SITE NEWS: We are moving all of our site and company news into a single blog for Sports-Reference.com. 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.

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

Pro Football Reference Blog

NCAA: SRS ratings through six weeks

Posted by Chase Stuart on October 11, 2009

PFR has used the Simple Rating System to grade college and NFL teams for years. All ratings or rankings are meaningless without explanation, and the link above explains what the SRS tries to do. The SRS version that I'm implementing below is most useful to predict future results; the SRS is predictive, not retrodictive. That means the SRS will have no trouble at all ranking a team that's undefeated and beat a team with one loss behind the very team it beat. Why? One, because we know that one game is just one game, and never is conclusive proof that one team is better than another; and two, because the SRS weighs each game equally. Of course, sample size issues are always present here; while I've waited for six weeks before presenting the SRS, we really need two or three more weeks in the books before we can have full faith in this system. For now, though, maybe they'll make you rethink your perception of a couple of teams.

So how am I calculating these simple ratings?

1) For each game, 3 points are given to the road team (unless it's a neutral site game). After that adjustment, all wins and losses of between 7 and 24 points are scored as however many points the team won by. So a 24-10 road win goes down as +17 for the road team, -17 for the home team.

2) Wins of 7 or fewer points are scored as 7-point wins and losses of 7 or fewer points are scored as 7 point losses, except that road losses of 3 or fewer and home wins of 3 or fewer are graded as 0 point ties. So a 21-20 home victory goes down as a tie for both teams. This is not as drastic as it sounds, because the SRS ultimately is not concerned with win/loss records. There is no distinction between a win and a loss (you don't need to make such distinctions in predictive systems) except for when the game is close. So three 10-point wins scores +30, just as two 20-point wins and a 10-point loss scores as +30. However, three 3 point wins (+9 before the adjustments, +21 after) is worth more than two 10 point wins and a 1 point home loss (+21 before, +13 after).

3) Wins/Losses of more than 24 points are scored as the average between the actual number and 24. This is to avoid giving undue credit to teams that run up the score. Texas' win over UTEP, 64-7, comes in as +39 for Texas. Why? Texas won by 57, but was at home, so that drops to +54. Averaging 24 and 54 gives you +39. Exception: In games against FCS/I-AA competition, there is no run-up-the-score modifier. Why? Otherwise, the elite teams could beat the FCS cupcakes by 64 points and go down in this system. Additionally, because of sample size reasons, I've combined all FCS teams into one "team" for SRS purposes: FCSTEAM. To the extent that a team plays a weak FCS team, they will therefore be overvalued (because they'll received credit for playing an average FCS team); to the extent that a team plays a top notch FCS team (relative to the average FCS team that plays an FBS team -- only inter-division games are used to calculate the average FCS team), they'll be undervalued.

Doug (and the readers) might argue that I've taken the first "S" out of the SRS, but I think these ratings are still simple. They're merely designed to get an approximate rating on how strong each team really is. By using just points scored and points allowed, we're obviously leaving lots of information behind. That's why I still think these are pretty simple, but a large improvement over looking at just points differential or won/loss records.

After translating the raw scoring differntial using the system above to a "SRS" victory margin, I simply add the victory margin for each game to the stength of schedule for the opponent. A 24-point win over an average team is equivalent to a 14-point win over a team that's 10 points better than average. I ran hundreds of iterations to figure out the true strength of schedule for each opponent and the true rating for each team; this involves adjusting the SOS of each team's opponents by the SOS of those opponents, and so on. Don't worry: at the end, the results converge, and we have true and legitimate ratings. My favorite part about the SRS is that everything adds up at the end. So after posting the ratings for each team, you could go through the schedule of any of the 120 FBS teams and understand exactly why their rating was what it was. I'll do that below with one team. First, the results:

```Gms    MOV     SOS    RAT     conf   team
5     28.4     3.7    32.1    SEC    Florida (5-0)
5     22.9     7.5    30.4    B12    Nebraska (4-1)
6     22.3     7.9    30.2    SEC    Alabama (6-0)
6     13.5    13.5    27.0    ACC    Virginia Tech (5-1)
5     27.8    -1.8    26.0    BigE   Cincinnati (5-0)
5     22.0     3.8    25.8    B12    Oklahoma (3-2)
5     25.2    -1.1    24.1    B12    Texas (5-0)
5     17.8     3.1    20.9    P10    Southern Cal (4-1)
5     21.7    -1.6    20.1    WAC    Boise St (5-0)
6     13.3     6.3    19.6    P10    Oregon (5-1)
6     15.6     3.1    18.7    B10    Ohio State (5-1)
5     21.7    -5.4    16.3    B12    Kansas (5-0)
6      8.4     7.6    16.0    SEC    LSU (5-1)
5     17.4    -1.9    15.5    MWC    TCU (5-0)
6      8.1     6.7    14.8    SEC    Tennessee (3-3)
5      8.7     6.1    14.8    SEC    Arkansas (3-2)
5      5.6     8.5    14.1    ACC    Miami FL (4-1)
5      9.4     4.5    13.9    WAC    Fresno St (2-3)
5     28.1   -14.5    13.6    BigE   South Florida (5-0)
5      7.0     5.8    12.8    INDY   Notre Dame (4-1)
5     12.4     0.4    12.8    SEC    Mississippi (3-2)
6     14.7    -1.9    12.8    BigE   Pittsburgh (5-1)
5     13.0    -0.2    12.8    B12    Missouri (4-1)
6     13.8    -1.3    12.5    MWC    Brigham Young (5-1)
6     10.2     1.8    12.0    SEC    Auburn (5-1)
4      6.5     5.2    11.7    SunB   Arkansas St (1-3)
6      9.8     1.8    11.6    P10    Stanford (4-2)
6     18.4    -7.0    11.4    MAC    Central Michigan (5-1)
6     19.0    -7.6    11.4    B10    Penn State (5-1)
6     14.9    -3.6    11.4    MWC    Air Force (3-3)
6     16.9    -6.0    10.9    B12    Texas Tech (4-2)
6      5.5     5.4    10.9    B10    Wisconsin (5-1)
5     10.6     0.3    10.9    BigE   Connecticut (3-2)
5     13.4    -2.8    10.6    P10    Arizona St (3-2)
6      8.4     2.1    10.5    B10    Iowa (6-0)
5      7.8     2.6    10.4    P10    California (3-2)
6      7.3     2.2     9.5    B10    Michigan St (3-3)
5     10.8    -1.3     9.5    CUSA   Houston (4-1)
5      5.6     3.6     9.2    ACC    Clemson (2-3)
5      7.6     1.5     9.1    B12    Baylor (3-2)
6      5.2     3.8     9.0    P10    Oregon St (4-2)
5     13.4    -4.4     9.0    B12    Oklahoma St (4-1)
5      2.0     6.9     8.9    P10    UCLA (3-2)
6      6.6     2.1     8.7    ACC    Boston College (4-2)
6     -3.2    11.7     8.5    P10    Washington (3-3)
6     -3.8    12.2     8.4    SEC    Georgia (3-3)
6      7.8     0.4     8.3    SEC    South Carolina (5-1)
6      6.7     1.6     8.2    ACC    Georgia Tech (5-1)
5      5.0     3.1     8.1    P10    Arizona (3-2)
6      7.6     0.4     8.0    B10    Michigan (4-2)
5     11.0    -3.1     7.9    MAC    Northern Illinois (3-2)
5     15.1    -7.3     7.8    BigE   Rutgers (4-1)
5     -2.3    10.1     7.8    SEC    Kentucky (2-3)
6      8.2    -0.5     7.6    INDY   Navy (4-2)
5      8.8    -1.2     7.6    MWC    Utah (4-1)
5     18.6   -11.6     7.0    CUSA   Tulsa (4-1)
6      6.5     0.2     6.7    ACC    Wake Forest (4-2)
6      2.7     3.8     6.5    B10    Minnesota (4-2)
6      8.7    -2.9     5.8    ACC    North Carolina St (3-3)
6     -0.3     5.5     5.3    ACC    Florida St (2-4)
5     -2.3     6.1     3.8    SunB   Troy (3-2)
5      4.2    -0.4     3.8    SunB   Louisiana-Monroe (3-2)
5      6.0    -2.5     3.5    B12    Texas A&M (3-2)
6     -0.5     3.5     3.0    SEC    Mississippi St (2-4)
5      9.2    -6.5     2.7    BigE   West Virginia (4-1)
6      4.8    -2.0     2.7    ACC    Duke (3-3)
6      6.5    -4.2     2.3    CUSA   Southern Miss (3-3)
6     -3.3     5.5     2.1    B10    Purdue (1-5)
6      7.7    -6.0     1.7    ACC    North Carolina (4-2)
6      4.8    -3.3     1.5    MAC    Ohio U. (4-2)
5     -1.0     2.0     1.0    WAC    Nevada (2-3)
6      4.5    -3.7     0.8    WAC    Idaho (5-1)
5      2.5    -1.9     0.6    ACC    Virginia (2-3)
6      3.2    -4.2    -1.0    SEC    Vanderbilt (2-4)
6     -0.8    -1.1    -2.0    MWC    Colorado St (3-3)
6     -5.7     3.5    -2.2    MAC    Bowling Green (2-4)
6      1.9    -4.6    -2.7    B12    Iowa St (3-3)
5     -1.2    -2.8    -4.0    WAC    Louisiana Tech (2-3)
5     -0.2    -3.8    -4.0    SunB   Middle Tennessee St (3-2)
6      5.3    -9.7    -4.3    B10    Northwestern (4-2)
5     -5.8     1.3    -4.5    BigE   Louisville (2-3)
5     -6.6     1.9    -4.7    SunB   Florida Int'l (1-4)
5     -6.0     1.1    -4.9    B12    Colorado (1-4)
6     -3.1    -1.9    -5.0    B10    Indiana (3-3)
6      0.9    -6.0    -5.1    B12    Kansas St (3-3)
6      1.1    -6.3    -5.2    CUSA   Marshall (4-2)
5     -7.1     1.4    -5.7    SunB   Louisiana-Lafayette (3-2)
6     -7.8     2.0    -5.8    CUSA   UTEP (2-4)
6      0.6    -6.4    -5.8    MAC    Western Michigan (3-3)
5     -1.8    -4.4    -6.2    MWC    San Diego St (2-3)
5    -12.1     5.7    -6.4    B10    Illinois (1-4)
5      3.0    -9.6    -6.6    CUSA   Central Florida (3-2)
5     -0.2    -6.9    -7.1    CUSA   SMU (3-2)
6     -3.7    -3.5    -7.2    MAC    Buffalo (2-4)
5     -3.8    -3.4    -7.2    CUSA   Alabama-Birmingham (2-3)
6     -8.7     1.0    -7.7    BigE   Syracuse (2-4)
6     -1.5    -6.4    -7.9    CUSA   East Carolina (3-3)
5     -4.2    -3.9    -8.1    WAC    Utah St (1-4)
4    -16.6     8.2    -8.5    SunB   Florida Atlantic (0-4)
5      2.8   -11.9    -9.1    MAC    Temple (3-2)
5     -8.4    -0.9    -9.3    MAC    Akron (1-4)
6     -5.9    -3.6    -9.5    MWC    UNLV (2-4)
6     -7.1    -2.5    -9.6    MAC    Toledo (3-3)
6     -2.5    -7.3    -9.8    MWC    Wyoming (4-2)
6    -11.3     1.0   -10.3    ACC    Maryland (2-4)
5    -10.7     0.3   -10.4    SunB   North Texas (1-4)
5     -3.8    -6.9   -10.7    WAC    Hawai`i (2-3)
5    -14.4     3.4   -11.0    WAC    San José St (1-4)
6    -18.7     7.5   -11.2    P10    Washington St (1-5)
6     -6.0    -5.4   -11.4    CUSA   Memphis (2-4)
6     -7.5    -4.4   -11.9    MAC    Kent St (2-4)
6     -3.5   -10.1   -13.6    INDY   Army (3-3)
6    -21.0     5.3   -15.7    MAC    Miami OH (0-6)
6     -8.6   -10.6   -19.2    WAC    New Mexico St (3-3)
5    -14.9    -4.6   -19.5    CUSA   Tulane (2-3)
5    -18.0    -1.5   -19.5    MAC    Eastern Michigan (0-5)
6    -24.0     4.4   -19.6    CUSA   Rice (0-6)
6    -11.3    -9.1   -20.4    MAC    Ball St (0-6)
6    -21.5     0.6   -20.9    MWC    New Mexico (0-6)
5    -24.0     1.5   -22.5    SunB   Western Kentucky (0-5)
84   -26.5     2.7   -23.8    FCS    FCSTEAM (5-79)
```

Clearly, Nebraska shows up as the biggest surprise up top. How can they rank ahead of Virginia Tech, after the Hokies beat them earlier this year? Or ahead of Alabama, who beat VT in week one? Let's go down Nebraska's schedule:

Nebraska beat Florida Atlantic at home by 46 points, which gets translated into a 33.5 point win. FAU is 8.5 points below average, so Nebraska gets +25 points for that win.

Nebraska beat Arkansas State by 29 points (again at home), which becomes a 25 point win. Arkansas State is actually pretty decent (3-point loss to Iowa on the road, 61 point-win over Mississippi Valley State), grading at +11.7 in the SRS. While that grade will probably fall a bit (and then hurt Nebraska), for now, the Cornhuskers get +36.7 for that game.

The Virginia Tech game (16-15, Hokies) goes down as a tie for both teams, since the game was in Blacksburg. VT has an SRS rating of 27.0, so that game gives Nebraska +27.0 points.

Nebraska beat Louisiana-Lafayette by 55 points at home; against a cupcake that's 5.7 points below average, Nebraska still gets 32.3 points for that game. Finally, after beating Mizzou 27-12 in prime-time (Jason), Nebraska gets 18 points for that road victory. Since Missouri is 12.8 points above average, Nebraska gets 30.8 points for that game.

So 25.0 points for beating FAU, 36.7 against Arkansas State, 27.0 against VT, 32.3 against Louisiana-Lafayette and 30.8 against Missouri, for an average of 30.4. That's what I mean when I say the SRS rating all adds up; once you know the ratings (which I will do for you), you can see how each team got its rating.

It should go without saying that these ratings are not perfect; they're simple. And it should be clear that I don't think they should replace the actual polls, because the polls should actually care about who wins the game. A poll that ranks Nebraska ahead of Alabama would be wrong, in my opinion. But a betting line that said Oregon is 3.6 points better than LSU would seem fine to me.

Let me know what you think of the SRS. No one knows that this is an NFL site more than me, so I hesitate to get too college-heavy on here. But if there's demand, I can post these SRS ratings every week (usually Sunday).

This entry was posted on Sunday, October 11th, 2009 at 10:59 am and is filed under College. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.