Selection Sunday is still a couple days away, but that doesn't mean you can't start narrowing in on your choice for the tourney's ultimate champ.
By monitoring just a few key statistical factors throughout the course of the regular season, you can reduce your list of champion candidates to four teams -- and be reasonably sure that the ultimate victor will be among them. Just look at last year: On my blog, I pegged Kansas as one of four potential champions -- three months before March Madness.
|
|
| Carmelo Anthony's Syracuse team was one of the few champs not to meet all the criteria. (Getty Images) |
The numbers show that fewer than 10 teams on average will possess these three attributes going into the tourney. In 24 seasons, only 228 teams have met the criteria; 68 of them have reached the Final Four (2.83 per dance), and 22 have cut down the nets.
Overall, this group of teams has performed solidly above seed expectations (PASE), posting an impressive +.175 PASE. Last year, only five teams met these criteria. Three of them reached the Final Four (Memphis, North Carolina and eventual champ Kansas). Tennessee and Duke were the other two.
By applying one additional filter, you can winnow your list of potential champions down to eight teams -- with a minimal decrease in predictive accuracy. Twenty one of the past 24 tourney champs (and 19 of the past 20) have met the seed, scoring and margin conditions and been to the previous year's tourney. The only champ that had the scoring and margin credentials but didn't go to the dance the year before was Syracuse in 2003.
If you believe this year's tournament winner won't buck this trend, then you'll only need to concern yourself with about eight candidates.
Since 1985, 195 teams have satisfied the four conditions described above (8.1 per tourney), 63 of them (2.6 per tourney) have reached the Final Four, and 21 have been crowned champion. Not surprisingly, this group's cumulative PASE value is even stronger than the previous group. They beat seed expectations at a rate of +.203 games per tourney. In 2008, the same five teams mentioned above fulfilled these conditions.
Want to cut the list to a handful of squads -- and still have a 70-percent chance of pegging the tourney champ? Stick with the seed and experience conditions; bump the points-per-game and margin requirements up to 77 and 11 points; and add coaching and conference affiliation rules. Seventeen of the 24 tourney champions (and 16 of the past 18) have met all the conditions described above plus they've come from one of the Big Six conferences (ACC, Big 12, Big East, Big Ten, SEC and Pac-10) and had a coach who had been to the dance at least five times.
Most of the exceptions to these conditions occurred in the modern tournament's early years. Low seeds Villanova and Kansas won in 1985 and 1988 and mid-majors Louisville and UNLV won in 1986 and 1990. The other exceptions were Michigan in 1989 (Steve Fisher was making his first trip to the dance), Michigan State in 2000 (Tom Izzo was making his third), and Syracuse in 2003 (the Orangemen hadn't made the previous year's tournament).
If you filtered out all the teams since 1985 that didn't meet these criteria, you'd be left with 113 squads (4.7 per tourney). More than 35 percent of them (40) reached the Final Four -- and 17 became champs. Overall, their PASE (+.261) was over a quarter-game per tourney above seed expectations.
So with six statistical filters, you can restrict your search for a tourney champ to less than five teams -- and be more than 70 percent sure the ultimate winner will be among them.
Let's go one better: Add pre-dance momentum and frontcourt strength to the mix. Sixteen of 24 champs -- and 15 of the past 18 -- have possessed all the previously mentioned conditions, won at least seven of their last 10 pre-tourney games without losing two or more in a row and gotten more than 35 percent of their points from forwards and centers. (The only new exception was Arizona in 1997; the Wildcats had lost two consecutive before the tourney.)
Only 104 squads have met all these criteria in 24 years (4.3 per tourney); 36 of them have been Final Four contenders and 16 have cut down the nets. Perhaps more impressive, these teams are consistently strong tournament overachievers, beating seed expectations at a rate of about one-third of a game per tourney (+.331 PASE). Last year, the only three teams to possess all these credentials were Tennessee, North Carolina and eventual champ Kansas. Two out of three ain't bad.
If all of this sounds like it will be too much late-night, last-minute number crunching, you can do one of two things. Let me do the heavy lifting and just visit my blog, where I'll be evaluating teams on these stats until the tourney starts.
Or keep your eyes on 26 teams that fulfill a more basic set of rules. Eighteen of 24 tourney champs -- and 15 out of 17 -- were seeded No. 1 through No. 4, went to the dance the previous year, had a coach with more than four tourney trips and were among the "Big Six" power conferences.
Before a single regular-season game was played, you could be reasonably sure that the tourney champ will be one of these 26 teams: Arizona, Clemson, Connecticut, Duke, Georgetown, Georgia, Kansas, Kentucky, Louisville, Marquette, Michigan State, Mississippi State, North Carolina, Notre Dame, Oregon, Pittsburgh, Purdue, Stanford, Tennessee, Texas, UCLA, USC, Vanderbilt, Villanova, West Virginia and Wisconsin.
Of course, as the season progressed, some of these teams played their way out of tourney contention, others showed themselves to be unworthy of a top four seed and many more failed to meet one of the other key champion conditions.
So this week, you should be looking at a list of about six teams that have the statistical chops to win the tourney. And there will be a three-in-four chance the ultimate champ will be among them.
Peter Tiernan has been using stats to analyze March Madness for 19 years. His insights into the NCAA tournament can help you build a better bracket. E-mail him at pete@bracketscience.com or visit Bracketscience.com
