wins last year, and 2 games in which they were dominated. Very much an up and down situation – but the 2 big losses were to AFC teams (one the Colts), meaning they were VERY good, but also prone to letdowns and particularly unlikely to win the Super Bowl.
So… first step, reduce all links by 1/3, which axes MIN=>CHI.
Second step, reduce the links that are still in beatloops by another 1/6, which axes CHI=>KC, CHI=>GB, SD=>CHI, KC=>SD, and SF=>ARI.
Finally, reduce the still standing loops by 1/2 (i.e. wipe them completely, which eliminates NO => SEA => TB => NO, CAR => ATL => HOU => CAR, PHI => DET => CHI => PHI, and TEN => JAX => TB => TEN.
Preserved victories:
At 2/3 weight:
KC => MIN
GB => MIN
PHI => MIN
At 1/2 weight:
ARI => PIT => SF
ARI => SEA => SF
SD => HOU => KC
GB => SD
With the possible exception of the Arizona wins (which would get wiped out when we resolve the 4 and 5 team loops) I think those preserved wins jive with most folks’ intuitive perceptions of team strength.
Moose, if you post the 4 and 5 team loops, I’d continue the analysis of what victories I would “preserve” among those. I’d assume that some additional 4 and 5 team loops exist if we don’t throw out the victories I list above.
There were no 4 or 5 team loops left after removing the original 3 team loops. Currently I’m doing an overhaul on my version of the program to make it more flexible and give me the ability to manipulate it better.
Just based on the graph though, the only additions I see that create new loops are ARI => PIT and ARI => SEA.
ARI => SEA => CIN => BAL => ARI
ARI => PIT => CIN => BAL => ARI
Patriots should be ranked higher.
hahahahaha
OK, cool. Those were the only ones I was aware of, too. So, you decrement those loops by 1/2, which takes out the ARI wins, leaving:
At 2/3 weight:
KC => MIN
GB => MIN
PHI => MIN
At 1/2 weight:
PIT => SF
SEA => SF
PIT => CIN
SEA => CIN
CIN => BAL => ARI
GB => SD => HOU => KC
The only effects on the graph are:
GB gets an arrow to SD (and loses the extraneous arrow to Denver)
KC and Philly both get arrows to MIN
SD=>HOU=>KC is added, which pushes the whole KC=>OAK=>CLE=>CIN=>etc chain down at least one tier. The JAC=>KC arrow becomes extraneous.
The JAC=>ATL arrow also becomes extraneous, as JAC=>HOU=>KC=>MIN=>ATL emerges.
And TEN=>OAK becomes extraneous as well. HOU=>KC=>MIN really cleans up the graph, as it turns out.
[...] 1. Beatpaths has the Redskins at No. 7 overall, No. 4 in the NFC. And that’s before the win at New York. I’ve written about this site before, but, basically, their rankings are based on “wins, losses, who beats who, and nothing else.” It doesn’t care that a win came in overtime or a loss was by 45. Just wins and losses, and, by that measure, the Skins are damn near elite. [...]