A Madness To The Method? A top-secret computer formula, the RPI, is often the difference between a team's receiving an NCAA at-large berth or going to the NIT. Its critics think there are better yardsticks--and they may be right

One day last May, Gary K. Johnson, senior assistant director of
statistics for the NCAA, boarded an airliner and greeted the man
seated beside him. Johnson's seatmate grunted hello back, lost
himself in a book and only upon landing noticed the NCAA logo on
Johnson's briefcase. The man asked what he did, and Johnson
explained that he was headed for Conference USA's spring meetings
to answer questions about the statistic to which Johnson devotes
most of his late winters--the Ratings Percentage Index, or RPI.
It's a measurement the NCAA men's basketball committee uses to
help decide which teams get into the national championship
tournament.

"You do the RPI?" the passenger blurted. All bustle stopped.
Heads turned. Jaws dropped from full upright and locked position.
For a moment Johnson might have been transported to one of those
old E.F. Hutton commercials. "I'm not famous," Johnson says. "But
I guess the RPI is."

Once a cryptic formula known only to the cognoscenti, the RPI has
become a set of numbers that will, from now through Selection
Sunday, March 16, dominate the blogosphere of the Web and the
blabosphere of sports talk radio. Until he started charging a fee
in 2002, Jerry Palm, the self-described geek who updates the RPI
daily at collegerpi.com, was attracting more than a million page
views per week during the run-up to the tournament. Says former
Big Ten commissioner Wayne Duke, chairman of the committee when
the RPI was introduced 23 years ago, "It's taken on a life of its
own."

The NCAA has always gone out of its way to play down the
importance of the RPI in determining the at-large entries in the
65-team championship field. The basic formula--25% your winning
percentage, 50% your opponents' percentage and 25% your
opponents' opponents' percentage--is, committee members insist,
only one evaluative tool among many used, and it has been relied
on less and less over the years. It's most useful, they say, in
separating teams into tiers and in differentiating one bubble
team from another.

Nevertheless, the RPI has made schools think hard about whom they
schedule. Nothing else can explain why Butler, bidless a year ago
despite 25 wins and a perfect nonconference record, paid a
midseason visit to Duke on Jan. 30, absorbing an 80-60 defeat
but measurably boosting its strength-of-schedule rating. What's
more, for fear that laggards might freeze their leagues out of
one or two of the NCAA tournament's lucrative at-large bids, the
Missouri Valley and Sun Belt conferences have been reserving
$500,000 of their annual NCAA tournament payout for an incentive
pool that rewards members who hit minimal targets in the RPI.
Last Saturday such historical postseason mischief-makers as
Creighton, Gonzaga, Kent State, Southern Illinois and Tulsa
joined 13 other teams for Bracket Buster Saturday, a nine-game
event designed to give seven mid-major conferences a late bump in
the RPI. (The biggest such bump went to Southern Illinois, whose
66-64 win over Wisconsin-Milwaukee raised their RPI number by
.0124 and vaulted them from 67th place to 55th.) No wonder
Johnson and his fellow NCAA statisticians call the RPI "the
monster."

However did we get here?

Beginning in the late 1970s, the NCAA tournament underwent a
welter of changes, including the expansion of the field from 40
to 48 teams, the lifting of the limit on how many bids a
conference could receive and the seeding of all teams in the
bracket. As committee members groped to sort out the many
candidates now in the at-large mix, they asked the late Jim Van
Valkenberg, then the NCAA's director of statistics, to come up
with some empirical measurement of schedule strength. If you
think of conferences as high schools, and teams as kids applying
to a selective college, Van Valkenberg's creation was to be a
kind of standardized test, something the admissions officers on
the committee could use to tell one 22-8 team from another.

The original RPI was based 40% on a team's own winning
percentage, 20% on its opponents' percentage, 20% on its
opponents' opponents' percentage and 20% on road winning
percentage. But that last factor was skewing the best teams far
ahead of the pack, because the elite tend to win regardless of
where they play. So within two seasons the factor for road record
was removed, and the formula became 20% a team's winning
percentage, 40% its opponents' winning percentage and 40% its
opponents' opponents' percentage. A decade later the unlikely
success of the College of Charleston--the best team in the
chronically weak Trans Atlantic Athletic Conference--forced
another change. "A very good team in a not-very-good conference
could hit league play and have its RPI drop even though it kept
winning, because strength of schedule was such a big factor,"
Johnson says. For the 1993-94 season Johnson and two math
professors again reworked the formula, this time reducing the
weight of opponents' opponents' winning percentage, which is
hardest for a team to control, and adjusting upward the values of
a team's own won-lost record and that of its opponents.

In 1991 a geologist moonlighting as a Big Ten newsletter
publisher, Jim Sukup of Carmel, Ind., began calculating the RPI
and publishing it in his Collegiate Basketball News. Sukup did a
brisk business with conferences and athletic departments, with
whom even the NCAA doesn't share the RPI until after the
tournament. By '93 Sukup was faxing daily updates to ESPN, and
two years later he began providing a weekly RPI to the Associated
Press. With the public announcement of the formula change before
the '93-94 season, several replications popped up on the Web,
most prominently the one posted by Palm, who worked as a systems
analyst for a bank in Chicago until being laid off a year ago.

"We made the big change in '94, but there's been a lot of
tweaking along the way," says Johnson, who now awards bonus
points for such factors as beating a team in the top 50 and
assesses penalties for, say, scheduling too many games against
teams ranked in the RPI's bottom half. These bonuses and
penalties determine the Adjusted RPI. At its meeting in December,
the committee voted to keep these adjustment criteria secret;
Sukup has developed algorithms that he believes approximates
them, while Palm simply posts the unadjusted RPI, taking Johnson
at his word that the adjustments can't be reverse-engineered. In
any case, Johnson says, "The bonuses and penalties don't change
the RPI a whole lot. We add them to try to get a little more
separation between bubble teams."

Though most of the changes over the years have been to the
advantage of the little guys, the RPI's many critics believe the
metric remains biased in favor of the power conferences. They
complain that the RPI doesn't take into account:

Common sense. Even after the reworking of the formula 10 years
ago, teams with major-conference pedigrees often lose yet,
because of schedule strength, go up in the RPI. Worse, a team
from a weak league can win and win--but go down. There's
something flawed about a ranking system in which a victor would
be better off not having played at all.

Point differential. The RPI ignores how much teams win or lose
by, for two reasons: Anything with a whiff of Vegas to it spooks
the NCAA, and the committee doesn't want coaches running up
scores to improve their RPI. "But you do keep score for a reason,
and the number of points you win or lose by has some
significance," counters Sukup, whose Collegiate Basketball News
has morphed into The RPI Report and rpiratings.com. "The NCAA
keeps track of scoring margin and publishes that." Sukup suggests
adding a factor that would reward victory margin, but not beyond
15 points.

Home court advantage. "There isn't enough weight to reward road
success, though everyone agrees that winning on the road is the
toughest part of college basketball," says Missouri Valley
commissioner Doug Elgin, who just finished a four-year turn on
the committee. A home team wins about two thirds of the time, and
the host in a nonconference matchup is much more likely to be the
higher-profile school. "The majors will play you, but only on
their home floor," says Morehead State coach Kyle Macy, whose
Eagles have lost at Arizona State, Alabama and Ohio State this
season. "But once your program starts to improve, some schools
won't play you again." Adds Elgin, "If [mid-major] teams could
schedule [more home-and-homes], it would be much more fair. [Two
seasons ago] Indiana made its first trip to Indiana State since
1908, and lost."

Johnson points to the RPI's early abandonment of the road-record
factor, and Palm says that the committee, having passed over
Georgetown last season and Alabama two years ago for their
reluctance to go on the road, makes sure to punish yellow
schedulers. But no metric can really evaluate schedule strength
unless it accounts for where games are played. "The major
conferences have a scam going," says Jeff Sagarin, whose own
ratings appear in USA Today and who proposes alternatives to the
RPI (box, page 55). "When you have huge imbalances between home
and away games, it's irrational to ignore them."

You don't have to be a conspiracy theorist to wonder about the
RPI. The NCAA parcels out millions to conferences according to
games won per tournament. Ergo, the more bids a conference lands,
the more money it stands to make. With the committee
disproportionately composed of representatives of the major
conferences (this season seven of the 10 members are from
major-conference schools) and the RPI consistently placing major
schools higher than other ranking systems, the metric looks like
the big-timers' errand boy. Counters Jim Livengood, chairperson
of the men's basketball committee and athletic director at
Arizona, "Every person on the committee shows compassion to all
schools and is dedicated to making our selections unbiased across
the board."

Palm believes that committee members, for the most part, are
conscientious. "Of course nobody believes they are, because
people are so cynical," says Palm, who began calculating the RPI
to better understand the selection process. "But only once has my
faith been shaken: In 1999 New Mexico got in when it had beaten
only one good team [No. 7 Arizona] all year, had a bad record in
the Western Athletic Conference and was 74th in the RPI--the
worst RPI to get an at-large bid since I started doing this."

Johnson is forever reminding people that the RPI isn't designed
to predict games, as some people do with the Sagarin Ratings.
"It's a tool to help the committee," he says. "If the RPI were a
say-all, there'd be no need for a committee." Nonetheless, he's
currently tabletopping yet another improvement to the
formula--though he won't say exactly what.

Until the RPI is perfected, Sukup and Palm hope fans know enough
not to shoot the messengers. "I get, 'You must be an Indiana fan,
because you rate them so high,'" Palm says. "I hate Indiana. I
went to Purdue."

Says Sukup, who holds degrees from Indiana and Wyoming, "A couple
of years ago someone from either Auburn or Alabama told me I
could shove the RPI up my Yankee you-know-what."

On his next flight south, Gary Johnson may want to check that
NCAA briefcase curbside.

COLOR PHOTO: JOHN BIEVER [T of C] SOONERS THE BETTER Guard Ebi Ere and Oklahoma rule in college hoops' critical RPI formula (page 48).

COLOR PHOTO ILLUSTRATION: PHOTOGRAPHS BY PETER GREGOIRE PHOTO; ILLUSTRATIONS BY JOHN BLACKFORD NUMBERS CRUNCHER As the NCAA's chief statistician, the embattled Johnson is constantly defending--and tweaking--the factors in the RPI.

COLOR PHOTO: BOB ROSATO

COLOR PHOTO: DARRON CUMMINGS/AP

COLOR PHOTO: BOB ROSATO STILL AFLOAT Kenny Walker and the slumping Tide haven't sunk far in the RPI.

COLOR PHOTO: BRENT SMITH BLOCKED Playing in a weak league has held back Joel Cornette and the Bulldogs.

COLOR PHOTO: PHOTOGRAPHS BY PETER GREGOIRE POINTS MAN Sukup would add a factor for margin of victory.

COLOR PHOTO: PHOTOGRAPHS BY PETER GREGOIRE TRUE BELIEVER Palm trusts the committee to do right--usually.

COLOR PHOTO: MANNY MILLAN KANSAS CAN The Predictor boosts Nick Collison and the Jawhawks.

COLOR PHOTO: JOHN BIEVER TO THE TOP According to the RPI, Hollis Price and surging Oklahoma are No 1.

COLOR PHOTO: DAVID E. KLUTHO AFTER KENTUCKY, KEYED by the stifling defense of Marquis Estill (50) and Chuck Hayes, beat then No. 85 Arkansas last week, it dropped from No. 2 to No. 6 in the RPI. But following Sunday's win over No. 20 Mississippi State, the Wildcats were No. 3--one slot lower than their AP ranking.

COLOR PHOTO: PHOTOGRAPHS BY PETER GREGOIRE ELO-MENTARY Sagarin has applied the rationale of world-class chess to hoops.

A SWEETER SCIENCE?
By factoring in the game's complexities, Jeff Sagarin says his
formulas improve on the RPI

After a spate of upsets broke open the NCAA tournament in 1986,
including Cleveland State's victory over Indiana and
Arkansas--Little Rock's win over Notre Dame, an NCAA basketball
committee member told a reporter, "You can throw these computer
ratings in the garbage." After reading that passage at home in
Bloomington, Ind., Jeff Sagarin bolted for the phone. He called
the committee member to point out that his ratings had Cleveland
State 20th, even if the RPI ranked the Vikings 60 places lower.
The member instantly apologized. "You're the only reason we took
Cleveland State," he confessed to Sagarin. "They should have put
you in uniform on their bench."

Better yet, the committee should put Sagarin in its conference
room. Using a math degree from MIT and an M.B.A. in quantitative
analysis from Indiana, he has designed the most widely respected
system for gauging the relative strengths of teams and their
schedules. And he's quick to credit the man who inspired a
crucial part of his formula, the author of Sagarin's much-thumbed
copy of The Rating of Chessplayers Past and Present--Arpad Elo, a
Hungarian emigre physicist who died in 1992.

After he began playing competitive chess, Elo noticed that it was
possible to rise in the rankings despite losing and fall even if
you won--precisely the anomaly that leaves so many fans baffled
by the RPI. So in the 1950s he came up with a ranking system that
foreclosed that possibility and struck a better balance between
whom you played and how you did, always in light of expectations.
Elo's system predicted almost exactly the number of games by
which Bobby Fischer beat Boris Spassky in their 1972 face-off.

Click on the Men's Ratings link at usatoday.com/sports/sagarin,
and you'll find three columns of information. One is Sagarin's
adaptation of the Elo chess-rating system. Another is Sagarin's
Predictor, a ranking based solely on point differential. (He also
calls this White Owl or Rheingold because it's the column the
cigar-chewing, beer-quaffing sharpies check out: Subtract the
lesser team's figure from the better team's, adjust several
points for the home court, and you have an instant line.) The
third is the set of numbers that receives the most attention, his
composite ratings, which weigh won-lost record and point
differential equally. (The ratings are also published each
Thursday in USA Today.)

Sagarin understands why a system based on points would give the
gambling-averse NCAA the heebie-jeebies. That's why he proposes
that his adaptation of Elo stand in for the RPI. Elo doesn't take
into account margin of victory because, in chess, there is little
significance to whether you win quickly or not; sometimes a
player will sacrifice a queen to the larger objective of winning
a game. Elo also recognizes that the higher-ranked team should
win, so its ranking ought to rise less when it does. Conversely,
the lower-ranked team should lose, so its ranking ought to fall
less when it does. These brakes above and below give Sagarin's
Elo a certain stability and guard against the stratification
between the power leagues and everyone else that mars the RPI as
the season progresses.

Most of all, Elo improves on the RPI by giving a win on the road
more value than a victory at home--because home teams win 67% of
the time. Over the 28 seasons Sagarin has analyzed the NCAA men's
basketball rankings, the RPI has outperformed Elo only six times
in forecasting how teams would do in postseason play. (Sagarin's
composite fared even better, losing to the RPI four times; White
Owl did best of all, losing just twice.)

"I respect what Jeff's doing," says Jim Sukup of The RPI Report.
So apparently does the basketball committee, which since 1984 has
requested his ratings for use during its deliberations. Sagarin
says a member once told him that the committee likes his
composite numbers precisely because they account for
scores--"though," the committee member added, "we can't say
that."

"If I were czar of the NCAA, I'd use Elo," Sagarin says. "It's
not like I'm suggesting affirmative action for the small
conferences. But the way the RPI is set up, it's like telling a
kid in a ghetto high school, 'No matter what you get on your
SATs, you can't go to college.'" --A.W.

WHAT'S BEHIND THE NUMBERS
A closer look at the RPI's Top 100 as of Monday explains some of
its strange ups and downs
STRENGTH OF SCHEDULE
Div. I Rating Rank RPI
Record

1 OKLAHOMA 19-4 .6322 4 .6636
2 TEXAS 18-5 .6434 3 .6614
3 KENTUCKY 22-3 .5821 36 .6595
4 ARIZONA 21-2 .5674 50 .6559
5 NOTRE DAME 21-5 .6248 7 .6556
6 GEORGIA [1] 16-7 .6606 1 .6528
7 DUKE 19-4 .5985 20 .6455
8 FLORIDA 22-4 .5764 44 .6439
9 OKLAHOMA STATE 19-5 .6059 15 .6430
10 UTAH 19-4 .5975 21 .6413
11 LOUISVILLE 19-4 .5790 41 .6413
12 WAKE FOREST 19-4 .5800 40 .6348
13 MARQUETTE 20-4 .5679 48 .6336
14 SYRACUSE 19-4 .5821 37 .6336
15 KANSAS 18-6 .6116 12 .6335
16 XAVIER 20-4 .5674 51 .6329
17 BRIGHAM YOUNG 18-6 .6115 13 .6272
18 DAYTON 19-5 .5774 43 .6244
19 STANFORD 21-6 .5757 45 .6238
20 MISSISSIPPI STATE 17-6 .5933 25 .6208
21 PITTSBURGH 19-4 .5567 61 .6178
22 CINCINNATI 15-8 .6245 8 .6162
23 ALABAMA 15-8 .6106 14 .6134
24 MISSOURI 16-7 .6026 18 .6129
25 PURDUE 16-8 .6119 11 .6118
26 MARYLAND 17-7 .5902 29 .6115
27 ILLINOIS 18-5 .5405 78 .6058
28 ST. JOSEPH'S 19-4 .5238 104 .6025
29 ARIZONA STATE 16-8 .5899 30 .6024
30 MEMPHIS 18-5 .5357 86 .6008
31 WISCONSIN 19-6 .5508 68 .6001
32 CALIFORNIA 19-5 .5191 113 .5970
33 AUBURN 17-7 .5650 53 .5945
34 SETON HALL [2] 13-9 .6190 9 .5932
35 CREIGHTON [3] 24-3 .4805 198 .5929
36 INDIANA 16-9 .5774 42 .5927
37 MICHIGAN STATE 14-11 .6136 10 .5867
38 CONNECTICUT 17-6 .5312 93 .5859
39 TEXAS TECH 15-8 .5648 54 .5850
40 BUTLER 20-4 .4939 165 .5839
41 UNLV 16-8 .5607 58 .5834
42 DEPAUL 14-9 .5808 38 .5827
43 COLORADO 16-9 .5675 49 .5821
44 MINNESOTA 16-7 .5382 83 .5791
45 MICHIGAN 16-9 .5594 59 .5784
46 GONZAGA [4] 20-7 .5219 108 .5782
47 VILLANOVA 14-10 .5941 22 .5773
48 BOSTON COLLEGE 14-9 .5864 32 .5770
49 OREGON 18-7 .5263 100 .5770
50 SAINT LOUIS [5] 12-12 .6269 6 .5769
51 TENNESSEE 15-8 .5446 73 .5768
52 WYOMING 18-6 .5146 121 .5766
53 WEBER STATE 21-4 .4903 179 .5764
54 NORTH CAROLINA [6] 14-12 .6033 17 .5753
55 S. ILLINOIS 19-5 .4948 163 .5734
56 OHIO STATE 13-11 .5996 19 .5724
57 C. MICHIGAN 17-5 .4977 159 .5686
58 VIRGINIA 13-11 .5912 27 .5684
59 LOUISIANA STATE 15-9 .5466 71 .5663
60 PROVIDENCE 13-11 .5922 26 .5644
61 LA.-LAFAYETTE 18-7 .5163 119 .5642
62 TEXAS A&M 13-10 .5840 33 .5640
63 GEORGIA TECH 12-11 .5882 31 .5627
64 WIS.-MILWAUKEE 20-6 .4944 164 .5616
65 CLEMSON 15-8 .5291 95 .5616
66 PENNSYLVANIA 17-5 .4790 201 .5603
67 VANDERBILT 10-13 .6282 5 .5601
68 SOUTH CAROLINA 11-12 .6058 16 .5592
69 KENT STATE 17-6 .4919 172 .5581
70 N.C. STATE 14-9 .5429 75 .5580
71 FRESNO STATE 19-6 .4847 183 .5570
72 MANHATTAN 20-5 .4615 232 .5549
73 UNC WILMINGTON 18-6 .4850 182 .5533
74 RICHMOND 13-11 .5702 47 .5523
75 W. KENTUCKY 18-8 .5058 148 .5517
76 VALPARAISO [7] 17-8 .5137 123 .5515
77 ILLINOIS CHICAGO 18-7 .4933 169 .5514
78 HAWAII 15-8 .5200 111 .5485
79 MISSISSIPPI 12-11 .5653 52 .5481
80 FLORIDA STATE 12-12 .5802 39 .5480
81 HOLY CROSS 20-4 .4426 269 .5478
82 CHARLESTON 19-6 .4704 217 .5475
83 W. MICHIGAN 15-8 .5132 125 .5467
84 WEST VIRGINIA 13-11 .5524 67 .5459
85 SAN DIEGO STATE 12-10 .5587 60 .5456
86 UNC CHARLOTTE 11-13 .5911 28 .5454
87 TROY STATE 20-5 .4587 238 .5446
88 BAYLOR 13-10 .5325 91 .5404
89 WAGNER 17-8 .5253 101 .5401
90 RHODE ISLAND 15-9 .5130 126 .5399
91 TULSA 15-9 .5126 129 .5389
92 IOWA STATE 12-10 .5392 81 .5385
93 ARKANSAS 7-16 .6551 2 .5384
94 FAIRFIELD 15-10 .5273 99 .5349
95 MIAMI (OHIO) 11-11 .5542 63 .5343
96 UC IRVINE 16-7 .4802 199 .5342
97 CHATTANOOGA 16-7 .4816 195 .5335
98 UTAH STATE 17-7 .4736 214 .5331
99 AUSTIN PEAY 15-7 .4754 211 .5327
100 DREXEL 15-9 .5183 114 .5325

[1] THE BULLDOGS topped the RPI for three weeks early in the season,
and they continue to ride high despite seven losses. That's
because Steve Thomas (55), Ezra Williams & Co. have played the
nation's toughest schedule, facing only three teams outside the
RPI's top 100.

[2] Life is good when you're in a major conference. The Pirates, 5-4
as they began Big East play, have gone 8-5 since--and boosted
their RPI by a whopping .0535. (To see how the other half lives,
see No. 76 Valparaiso.)

[3] BY A QUIRK of scheduling, the Bluejays' final three conference
games are against upper-division Missouri Valley teams. If Kyle
Korver and his mates can win them, they should move up in the
RPI. Still, this is a team that the polls and other rating
systems rank much higher.

[4] A year ago, after tying for the West Coast Conference
regular-season crown with Pepperdine, the Zags drew the
seventh-place team in the first round of the WCC tournament--a
stinker RPI game. This year the league has given the top two
regular-season finishers (Gonzaga figures to be one of them) byes
into the tournament semifinals so they can avoid taking a
late-season hit in their RPIs.

[5] The Billikens at week's end had the best RPI of any team without
a winning record. Why? They'd beaten Louisville and
Cincinnati--wins that helped boost them from 85th in the space of
a week.

[6] The Tar Heels' rating dropped .0125 after their three-point loss
to Clemson on Feb. 15. But because the RPI doesn't factor in
point differential, they fell a mere .0036 after a 40-point
walloping by Maryland last Saturday.

[7] By virtue OF their lowly Mid-Continent pedigree, the Crusaders
have added only .0091 to their RPI since their 5-7 start,
despite going 12-1. In other words, the RPI values Seton Hall
going 6-5 in its league almost six times as much as it values
Valpo going 11-1 in its league.

Copyright 2003 Collegiate Basketball News Co.

SOMETIMES, YOU CAN LOSE FOR WINNING
As the cases of Alabama and Butler prove, it's not just how you
play, it's whom you play

To illustrate how the RPI favors schools from major conferences
and is stacked against members of low-profile leagues, take a
look at Alabama and Butler, which began January with nearly
identical records (Alabama was 9-1, Butler 10-1) and with RPI
ratings of .6225 (ranked No. 25) and .5922 (No. 49),
respectively. Since then, as of Monday the Crimson Tide had gone
6-7, but its RPI had suffered only a gentle net decline, to
.6134 (No. 23). The Bulldogs, on the other hand, had gone 11-3
during that span, yet they'd lost ground in the RPI, to .5839
(though their ranking had improved, to No. 40). This despite a
brief jump after an 80-60 loss at Duke on Jan. 30, pointing out
the importance of a team's nonconference schedule.

The Bulldogs are members of the Horizon League, which was 14th
in the RPI conference ratings through Sunday (to No. 1 for
Alabama's SEC), and their RPI will probably continue to sink,
mostly because of their conference's relative weakness. Worse,
they can't do much more to improve their nonleague schedule.
"Butler's not going to get a power team to come to Indianapolis
to play them, because there's a real probability of losing,"
says Jim Sukup, who produces The RPI Report.

Last year the Bulldogs had nonconference wins over Indiana,
Purdue, Washington and usually strong Ball State, finished first
in the Horizon and completed the regular season 25-5. After
being upset in the first round of the Horizon tournament,
however, Butler sat at No. 77 in the RPI and was shut out of the
NCAA tournament. If the Bulldogs want to make sure the same
scenario doesn't play out this March, they had better win the
conference tournament--or arrange to lose to Duke every week.
--A.W.

DIVISION I WON-LOST PCT.

12/31 1/7 1/14 1/21 1/28 2/4 2/11 2/18 2/23

Alabama 9-1 10-1 11-2 12-4 12-5 13-5 13-7 14-8 15-8
Butler 10-1 11-1 13-1 15-1 16-2 17-3 18-3 19-4 21-4

NUMERICAL RPI

12/31 1/7 1/14 1/21 1/28 2/4 2/11 2/18 2/23

Alabama .6225 .6563 .6414 .6322 .6303 .6302 .6078 .6068 .6134
Butler .5922 .5910 .5888 .5884 .5814 .5791 .5875 .5859 .5839

THE RATINGS GAME
Here's how this week's AP Top 10 stacked up on Monday against
the RPI and two Sagarin rankings: Elo Chess, which takes into
account home court advantage, and the Predictor, which uses
margin of victory.
SAGARIN SAGARIN
AP TOP 10 RPI ELO CHESS PREDICTOR

Arizona Oklahoma Kentucky Arizona
Kentucky Texas Arizona Kansas
Oklahoma Kentucky Florida Louisville
Florida Arizona Wake Forest Kentucky
Texas Notre Dame Notre Dame Pittsburgh
Duke Georgia Louisville Texas
Kansas Duke Marquette Maryland
Pittsburgh Florida Texas Oklahoma
Notre Dame Oklahoma State Duke Notre Dame
Marquette Utah Oklahoma Florida

"The major conferences have a scam going," says Sagarin. "When
you have HUGE IMBALANCES between home and away games, it's
irrational to ignore them."