SITE NEWS: We are moving all of our site and company news into a single blog for We'll tag all PFR content, so you can quickly and easily find the content you want.

Also, our existing PFR blog rss feed will be redirected to the new site's feed. ยป Sports Reference

For more from Chase and Jason, check out their work at Football Perspective and The Big Lead.

2009 Yardage Differential SRS

Posted by Neil Paine on August 13, 2010

Just for kicks, I plugged last year's results (regular-season + playoffs) into the SRS formula -- except instead of using point differential as the inputs, I used per-game yardage differential, a quick-n-dirty favorite of oddsmakers when determining if a team has played above or below what its W-L record would indicate. Broadly speaking, the outcome of a game can be retrodicted reasonably well using some combination of yardage differential and turnover differential, and we know that turnover differential is somewhat unreliable from year to year, so this SOS-adjusted ypg differential could provide insight into which teams will play better or worse than expected in 2010:

Rank Team Rating
1 Dallas Cowboys 88.2
2 New England Patriots 77.8
3 Minnesota Vikings 74.0
4 Green Bay Packers 69.4
5 New York Jets 62.5
6 New York Giants 55.8
7 Pittsburgh Steelers 51.0
8 Baltimore Ravens 50.6
9 Houston Texans 44.1
10 Philadelphia Eagles 43.0
11 Carolina Panthers 34.6
12 Indianapolis Colts 32.1
13 San Diego Chargers 31.3
14 New Orleans Saints 30.7
15 Denver Broncos 26.8
16 Miami Dolphins 7.3
17 Atlanta Falcons 5.5
18 Cincinnati Bengals 3.3
19 Washington Redskins -1.4
20 Tennessee Titans -11.3
21 Arizona Cardinals -12.6
22 Jacksonville Jaguars -29.0
23 Chicago Bears -33.2
24 Tampa Bay Buccaneers -49.5
25 San Francisco 49ers -50.5
26 Seattle Seahawks -54.6
27 Buffalo Bills -57.8
28 Oakland Raiders -82.1
29 Kansas City Chiefs -83.2
30 Detroit Lions -92.7
31 St. Louis Rams -103.0
32 Cleveland Browns -127.0
Home-Field Advantage 24.3

What this basically says is that to produce a prediction for a game's yardage differential last season, you subtract one team's rating from the other's and add the HFA term to the home team. Doing this for every game last season, this set of ratings produced the smallest squared error.

What jumps out? The Super Bowl participants (Indy & New Orleans) rank surprisingly low; the Cowboys could very well validate Chase's gut feeling about a big year in 2010; the Jets were probably better than their record would indicate (as evidenced by their playoff run); the Patriots might not be finished quite yet; and trendy hopes for a 49ers resurgence (beyond winning the weak West) could be misguided.

Of course, none of this takes into account offseason player movement or the million other factors at play here. But it is interesting as an additional piece of info about which teams played better or worse than you might have thought last season.

This entry was posted on Friday, August 13th, 2010 at 12:24 pm and is filed under Insane ideas, Statgeekery. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.