The NCAA Tournament and the RPI

By Ben Cohen

With the NCAA tournament?s selection committee meeting in less than a month to choose the field?s 68 teams, the Journal spoke Wednesday with Xavier athletic director Mike Bobinski, the committee?s chair, about the RPI, NCAA tournament expansion and how the committee prepares. The interview has been condensed and edited.?

How do you prepare for selection weekend, and when do you start?

It begins when the absolute first game tips off. We meet as a committee in the beginning of November and set the strategy and give everybody their marching orders.

The single biggest thing a committee member has to figure out is how to process all the information that?s available to you and how you individually can be best prepared. It?s a huge data management task.

The committee says that the Ratings Percentage Index (RPI) is just one of the tools that it uses for selection. It seems to be the benchmark used most often?even though it has been criticized.?Why do you think the RPI is the best evaluative metric?

Whether it?s the best is a matter of opinion. I?ve become more comfortable with our use of it over time because I?ve really come to know now that I?m in the room that we don?t ever pick teams based on their RPI. We don?t ever look at teams and say, their RPI is a certain number, therefore they must be in the field. It doesn?t work that way.

We do absolutely use the RPI when we evaluate a team?s performance against the top 50 or top 100 or their strength of schedule. It?s absolutely a tool we use to organize the field. We use it to group teams in certain ways. That?s undeniable that we use it from that perspective. What people sometimes believe is we use the RPI as a selection tool, and we never, ever do that.

It?s a consistent way of organizing the data and slicing and dicing things, but we look a lot deeper than that.

The RPI is descriptive, not predictive.

It?s based on history. It?s based on what happened.

We?re not in the business of predicting what will happen in the tournament. Yes, we seed the field, but we seed the field based on history and how we?ve evaluated what teams have accomplished, not based on what they?re going to do. We don?t ever try to predict the results of the tournament. The beauty of the tournament is its unpredictability.

You said recently that the NCAA conducted a study comparing the RPI to other metrics and found that ?the RPI actually did end up with the highest level of predictive value and the highest correlation with ultimately success in the tournament.? But that was refuted yesterday.

I misspoke when I said that. I should?ve kept my mouth shut.

It may not be the best predictive measure, but it?s very, very close. But that?s purely for information. We?re not trying to predict the result.

Most of the smartest analysts of college basketball prefer metrics other than the RPI. If the committee considers the RPI one of the most reliable systems, why do so few others use it?

I think part of that is certain folks gravitate toward other ranking systems because of particular features they may prefer that may not be present in the RPI. Some people like margin of victory or different weighting on style of play.

I use a combination of different systems when I look at teams and also weigh that with my own observations. I want to see a team play and get a sense of who they are and how they look.

Will the committee ever factor in margin of victory, or is that a non-starter?

I don?t see that becoming a criteria that we would ever place any great weight on. The bottom line is you win or you lose, and a close win is every bit as good as a big win. You ask any coach in America and I?d guarantee they?d say the same thing.

It?s possible to game the RPI by scheduling in a way that would inflate a team?s RPI ranking. Is that a concern?

People may think they?re gaming it. But we never look at that number in and of itself. It?s always what makes up that number ? particularly if it?s a bit of a head scratcher, either a little bit higher or lower than you might think. We dig much deeper than the pure numbers.

So there are ways for sure to come up with very favorable looking RPIs by how you organize the schedule. We?ve all figured that out. It?s a well-known fact in the committee room how that gets done. We really look at the essence of a team?s profile and not so much the pure numbers. We have the time and commitment and resources to dig a lot deeper.

Would the committee?s job be more difficult if the NCAA tournament were 96 teams instead of 68?

I like where we are in terms of 68. I think 96 would present the same problems because there will always be a 97th team, just like there?s a 69th team right now. That will never change. I don?t think the number matters as much as we want the games to be of such interest that people will care about who did and didn?t get in the field.

Source: http://blogs.wsj.com/dailyfix/2013/02/20/the-ncaa-tournament-and-the-rpi/?mod=WSJBlog

christina hendricks lifelock camp david hawaii weather the jerk lake havasu halo 4