The picks were 9-5 this week. One thing that is interesting is that there would have been a lot of beatflukes up until last week, but this week there would be only one. I think that means the graph might be getting more stable now.
Another thing I noticed is that right after the Denver win, with only that additional game figured in, the rankings changed quite a bit. But with the rest of the week’s matches included, the rankings are actually fairly static – not as many huge leaps and jumps as I’ve seen in prior weeks.
My graphs are up – NEW SITE: http://www.beatgraphs.com
Features:
All 3 graphs for every week this year.
Historical search, now goes back to 2004.
Helpful descriptions of how the process works.
NCAA Division I-A Football rankings.
Forums for discussing old results and new ideas.
Things to come:
More NFL history to be loaded going back to 1970.
Other leagues may be added – NBA, NHL, NCAA basketball.
Doktarr and TT, you have reserved slots as moderators on the forums should you choose to claim them.
I think this week, despite many of the close picks, provides considerable support for a “confidence” ranking of picks based on beatpower. Based on my original ranking of the picks, we can see that the incorrect picks are strongly clustered in the low-confidence bottom half:
NY Jets-St. Louis 100.0 (100.0-0.0) CORRECT
Carolina-Oakland 97.7 (100.0-2.3) CORRECT
Arizona-San Francisco 84.0 (90.0-6.0) CORRECT
Atlanta-New Orleans 54.6 (72.0-17.4) CORRECT
Baltimore-Houston 38.4 (84.4-46.0) CORRECT
Miami-Seattle 33.3 (40.4-7.1) CORRECT
Pittsburgh-Indianapolis 32.4 (91.7-59.3) WRONG
Tennessee-Chicago 29.2 (100.0-70.8) CORRECT
Jacksonville-Detroit 28.0 (28.0-0.0) CORRECT
Cleveland-Denver 14.9 (38.9-24.0) WRONG
Green Bay-Minnesota 14.5 (66.7-52.2) WRONG
NY Giants-Philadelphia 5.6 (77.8-72.2) CORRECT
Buffalo-New England 5.4 (36.0-30.6) WRONG
Kansas City-San Diego 0.3 (7.1-6.8) WRONG
I suggest that this method be adopted as the official pick ranking for beatpaths.
Very cool Tom – you want posting privileges to post this stuff on the main page? Go ahead and drop me an email through curtsiffert.com so I have your email address.
Just so that it’s said before tonight – the Jets have a long beatpath to New England, and are heavily favored.
The iterative algorithm ranks the Jets much, much lower, and as such has this game very close. The teams are playing for a beatpath. Moose’s path score rankings have the Jets very slightly favored, although it’s close enough that you would expect home field to swing it the other way.
Man, look at CHI vs. GB. Those two are almost in the exact same situation (only difference is GB is under ATL). That’s got to be a tough one for the rankings to pick.