Saturday, August 29, 2009

Change the RPI Formula ??

It's Back-To-School time, so here's some heavy reading and links for even heavier reading.

Greg Van Zant, the head coach at West Virginia University, recently sent a detailed recommendation to the NCAA Division I Baseball Committee to change the formula for calculating the Ratings Percentage Index (RPI).

CollegeBaseball Insider includes the complete text and details in Proposal for Changes in Computing RPI:
As you know, the Ratings Percentage Index (RPI) is a very important tool that ranks NCAA Division I Baseball teams from 1 to 302. This RPI ranking aids the NCAA Division I Baseball Committee in a very difficult job of selecting at-large teams for the NCAA Baseball Tournament. Therefore, it is extremely critical that the RPI formula produces the most accurate ranking possible.
. . .

There are four different facets of the RPI that need to be examined:
1. The weighting of the three RPI components or factors.
2. The weighting of home wins and road wins.
3. The four exempted contests against non-Division I opponents.
4. The bonus and penalty system.

My proposal is as follows:
1. Change the weighting of the three RPI factors to 50-35-15.
2. Count a road win as 1.25, a neutral win as 1.0 and a home win as 0.83333333.
3. Penalize teams for playing non-Division I opponents.
4. Eliminate the bonus and penalty system.

. . .
■ Baseball America summarizes the proposal in with Time for a Tweak? (
His solution is to replace winning percentage in the RPI with "adjusted winning percentage." Simply put, a win on a neutral field would still count as 1.0 win, but road wins would count as 1.25 wins, and home wins would count as 0.833 wins. That way, if two equally matched teams played 20 games at one team's home site, and the home team won 12 of the games while the visitor won eight, both teams would wind up with identical adjusted winning percentages.
■ In Why does RPI Matter? , Kelly Kulina argues that northern schools are disadvantaged by the current RPI formula (
As with anything involved with the postseason, the RPI and its role in postseason selection does not come without controversy. Critics of the RPI point out that it compounds the disadvantage cold weather teams face against warm weather teams. Since they must spend the first three weeks of the season on the road, many cold weather teams will start with a losing record and play against other teams in their conference with losing records. This creates the compounding effect causing the RPI disadvantage. In fact, many lower tier teams in warm weather conferences will get an at-large nod over upper tier teams from cold weather conferences. If you take a look at the past few years’ College World Series participants and winners, you’ll see a major disparity between cold weather teams and warm weather teams.
■ Boyd Nation of, along with some other interested followers of the game, express their opinions in a message board thread on, Proposal to change the RPI in baseball:
boydnation: There was some support, but mostly just in the sense that everyone but the Southeastern beneficiaries of the RPI want it to be changed. There's no sense that anyone, Western coaches included, actually care about going to a more accurate system; everyone's just looking for the solution that most benefits them. . . . As to Van Zant's proposal, there's a lot of thought in it and a lot of really bad statistical work -- mostly he's doing what we call wishcasting, where he decides what he wants the answer to be and then declares it to be correct. If I get bored, I'll actually code up his proposal and see what it does.

No comments:

Post a Comment