This is everything you need to know about the RPI. Probably more. Please check here
first before sending me e-mail with questions about the ratings. It'll save us both time.
Thanks to NCAA Senior Statistics Coordinator Gary K. Johnson, keeper of the official RPI, for providing some of this information.
The RPI (Rating Percentage Index) is a measure of strength of schedule and how a team does against
that schedule. It does not consider things like margin of victory, only
whether or not a team won and where the game was played. It is used by the NCAA as one of their factors in deciding which
teams to invite to the NCAA tournament and where to seed them.
It was created in 1981 and is maintained by the NCAA Division I Men's Basketball Committee. They have always
placed a premium on schedule strength when selecting teams for the tournament, so they wanted a relatively
simple way to measure that and the RPI was born. This page presents an independent duplication of those ratings.
Just how important is the RPI?
As far as getting into the tournament, it appears to be more important to bubble teams than the top powers.
It is also important in the seeding process.
What is the formula?
The basic formula is 25% team winning percentage (WP), 50% opponents' average winning percentage (OWP),
and 25% opponents' opponents' average winning percentage (OOWP).
For the 2004-05 season, the formula was changed to give more weight to road wins vs home wins. A team's win total for RPI purposes is 1.4 * road wins + neutral site wins + 0.6 * home wins. A team's losses is calculated as 0.6 * road losses + neutral site losses + 1.4 * home losses.
For example, a team that is 4-0 at home and 2-7 on the road has a RPI record of 5.2 wins (1.4 * 2 + 0.6 * 4) and 4.2 losses (0.6 * 7). That means that even though it is 6-7, for RPI purposes, it is above .500 (5.2-4.2).
This "weighted" record is only used for the 25% of the formula that is each team's winning percentage. The regular team records are used to calculate OWP and OOWP.
As always, only games against Division I opponents count in the RPI.
Do you adjust the OWP to account for the fact that they have played the team in question?
Yes. Games against the team in question are ignored.
Does the NCAA make these adjustments too?
Yes.
In the OWP, do you include the records of teams scheduled in the future?
No. A team's is included as an opponent in the calculations only after the game is played.
This is because a team's entire schedule is not usually known, mostly due to games added because of in-season and conference
tournaments. Also, sometimes games get cancelled.
Do you use the record of an opponent at the time the game is played or at the time the RPI is calculated?
The RPI is always calculated based on current records, not date-of-game records.
If a team plays somebody more than once, is their winning percentage included more than once?
Yes. Once for each time a team plays that opponent. And all games against the team in question are removed from each occurrence of that opponent's record when calculating OWP.
What about the NCAA's secret adjustments?
The secret adjustments no longer exist under the new formula.
How do you figure a conference's RPI?
It's simply the average of the RPIs of the teams in the conference. The NCAA does this both ways: with and without conference games.
I provide both numbers as well, but rank them by non-conference. I feel that conference games are like a team playing itself, so they
shouldn't count. It should be noted that this data is among the least relevant in the selection process.
How do you figure the Strength of Schedule?
It's 2/3 opponents' winning percentage and 1/3 opponents' opponents' winning percentage, the same ratio as in the RPI formula.
Why do the rankings seem so weird early in the season?
The RPI is primarily a measurement of strength of schedule and how the team did
against that schedule and each season stands on its own. There are no preseason expectations - everyone starts at zero.
Consequently, the numbers can really skewed early in the season because teams have played so few games.
For example, a team can be 2-0, but its opponents haven't played anyone
else yet, so its strength of schedule is 0. That's 75% of the formula, so the team won't be
rated very highly. The data starts to take shape and look reasonable as conference play
gets into full swing around the end of December. And the more games played, the better the data gets.
Why is my team's record wrong?
The RPI is calculated using games between full-fledged Division I (D-I) opponents only. All full-fledged D-I teams are listed in the rankings. If a school plays a team not listed in the rankings, that game doesn't count and won't be reflected in the records.
Why isn't your team name here listed in the RPI?
Some schools are in the process of moving up to full Division I status, but are not quite there yet. It used to be that teams would not be considered D-I for the purposes of the RPI until they became full-fledged members of Division I, but now they will count in the RPI if they are far enough in the reclassification process and play a full Division I schedule. That is defined as one with no more than four games against teams outside of Division I. Only games against teams listed in the RPI count in the RPI calculation.
Why is my team ranked behind this other team when we beat them?
The RPI is a measurement of strength of schedule and how you did against that schedule. The ENTIRE schedule. To look at
only one game ignores the "any given day" aspect of sports. In 1998-99, for example, Iowa lost to Creighton, who lost to
Drake, who lost to Evansville, who lost to East Carolina, who lost to James Madison, who lost to Maryland-Eastern Shore,
who lost to Delaware St, who lost to New Hampshire, who lost to Yale. No one in their right mind would have rated Iowa behind
any of those teams, let alone all of them.
Why did my team move up after a loss (or down after a win)? Why did my team move up (or down) when they didn't play?
There could be several reasons. One is that since the RPI measures strength of schedule, so it is possible to have a higher RPI after playing a good team, win or lose. Also, a team could have its RPI move down by playing a bad team, win or lose. A team's RPI can also change when they don't even play if one of the teams on their schedule played.
Another thing is that rankings are relative, meaning that a team's ranking is affected by the results of teams around them in the rankings. A team could move in the rankings without even playing if the teams around them played. This is the way computer models can work and is often the most confusing difference for people used to following polls.
How is the RPI different from the Sagarin ratings?
There are two main differences between the RPI and Sagarin's ratings:
The RPI only cares whether or not a team wins, whereas Sagarin takes into account other things like the margin of victory, etc.
Sagarin's ratings also have a starting point, based on the previous few years. That starting point is no longer included by the end of the year, however, so that
each year can stand on its own. The RPI starts everyone from scratch. That's why it's not terribly
informative until around midseason.
Also, Sagarin doesn't really release any information about his formula, since he invented it and sells the
results, while the basic formula for the RPI is available to anyone who has a computer and too much time on their hands.
Sometimes my paper has the RPI in it. Is that your list?
No. Someone else is doing the same calculations I am and selling it to the media. This site is the only place to see my numbers.
Why are your numbers different than the ones I see somewhere else?
I cannot answer this question specifically, because I don't have time to troubleshoot other people's numbers. However, there are only two ways for the numbers to be different - either the formula is different or the data is different. Well, three if you count rounding error.
Under the new formula (starting with the 2004-05 season), the calculation of the RPI is pretty easy, and I am doing it correctly. However, game location is now important in the formula, and determining whether a game is a home game or a neutral site game can be tricky. If I have it one way and someone else has it the other, we would get slightly different results in our RPI calculations. When there's doubt about game location, I ask the NCAA for clarification to make sure my data is correct.