Podcastin': SMQ on Football Outsiders
by SMQ on Nov 15, 2007 5:41 PM EST
in News
Email
PrintRussell Levine, one of the fine NFL-centered gurus of the inimitable Football Outsiders, called up Wednesday night to talk point spreads on the biggest games of the weekend, then edited it together using sophisticated audio technology and a dash of newt's blood into a strange new medium known to those who practice the nebulous form as a "pod-cast." Startling accents and stuttering ensued.
Update [2007-11-21 10:58:3 by SMQ]: Below the jump now, to get the beautiful, beautiful Southern Miss fight music off the home page...
In case it wasn't totally clear, I don't really know what happened to Kansas State after the Wildcats lost to Kansas. Just to make sure that got across. I don't know how to explain it, alright? They just stopped tackling people. Just to clear up that point.
6 comments
|
0 recs |
Do you like this story?
Comments
Re: Scheduling
FIU?
Scheduling is always chicken and egg - maybe those teams' stats and records are so good because they're playing terrible teams like Notre Dame, Washington and FIU?
That's the problem
That's not the problem with the RPI
if Sun Belt teams never played anyone else they'd all have .500 SOS, and we know that's not an accurate description of an all-Sun Belt schedule. (This has always been a problem with the RPI in college basketball.)
The problem with the RPI is that there's duplication between WP and OOWP and OWP and OOWP. There is no analogy at all to the putative "if the Sun Belt teams only played Sun Belt opponents every team would have an OWP of .500" (and that's not true, anyway. They'd all differ from .500 by an amount related to their winning percentage - the teams with a winning record would have a less-than .500 SOS and vice versa.)
Duplication is part of it
Sure, SOS wouldn't be exactly .500 ... but it does artificially push their SOS much higher than it should be. Hockey has tried all kinds of formula tweaks with the RPI because there are two conferences that are really bad (based on head-to-head results) compared to the other four, but the RPI doesn't show it - largely because their schedules are so insular that the ratings can't adequately be compared against the big four.
A smarter rating system would use the ratings themselves instead of just opponents' winning percentage as a factor (an iterative approach). Then the amount of direct competition between the two groups doesn't have to be a large portion of the schedule, just a large enough (absolute) number of games to get an idea of the groups' relative strength, in order to get an accurate rating.
Well, duh
In any case, the 116 computer rankings I use are all very cognizant of the different ways to characterize SOS. Several of mine break it down by scoring prowess vs prowess at preventing scores. They all come out about the same when you use SOS defined by opponents' rating as opposed to opponents' WP.
Certain photos copyright © 2012 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
