Virtually VarsityBeta

Understanding RPI (Ratings Percentage Index)

The RPI is a ranking system used in various sports to help compare teams' strength of schedule and overall performance. Let's break down how it works using a few examples.

Basic Components of RPI

RPI consists of three main components:

  • Winning Percentage (WP) - 25% of RPI
  • Opponents' Winning Percentage (OWP) - 50% of RPI
  • Opponents' Opponents' Winning Percentage (OOWP) - 25% of RPI

1. Winning Percentage (WP)

This is the simplest component - it's just wins divided by total games.

Team A has 8 wins and 2 losses

WP = 8 ÷ 10 = 0.800 (or 80%)

2. Opponents' Winning Percentage (OWP)

This is where things get interesting. Let's walk through a detailed example:

Suppose we're calculating the RPI for "Team A". Here are Team A's opponents:

Team B: 7-3

Team C: 5-5

Team D: 3-7

Important Caveat

When calculating OWP, we remove all games against Team A from each opponent's record. This prevents circular influence. Here's why:

Let's say Team A played Team B twice and won both games. Team B's original record is 7-3, but two of those losses were against Team A. So for OWP purposes, we remove those games:

Team B's Original Record: 7-3

Remove games vs Team A: -0 wins, -2 losses

Team B's Adjusted Record: 7-1 (87.5%)

Why Remove Games Against Team A?

Let's illustrate why this is important with an example:

Scenario 1 (WITHOUT removing games):

Team A is 10-0

Their opponents' combined record is 30-70

Raw OWP would be 30%

Scenario 2 (WITH removing games against Team A):

Team A is 10-0

Their opponents' combined record was 30-70

After removing the 10 losses against Team A:

New opponent record: 30-60

Adjusted OWP is 33.3%

3. Opponents' Opponents' Winning Percentage (OOWP)

This follows the same principle but goes one level deeper. For each of Team A's opponents, we:

  1. Identify all of their opponents
  2. Calculate their winning percentages
  3. Average them together

Final RPI Calculation

Once we have all components, we apply the weights:

RPI = (0.25 × WP) + (0.50 × OWP) + (0.25 × OOWP)

Example:

WP = 80%

OWP = 56.9%

OOWP = 55%

RPI = (0.25 × 0.800) + (0.50 × 0.569) + (0.25 × 0.550)

= 0.200 + 0.2845 + 0.1375

= 0.622 or 62.2%

Counter-Intuitive Results

Sometimes, removing games against a winning team can actually lower the OWP. This happens when:

  1. The team has beaten several strong opponents
  2. Lost to several weak opponents
  3. Removing these games disproportionately affects the strong teams' records

Team X (8-2) has beaten:

- Top Team (9-1, becomes 8-1 after removal)

- Good Team (7-3, becomes 6-3 after removal)

- Average Team (5-5, becomes 4-5 after removal)

Original average: (9-1, 7-3, 5-5) = 70% win rate

Adjusted average: (8-1, 6-3, 4-5) = 67% win rate

This shows how the adjustment process, while more accurate, can sometimes lead to seemingly counter-intuitive results while still reflecting the true strength of schedule.