RPI numbers just don't add up

RPI numbers just don't add up

Published Feb. 3, 2011 12:00 a.m. ET

When college basketball feared the pace of the game was dragging, it added a shot clock. When the lane got too crowded, the three-point shot arrived.

When more programs started going gaga about the sport, the NCAA Tournament was expanded. Multiple times. Look for a record 68 teams this March.

When the NCAA Tournament Selection Committee needed help picking the at-large teams and seeding the field, the NCAA created a computer formula.

Hey, nobody’s perfect.


That certainly includes this formula, the Ratings Percentage Index (RPI), which was created in 1981. The RPI has had coaches reaching for antacids ever since, especially after they learned this Really Puzzling Information is one factor the committee consults on Selection Sunday.

If the shot clock, three-point line and expanded tournament field have enhanced the game, the RPI has been whipping up a blizzard of biting criticism for 30 years.

Including 2011.

There is one remaining unbeaten college basketball team today. That team plays in the Big Ten, a conference generally considered America’s best or second-best league.

That team has already won half of its conference games. It has won by 18 at Florida, which looks like the best team in the Southeastern Conference. It has also won by 14 at Florida State, a formidable contender in the Atlantic Coast Conference.

That flawless team, of course, is 22-0 Ohio State.

The Buckeyes are ranked first in the coaches’ poll and the writers’ poll as well as in the three alternative computer formula polls created by Jeff Sagarin, Ken Pomeroy and Peter Wolfe.

There is one bizarre, persistently dissenting voice — the RPI.

The RPI tries to tell you the Buckeyes are no better than the fifth-best team in the nation, behind Kansas (one loss), Georgetown (five — count 'em — five defeats), Brigham Young (two losses) and San Diego State (one loss).


That’s only the beginning. The RPI has Georgetown second in the nation. Patrick Ewing and Alonzo Mourning would have a tough time arguing that one. The Big East standings show the Hoyas tied for sixth in the league.

Duke is the best team in the ACC, a likely top seed for the NCAA Tournament. Don’t believe any of that, Mike Krzyzewski. Your Blue Devils are swimming at No. 11 in the RPI.

I could go on. And I will go on. In a paragraph or two.

“You could say there are some inconsistencies and issues with the RPI,” said Pomeroy, a Virginia Tech grad who created his mathematical ratings formula in 1999.

This is what the RPI is: A formula that ranks teams primarily on the basis of wins and losses and strength of schedule.

A team’s winning percentage accounts for 25 percent of its rating. The winning percentage of its opponents is responsible for 50 percent. The final 25 percent is provided by the winning percentage of its opponents’ opponents.

Here is one critical thing the RPI does not do: put any weight on margin of victory.

An overpowering, no-doubt-about-it, 25-point victory is of no greater value than a game won on a banked-in three at the buzzer. That’s silly. But that’s the RPI.

Looking for another nit to pick with the RPI?

Try this: The formula was altered several years ago to provide more credit for road wins than home wins. There is some merit to that. Home-court advantage is generally considered worth three or four points in college hoops.

But here is the problem: A Big East team like Georgetown can generate more juice from beating Rutgers, Seton Hall and Villanova on the road than Pittsburgh has obviously earned from defeating Connecticut, Marquette and Syracuse at home.

Remember this too: On Jan. 12, Pitt, the Big East leader, went to Georgetown and won by 15 points.

Apparently that news failed to resonate within the RPI formula because coach Jamie Dixon’s Panthers are ranked eighth in the RPI, an inexplicable six spots behind the Hoyas.

There’s more. There’s always more. Six road victories have Brigham Young tucked third in the RPI — even though the Cougars are not ranked higher than eighth by Sagarin, Pomeroy or the two human polls.

The RPI loves Tennessee (18th) and its seven defeats. The Vols can’t crack the Top 25 in Pomeroy, Sagarin or the human polls.

Looking for the flip side? Washington sits in the top 20 in both human polls as well as in Pomeroy and Sagarin. You’d have to scroll to No. 23 to find the Huskies in the RPI — and that’s after they started this week outside the Top 30

“I believe my formula is closer to Jeff’s than the RPI,” Pomeroy said.

Sagarin has an advanced degree in mathematics from the Massachusetts Institute of Technology. He lives in Bloomington, Ind. Pomeroy's background is actually in meteorology. He works for the National Weather Service in Salt Lake City. Both guys love the game.

Like Sagarin in his preferred Predictor ratings, Pomeroy absolutely believes that margin of victory counts. Pomeroy’s formula also places increased weight on games played later in the season because he believes they are a more accurate indicator of how a team is performing.

And why would the RPI like Brigham Young better than Ohio State?

“Probably because BYU has those five road wins and because Ohio State doesn’t really have a marquee conference road win yet,” Pomeroy said.

“But I think we’ve got Ohio State (first) and BYU (10th) rated where they should be rated.”

So do I.