What is this?
Omni Rankings publishes ratings and projections for each Division I team in men’s basketball, women’s basketball, and football. Wait, it gets even spicier. The ratings and projections are generated by mathematical models gauging the current strength of each team. The abridged version: It’s an orgy of math and college sports.
Who is behind it?
Omni Rankings was founded by Matt Olson, a Minnesota native and Northwestern University alumnus with a professional background in structural engineering and broadcast journalism (the natural career progression into manatee dermatology is pending). Matt has developed sports rating algorithms and mathematical models since 1990.
How frequently are the ratings updated?
Basketball ratings are recalculated daily, while football ratings are updated weekly. Each Omni Rankings ratings page includes a time stamp indicating the most recent update.
What is the difference between the Macro and Micro ratings?
Two fundamentals determine each team’s Macro rating: The strength of the opposition and the margin of victory or defeat in each game played. Micro ratings are derived from multiple team offensive and defensive statistics (24 for basketball, 22 for football). Each team’s current Macro and Micro ratings are aggregates of the individual ratings earned for each game played in a particular season.
Which games do the models use to generate the rankings?
The models consider the results of every game between Division I teams in a particular sport for a given season. Results of any game involving a non-Division I team are not included.
Are all of a team’s games given equivalent significance in determining that team’s rating?
The models assign moderately greater weight to the most recent games. This weighting affords the models flexibility (as each team’s performance varies due to improved/diminished level of play, injuries, coaching changes, or other factors) and balance (not allowing the result of a single game to skew the ratings of the teams involved).
Are results from previous seasons used in generating a team’s current rating?
No, each season stands alone. An iterative algorithm uses the results of the first series of each team’s games to generate a “best guess” initial rating for that team. This approach yields a more accurate and less arbitrary comparison of teams at the beginning of a season without relying on previous season results.
Do the models take into account home court or home field advantage?
Yes. In each season and in each sport, the models quantify the benefit of playing at home. The Macro ratings home advantage is indicated on each Division I and conference rankings page and may be added to each home team’s rating when projecting margin of victory between two teams in a hypothetical game. The Micro ratings home advantage is a series of factors — one each applied to the individual statistical components of the Micro rating — and varies depending on the statistical attributes of the two teams involved. The home court/field effect is neglected for neutral site games in both the Macro and Micro ratings.
Does running up the score on an opponent benefit a team’s Macro rating?
The model applies a diminishing returns algorithm to limit the effect of “meaningless” points in no-longer-competitive games. A similar algorithm tempers the margin of victory in overtime games.
What are the Micro offense and defense ratings?
The Micro offense rating indicates the projected points that team would score against a statistically average Division I opponent. Similarly the Micro defense rating reflects the projected points that team would allow against a statistically average Division I opponent.
How do the Micro offensive and defensive statistics differ from each team’s raw statistics?
The Micro offensive and defensive statistics are termed ”equitable statistics” as they allow objective comparison of all teams in each statistic. The equitable statistic is developed by adjusting the raw statistic considering opponents’ strength in the reciprocal statistic. For example, a team facing opponents with strong FG% defenses would have its raw FG% adjusted upward, while a team facing facing opponents with poor FG% defenses would have its raw FG% adjusted downward. The resulting equitable statistics represent projected performance versus a statistically average Division I opponent. Additionally each statistic is weighted for greater emphasis on more recent games and accounts for home and road results.
Why are ratings and projections not available for each team’s early season games?
The iterative process described above to establish initial ratings for each team requires four games’ worth of results, a sample size considered sufficient to represent team performance. Due to this approach, Omni Rankings does not publish projections for any game involving a team playing one of its first four games, while each team’s ratings are withheld until that team completes its fourth game.
My team has a better win-loss record than our rivals and defeated them earlier in the season. Why are our rivals currently ranked higher?
The ratings reflect the model’s assessment of each team’s current strength, not a grade of the team’s overall performance for the season. If the rival team has performed better than your team in recent games — based on the ratings criteria, not winning percentage — the rivals may overcome that earlier loss to your team and an inferior win-loss record to achieve a higher current rating.
How are the conference basketball tournament projections generated?
Omni Rankings projects each team’s conference record to establish the seeding, bracket, and game sites (with current Omni Rankings ratings serving as tiebreakers until the conference standings are final). Game projections are made based on current Omni Ranking ratings and the game site until a projected tournament champion is determined. This projected champion is considered the automatic bid qualifier for the respective conference in Omni Rankings’ NCAA tournament field.
How are the top performances, surprise performances, and upset wins determined?
Top performances represent the highest individual game ratings in a particular season, surprise performances indicate the largest exceedance of pre-game ratings projections, and upset wins reflect the greatest pre-game ratings deficits overcome.
If you have additional questions or comments about Omni Rankings or the ratings methodology, feel free to email via email@example.com. Thanks for visiting OmniRankings.com, named seventh-finest college sports numerical analysis website by Minutiae Obsession magazine.