WINTER GAVE US a bad end to the college football season, with BCS bowl games decided by an average margin of more than two TDs and an embarrassing national title game. And now I'm here to warn you that college basketball is heading for the same defective conclusion this spring. Like their counterparts on the gridiron, NCAA hoops teams need to impress a series of so-called experts -- writers, coaches and ultimately an evaluation committee -- to make the postseason and earn an appropriate seeding. But time and again, voters almost comically misjudge certain kinds of programs.
To evaluate teams objectively and to show how bizarre poll results can be, we'll use ESPN's College Basketball Power Index (BPI), a rating that incorporates all D1 games, adjusting a team's score for, among other things, location, result and opponent strength. As of Jan. 28, eight teams were either in the AP's Top 25 but out of the BPI's, or vice versa. Butler, the ninth-best team in the nation, according to the AP, ranked 33rd in the BPI; Kansas State, the AP No. 18, was just 45th in the BPI. Conversely, Pittsburgh was 14th in the BPI but got just 10 votes in the AP poll, landing below 31 other clubs, and Oklahoma State (BPI: 24) didn't earn any votes at all.
These distortions matter. The five most underrated programs, as measured by the gap between their BPI and AP rankings as of Jan. 28 (Pittsburgh, VCU, Minnesota, Colorado State and Creighton), were actually 11.1 points per game better than the five most overrated (Louisiana Tech, Villanova, Arizona State, Kansas State and Butler), according to the BPI. But the overrated teams had an average ranking of 24.8 in the AP poll, while the underrated teams, if they showed up at all, often fell in the category of "others receiving votes." So this is a case of decent teams climbing over much better competitors, knocking them out of the AP Top 25, and therefore the public eye, every week the new poll comes out.
Why does this happen? I see three reasons. First, voters tend to overrate teams from power conferences. Along with Villanova and K-State, overrated clubs include Georgetown, Missouri and Michigan State. Voters also like gaudy records and discount the value of close losses -- even against tough opponents -- because schedule strength is hard to quantify. Michigan (BPI: 4) climbed to No. 1 in the Jan. 28 AP poll because the Wolverines had a 19-1 record, better than any other D1 team, even though their strength of schedule ranked 98th in the NCAA. Meanwhile, advanced metrics said Minnesota (BPI: 10) was one of the 10 best teams in the country, but who believed in the Gophers? After all, they'd lost their past four games -- never mind that three of them were to outstanding teamsand by a total of 16 points.
The last factor: pace of play. Overrated clubs move at a medium tempo. Michigan State (AP: 13; BPI: 20) was averaging 65.7 possessions per game, and Missouri (AP: 17; BPI: 30) was averaging 68.4 -- that's the range of pace for most teams. By contrast, many underrated teams are more extreme. At 60.7 possessions per game, Pitt was one of the slowest in the country, and Florida (AP: 4; BPI: 1) also was playing at a crawl. Then there were UNLV (AP: unranked; BPI: 18), running 70.3 possessions per game, and Kentucky (AP: unranked; BPI: 36), pushing 70. Apparently, it's hard for voters to evaluate teams if they're slow and easy to dismiss them if they're fast.
There's a common thread running through these explanations: Voters favor the familiar, whether it's schools or athletes or styles of play. As a result, Kansas State, an experienced Big 12 team that had gone 15-4 against opponents like North Dakota and Texas Christian, ranked 18th in the AP poll even though, objectively, the Wildcats were barely a top-50 team (BPI: 45, one spot behind Middle Tennessee). Meanwhile, Virginia Commonwealth, a speedy A-10 squad with three narrow losses to outstanding opponents, was one of the 25 best in the NCAA (BPI: 22) but only briefly cracked the AP poll this season. Hoops metricians need to start explaining why VCU deserves a high seed at the Big Dance and K-State shouldn't get in at all. And voters -- and fans -- need to pay attention to the numbers beyond wins and losses. Otherwise, March Madness could look uncomfortably like the BCS national championship.
Follow The Mag on Twitter (@ESPNmag) and like us on Facebook.