VRI was my original purpose for creating VaMatrix. I wanted to find an alternative to the VHSL Power Point system, a system that arbitrarily assigns values to certain teams based on classifications and raw records. No shade to that system, it largely serves its purpose, but in my opinion, must make many assumptions that could be better solved with math. I suspected there was a way to create a logic/algebraic rating system that better evaluated a team's resume.
VRI is based on the Colley Matrix, a ranking system created by Dr. Wesley Colley. What attracted me to it was the simplicity of the idea behind it. Instead of asking how good a team is, it asks how impressive that team's record is once you account for the schedule it played. The ratings are simply the algebraic solution that best fits every game's result (across the state) at the same time. I have tweaked the formula slightly but the overall system is very similar to the Colley Matrix.
Each team's rating is the number that best explains its results within the context of every other game played. Since every opponent has played other opponents, and those opponents have played others, every game becomes part of one large interconnected web that the math solves all at once.
That's essentially what the Virginia Resume Index does.
If there are 310 football teams, there are 310 equations that all depend on each other. Rather than assigning values by hand, we solve the entire system at once. The solution is the best set of ratings that satisfies all of those relationships simultaneously.
Nobody has to arbitrarily assign values to figure out how much a win should be worth. Algebra figures that out based on the results of all games.
Suppose there are only five teams in Virginia with the following results:
| Team | W | L | Result |
|---|---|---|---|
| Team A | 1 | 1 | Team B |
| Team B | 1 | 2 | Team D, Team A, Team C |
| Team C | 3 | 0 | Team A, Team B, Team E |
| Team D | 1 | 2 | Team E, Team B |
| Team E | 0 | 2 | Team C, Team D |
Matrix:
| Result | Team A | Team B | Team C | Team D | Team E |
|---|---|---|---|---|---|
| Team A | — | W | L | — | — |
| Team B | L | — | L | W | — |
| Team C | W | W | — | — | W |
| Team D | — | L | — | — | W |
| Team E | — | — | L | L | — |
This produces the following records and ratings:
| Team | Record | Rating |
|---|---|---|
| Team C | 3–0 | 0.679 |
| Team A | 1–1 | 0.523 |
| Team B | 1–2 | 0.415 |
| Team D | 1–2 | 0.371 |
| Team E | 0–2 | 0.262 |
Although Teams B and D both finish with identical 1–2 records, Team B receives the higher rating because its losses came against higher-rated opponents. Rather than assigning values by hand, the matrix determines each team's rating by solving every team's equation simultaneously, allowing the strength of each team's resume to emerge naturally from the results.
One consequence of this approach is that the rankings can look a little odd during the first few weeks of the season. If Team A beats a team that eventually turns out to be very strong, the system doesn't know that in Week 1. At that point, a win over a strong team and a win over a weak team are treated exactly the same because there isn't enough information to distinguish between them yet.
Personally, I think that's a feature rather than a flaw.
The system isn't allowed to assume that a traditionally successful program is going to be good again this year, just like it isn't allowed to assume a program that struggled last season will struggle again. Everyone starts with the same information: zero. As more games are played, the picture gets clearer, schedules become connected across the state, and the ratings settle into something much more meaningful.
Another thing I like about the Colley method is that it doesn't use margin of victory. Once you've won the game, you've accomplished the objective. Winning by 35 doesn't count more than winning by 3. That keeps the focus on results rather than encouraging teams to pile on points.
I don't think VRI is the perfect way to rank team strength. But if the goal is to measure what teams have actually accomplished, not who Vegas would favor or who looked the most impressive on film, I think it's one of the fairest methods available because it blindly lets the results of the season speak for themselves.