A pretty memorable week overall, even if just for the NE@IND game.
A few more upsets this week, including a huge one with #31 defeating #2. The graph is overall a bit flatter. Beatpaths was 9-6 this week, again leading all variants. ITB was 8-7, while IT was 7-8. I was 2-1 on my personal picks, so I am now two ahead of ITB for the season. The BeatPicks were 5-3 this week, or 28-11 for the season.
The following games/paths are considered flukey this week:
- NYJ->TEN->JAC->HOU->CIN
- OAK->KC->WAS->DEN->DAL->ATL
- BUF->CAR->ATL
- STL->DET->WAS
- TB->GB->DAL
- CHI->PIT->MIN
- SF->ARI
- NYJ->HOU
- NYJ->NE
- NYG->DAL
- NYG->WAS
- OAK->PHI
- DEN->CIN
- DEN->NE
- DEN->SD
- TEN->SF
- CAR->WAS
- CAR->ARI
- KC->WAS
- SEA->JAC
And now, the NFL Week 10 Beatpath Graph:
It’s really bizarre to see Washington with no beatlosses. This is a good example of where the iterative approach makes a lot of sense. One good victory should not be able to wipe out a season of poor results. The Denver win should not be able to wipe out losses to the Giants AND Carolina AND Kansas City AND Philly AND Atlanta. (The STL win already took care of the DET loss; that makes sense.)
This is another interconference issue. Even if Denver had beaten Baltimore, it would also have been wiped out by the WAS win.
The graph is noticeably flat considering how deep in the season we are. Do you think it would be useful (and easy) to chart a graph showing the number of layers of the graph from week to week? I’d be curious to see if we’ve ever had only 6 layers (or fewer) after 8 or more games.
Yeah, it’s interesting – actually, I think I’ve seen this before. Once the second game of the division series starts figuring in, things seem to flatten out sometimes because the season splits are common. And then by the end of the season, they start getting vertical again.
Looking though the past years quickly I don’t see any week 10′s that are this compacted, although I might have missed it. This actually relates to my point as well – a lot of games are getting looped away by a relatively small number of upsets.
The iterative version is still 13 layers tall; the only teams with no beatlosses are Indy, NO, and Cincy. The only teams with no beatwins are STL, DET, and Tampa.
As far back as my graphs currently go, this week has the shortest “longest paths” of any week 10. Shorter graphs lead to smaller path counts, and this week I have IND in first place with only 54 paths. Compare this to last week when #1 NO had 184, or even more exaggerated, week 10 of 2007 when #1 NE had 3118 paths.
IND’s 54 is the only the second lowest for week 10 though. The lowest was also IND in 2003 when they had only 42 paths. It got worse the following week with NE and IND tied at #1 with only 30 paths each in week 11 of 2003.
I check out this site regularly and was just wondering if you could post all the beatloops that are in this weeks rankings. I do my own rankings (on an outdated microsoft spreadsheet) and can usually figure them all out myself, but im having a little difficulty this week.
Thanks, Steve
Sure, you bet – I don’t know why I forgot this week.
Removing TEN split from JAC
Removing OAK split from KC
Removing ATL split from CAR
Shrinking circle length to 5
Shrinking circle length to 4
Shrinking circle length to 3
12 Circles…
Loop: TEN=>BUF=>NYJ=>TEN
Loop: ARI=>HOU=>SF=>ARI
Loop: HOU=>BUF=>NYJ=>HOU
Loop: NYG=>DAL=>PHI=>NYG
Loop: NYG=>OAK=>PHI=>NYG
Loop: WAS=>STL=>DET=>WAS
Loop: DEN=>CIN=>PIT=>DEN
Loop: BAL=>DEN=>CIN=>BAL
Loop: BAL=>DEN=>NE=>BAL
Loop: TB=>GB=>DAL=>TB
Loop: BUF=>NYJ=>NE=>BUF
Loop: NE=>MIA=>NYJ=>NE
Done
Shrinking circle length to 5
Shrinking circle length to 4
9 Circles…
Loop: KC=>WAS=>DEN=>DAL=>KC
Loop: NYG=>WAS=>DEN=>SD=>NYG
Loop: CHI=>PIT=>TEN=>SF=>CHI
Loop: CHI=>PIT=>MIN=>SF=>CHI
Loop: CHI=>PIT=>MIN=>GB=>CHI
Loop: DAL=>CAR=>WAS=>DEN=>DAL
Loop: KC=>WAS=>DEN=>SD=>KC
Loop: DAL=>ATL=>WAS=>DEN=>DAL
Loop: PHI=>WAS=>DEN=>SD=>PHI
Done
Shrinking circle length to 8
Shrinking circle length to 7
Shrinking circle length to 5
2 Circles…
Loop: JAC=>HOU=>CIN=>CHI=>SEA=>JAC
Loop: ARI=>CHI=>CLE=>BUF=>CAR=>ARI
Done
I also think it’s interesting that (according to the graph) all we know about Denver is that they are better than Cleveland and Oakland. In addition to beating directly connected to only 2 teams, the 2 teams they are connected to are lousy.