RPI Calculation Steps

September 27, 2019 Baseball Blog Updates

RPI Calculation Example using Ashland Blazer 2018 Football Regular Season

RPI Background

• RPI stands for “Ratings Percentage Index.” In short, it is a way to measure a team’s strength relative to other teams, based largely on the strength of their schedules.
• The RPI formula is used by both the NCAA and NAIA, along with many high school associations and among other organizations, as part of their postseason system.
• RPI is calculated from the team’s Winning Percentage (WP), the Opponent’s Winning Percentage (OWP) and the Opponents’ Opponents’ Winning Percentage (OOWP).
• The KHSAA is uniquely positioned to be the sole and authenticated source of RPI data due to the maintenance of required score reporting and scoreboard data for Baseball, Basketball, Field Hockey, Football, Soccer, Softball, and Volleyball.
• These three components are combined to produce the RPI using the following formula:
  • RPI = (WP * WPVAL) + (OWP * OWPVAL) + (OOWP * OOWPVAL)
  • The KHSAA has adopted for its RPI using the following final calculation:
    • WPVAL shall be .35 (or 35 percent)
    • OWPVAL shall be .35 (or 35 percent)
    • OOWPVAL shall be .30 (or 30 percent)
• In all of the calculations the following values are assigned to game results when calculating WP throughout the compilation:
  • A win is assigned a value of 1.0 as a basis for comparison;
  • A tie is assigned a value of 0.5;
  • A loss is assigned a 0.0 value.
• Games against out-of-state teams are automatically assigned a value of .50000 for the winning percentage in the final calculation for all sports except football. Thousands of results throughout the country in a variety of states that have used the RPI reveal that in the end, the relatively few games played against out-of-state teams over the years average out and comparisons involving actual schedule results vary minimally statistically from that .50000 result.
• This out-of-state value is also assigned in the WP calculation for an in-state home school competing against a member school.
• For football only, the out-of-state value is set to the previous year’s average winning percentage for in-state games. For 2021, that value is .51060 based on the 2019 season.
• For football only, a game value factor is included in each WP calculation as described in the football calculation example, based on the class of the team being played for each calculation and is only relevant to in-state games.
• This weighting is based on a 15% difference for the game value as class sizes become larger.
• Those game value factors are:
  • 1A = 1.323
  • 2A = 1.521
  • 3A = 1.749
  • 4A = 2.011
  • 5A = 2.313
  • 6A = 2.660
  • There is currently no special weight utilized for 6-player games, which would be the baseline of 1.0 for a game value, or for 8-player games, whose value would be 1.15
• An RPI of 0 is likely not possible in any scenario. From a mathematical perspective, a team would have to be winless, each of their opponents winless, and each of their opponents’ opponents winless.

Basic RPI Calculations (Not including the Game Value factor calculation exception for Football)

For sports other than football, here’s how WP, OWP, and OOWP are calculated.

• WP is the Winning Percentage of the team being considered.
  • The WP is computed by taking the number of wins and dividing by the number of games played. A 15-4 team would have a WP of 15 wins/19 games = 0.789.
  • Ties are counted as 1/2 of a win (a win value of .5), so a 15-3-1 team would have a WP of (15+.5)/19 = 0.816.
• OWP is the Opponent’s Winning Percentage which is the average of the WP of each of the opponents the team has played.
  • The WP for the each of the opponents is calculated like the WP for the team with one big exception: none of the games with the team whose RPI is being determined are included when calculating the WP of the opponents.
  • As an example, consider an opponent has a 9-1 record and that the only loss is against the team under consideration. Excluding that game, the opponent has a 1.000 record, which is what will be used in computing its relative schedule strength.
  • Click here to view the example from Ashland Blazer which illustrates that when calculating the OWP, each of Ashland Blazer’s opponents’ winning percentage is calculated. But in each one of those, the game against Ashland Blazer is excluded. A win over Ashland Blazer does not help the WP value used in the calculate Ashland Blazer’s OWP, nor does a loss hurt that WP. That game is not included in the calculation.
• OOWP is the Opponent’s Opponents’ Winning Percentage and is calculated by looking at each opponent and calculating their OWP, similarly excluding the games between the opponent and their opponents. The OOWP is the average of the OWP for each opponent.
  • The exclusion of the considered team extends to the calculation of the OOWP as well.
  • Click here to view the example from Ashland Blazer example to see how each of their opponents now becomes the compared team. So when performing the calculation for Raceland, the WP for each of their opponents is calculated. Therefore when performing the calculation, the WP for Greenup County must be computed. The game between Raceland and Greenup County will not be used in the calculation for Raceland’s WP. However, the game between Ashland Blazer and Greenup will be calculated within the OOWP.

Winning Percentage Calculation for Football

Football is a special case, and enhancements are made to the winning percentages. These percentages are adjusted by multiplying by a game value based on the class of the opponent in football, primarily due to the limited number of opportunities to play contests and to compensate for schools playing in a different class than they might otherwise be placed by enrollment.

• It should also be noted that by using class weights, as is done in football, it is possible to get WP greater than 1.000. It is possible to get OWP, OOWP, and even RPI that is greater than 1.000 due to class weights.
• Kentucky has six football classes, 1A through 6A, which generally reflects the enrollment of the school. The WP calculations are adjusted to give more weight to a game when a school “Plays Up” by playing a higher class and give less weight to games when a school “Plays Down” by playing a lower class.
• This weight is used to adjust the value of a win depending on the team’s class and the opponent’s class. If the game is a win for the team, the Winning Percentage of the game is the class weight (the Winning Value in the example) for the opponent divided by the class weight of the team (the Game Value in the example). This gives the Winning Percentage for that game.
• Click here to view the example to better understand the calculation and the use of the Game Value factor, look at the first table in the Ashland Blazer example.
• Consider the game between Ashland Blazer and Harlan County. Ashland Blazer (a 4A class team) won against Harlan County (a 5A class team). So, the Winning Percentage for this game is 2.313/2.011 = 1.15017. On the other hand, consider the game between Ashland Blazer and Russel (a 3A class team). Ashland won this game, too. So, the Winning Percentage of this game is 1.749/2.011 = 0.86972.
• This produces the desired outcome as Ashland Blazer “played up” in class when it played Harlan County, so the Winning Percentage of that game is greater than the win against Russell because Ashland Blazer “played down” in class.

Click here for a complete example using the schedule of Ashland Blazer during the 2018-19 football season to fully understand the RPI calculations