This post is the second in a series I am doing on the RPI. My hope is that it will help clarify the mystery surrounding the formula. If you have questions or comments, I would love to hear from you and attempt to address your concerns in a future post. All posts in this series will be available under the “My Thoughts” tab at the top of the page.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
NOTE: THIS POST WILL ONLY BE DEALING WITH THE BASICS OF THE RPI FORMULA. I FELT IT NECESSARY TO START SIMPLE, BUT I PLAN ON GOING INTO MUCH MORE DETAIL IN FUTURE POSTS.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Most college basketball fans know about the RPI. They know it is a computer ratings system that the Selection Committee uses to help make decisions. They know that it gets quoted about 10 times a minute on ESPN during the months of February and March.
Most college basketball fans also know about the movement of the RPI from day to day. A very common conversation on message boards is one trying to answer the question "How much will we move up/down after a win/loss to team X?" Often, a common frustration with the RPI is that a team can win a game and actually drop in the next day's RPI ratings. In the same way, a team can move up after a loss. Obviously, this seems counter intuitive, and many fans dismiss the RPI entirely as soon as they find out that these types of things are possible.
Despite all this knowledge about the RPI, many fans don't really understand what the formula is or how it works. Let's start with the formula. It is actually quite simple. Here is the formula for the RPI of Team A.
(0.25 x Team A's winning %)
+
(0.5 x Team A's opponents' winning %)
+
(0.25 x Team A's opponents' opponents' winning %)
=
RPI
Does this look confusing? It really isn't all that complicated. For example, if Team A won 80% of their games, they played opponents who won 60% of their games, and the opponents of their opponents won 48% of their games, then Team A's RPI would be:
(.25)(80) + (.5)(60) + (.25)(48) = .62
The number 0.62 would then be compared to the RPI of every other team in the country and they would be ranked from highest to lowest. For example, 0.86 might be the highest RPI value of any team in the country, making that team #1 in the RPI. A team with an RPI of 0.50 would probably be very average. A team with an RPI of 0.21 might be near the very bottom. RPI's can range from 0 (worst) to 1 (best). When you hear that a team's RPI is 39, that means that their RPI value is the 39th highest out of all the teams in the country.
Finally, there is one more component that I have to mention. Starting in 2004, the NCAA changed the RPI formula slightly to help differentiate between home games and road games. This was an effort to level the playing field since some teams (usually from BCS conferences) played almost all of their non-conference games at home while other teams (usually from smaller conferences) played almost all of their non-conference games on the road. This final component is a multiplier that makes home wins less valuable than road wins, and road losses less penalizing than home losses. It is also a simple formula.
Home wins count as 0.6 wins
Home losses count as 1.4 losses
Road wins count as 1.4 wins
Road losses count as 0.6 losses
This affects all of the percentages listed in the orginal formula. It means that a 10-10 team may or may not have a .500 winning % in the RPI formula. If they played 10 home games and 10 road games, it will indeed be .500, but if they were 9-5 at home and only 1-5 on the road, they still finished 10-10 (.500) but the number entered as their winning % in the RPI formula would be different. The RPI would see their record as:
(0.6 x 9) + (1.4 x 1) = 6.8 wins
(1.4 x 5) + (0.6 x 5) = 10 losses
6.8 + 10 = 16.8 games
6.8 / 16.8 = .405 winning %
Notice the difference between the 2 examples. Our 10-10 team that played 14 games at home and only 6 on the road actually had a winning % that looks like they only won 8 of their 20 games. This is the "penalty" that the NCAA has placed on teams that play more than half of their games at home. This also works the other way, as teams who play more than half of their games away from home are "rewarded".
Hopefully this helps you to better understand how the RPI works in terms of the formula itself. If you still have questions, feel free to comment on this post and I'll try my best to answer.
No comments:
Post a Comment