Bracket Science: Your guide to every seed matchup

EDITOR'S NOTE: This is the HTML version of a rich-media feature. Find the original Flash feature here .

ROUND 1

1 vs. 16 | 100-0, 1.000 | Higher seed scoring margin: +25.4 ppg
For 25 years and 100 games, top seeds have maintained their perfect record of dominance against 16 seeds. Here's an idea: Let's put a moratorium on talk of expanding the number of teams in the tourney until a 16 finally springs the ultimate upset. Then we'll know that there's enough parity in college basketball to warrant a bigger dance.

Near misses: Purdue 73, West Carolina 71 (1996). Michigan State 75, Murray State 71 OT (1990). Oklahoma 72, East Tennessee State 71 (1989). Georgetown 50, Princeton 49 (1989 ... and Tigers partisans still insist Mourning's last-second block was a foul).

2 vs. 15 | 96-4, .960 | Higher seed scoring margin: +16.8 ppg
Once every six years or so, a 15 seed shocks a two seed. The last victim was Iowa State ... nine tourneys ago. Winthrop darn near bumped off Tennessee in 2006, losing by just two at the buzzer. None of the games in 2009 was particularly close; Memphis struggled against Cal State Northridge but ultimately prevailed 81-70. The closest 2008 2 vs. 15 game saw Duke barely get by Belmont 71-70. That close call was a harbinger of trouble to come; the Blue Devils got upset by seventh seed West Virginia in their next game. While the time might be ripe for another upset, I wouldn't pencil it into your bracket.

Upset watch: All the 15 seed Cinderellas came into the tourney winning nine of their past 10 games and at least three in a row. They all had regular-season records between .600 and .800, indicating that they played their share of tough teams. They scored more than 69 and had a scoring margin of three or more per game. They all had at least one senior starter and got more than 70 percent of their points from the starting unit. And they all got balanced scoring from the back- and frontcourt, averaging between 37 and 58 percent of their scoring from guards. 15 seeds satisfying all these attributes are 4-7 against their two seeded opponents. The rest of the 15 seeds are 0-89.

Upset history: Hampton over Iowa State, 2001. Coppin State over South Carolina, 1997. Santa Clara over Arizona, 1993. Richmond over Syracuse, 1991.

3 vs. 14 | 85-15, .850 | Higher seed scoring margin: +11.4 ppg
Two out of every three tourneys are bad news for three seeds. The 2006 Iowa Hawkeyes and 2005 Kansas Jayhawks were most recently victimized. Three seeds are nearly four times more prone to first-round upsets than two seeds. More amazing, they're less likely to win one game in the tourney than top seeds are to win two. (Only 12 top seeds have failed to reach the Sweet 16.) But it still doesn't make any sense to pick three seeds to lose in Round 1. Too many have made deep runs to eliminate them early.

Upset watch: The 14 seeds most likely to spring upsets are high scoring (averaging more than 76.5 points a game), coming into the tourney with solid momentum (more than six wins in their past 10 games and at least three consecutive victories). They're 13-23 (.361) while their lower scoring counterparts are just 2-62 (.031). If you further refined your criteria to teams relying on unbalanced scoring (more than 55 percent of scoring from either the frontcourt or backcourt), you'd improve your upset-picking odds to 42 percent (11-15). The tell-tale sign of a three seed victim is a tight margin of victory and a so-so record. Three seeds that score less than 17 percent more than their opponents, have a winning percentage lower than .840 and have won fewer than nine of their past 10 games are more prone to upsets (12 losses in 44 tries for a 27.2 percent upset rate) than all other three seeds (only three losses in 56 tries, a 5.4 percent upset rate).

Recent upsets: Northwestern State over Iowa, 2006. Bucknell over Kansas, 2005. Weber State over North Carolina, 1999. Richmond over South Carolina, 1998

4 vs. 13 | 79-21, .790 | Higher seed scoring margin: +9.3 ppg
Four seeds perform almost as solidly as three seeds -- and far better than five seeds. With nearly 80 percent of four seeds advancing to Round 2 -- and less than one per tourney getting upset -- it's too risky to pick a 13 seed in Round 1. That said, last year, Cleveland State knocked off Wake Forest. And two Cinderellas surprised four seeds in 2008: Siena defeated Vanderbilt and San Diego shocked UConn.

Upset watch: The key indicators of a 13 seed Cinderella are balance, momentum and scoring margin. Thirteen seeds that 1) get between 32 and 67 percent of their points from guards, 2) rely on starters for less than 88 percent of their points, 3) have an average scoring margin above 4.7 points and 4) win more than six of their past 10 pre-tourney games are 10-10. All other 13 seeds are 11-69 (.138). The four seeds most likely to be victimized get fewer than half of their points from guards and have a winning percentage worse than .840. They get upset 34.8 percent of the time (16 of 46 games). Four seeds with better backcourts and records have only been upset five times in 54 matchups -- a 9.2 percent upset rate. Last year, Cleveland State had the qualities of a 13-seed dark horse, while Wake Forest, Xavier and Washington all had the vulnerabilities of fourth-seeded victims. Guess what? Cleveland State knocked off Wake Forest in the biggest upset of 2009.

Recent upsets: Cleveland State over Wake Forest, 2009. Siena over Vanderbilt, 2008. San Diego over Connecticut, 2008. Bradley over Kansas, 2006. Vermont over Syracuse, 2005.

5 vs. 12 | 66-34, .660 | Higher seed scoring margin: +4.8 ppg
The 5 vs. 12 matchup marks the point in Round 1 where it no longer pays to give higher seeds a free pass in your bracket. Over the past nine years, five seeds are just 20-16 against their lower-seeded opponent. Last year demonstrated the danger of 12 seeds. Three of them sprung upsets: Arizona beat Utah, Wisconsin knocked off Florida State and Western Kentucky surprised Illinois -- the Hilltoppers' second consecutive 12 seed surprise.

Upset watch: The two factors that matter most in identifying 12 seed spoilers are team experience and frontcourt scoring. Twelfth-seeded teams that have been to the tourney the previous year are 18-14 (.562) against five seeds; all others are 16-52 (.276). More significant, 12 seeds that get 55 percent to 75 percent of their scoring from forwards and centers are 20-15 (.571); the rest are 14-52 (.215). To pick a five seed victim, look at backcourt scoring and a lack of momentum. Fifth-seeded squads that get between 25 and 50 percent of their points from guards and are coming into the tourney with fewer than three consecutive wins are just 21-20 (.512); the rest are 45-14 (.763). In 2009, Arizona, Wisconsin and Western Kentucky had all been to the previous dance -- and they all won. (Only Arizona had the requisite frontcourt scoring.) Meanwhile, only Illinois was both backcourt-challenged and stumbling into the tourney; the Illini lost.

Recent upsets: Arizona over Utah, 2009. Wisconsin over Florida State, 2009. Western Kentucky over Illinois, 2009. Villanova over Clemson, 2008. Western Kentucky over Drake, 2008.

6 vs. 11 | 69-31, .690 | Higher seed scoring margin: +4.3 ppg
Six seeds are more likely to advance to Round 2 than five seeds, but that doesn't mean you should automatically ink them into your bracket. Sure, six seeds are 20-8 over the past seven years -- and they're notorious three seed killers in Round 2. Still, tourney pool success usually comes from accurately identifying the 11 and 12 seed surprises. Last year, the only sixth-seeded victim was Bob Huggins' West Virginia Mountaineers, which lost to upstart Dayton.

Upset watch: Offensive punch, scoring margin percentage and winning rate are the three keys to success for 11 seeds. Teams that score more than 73 points a game and 8 percent more points than they allow with a record better than .640 are 24-22 (.522). All other 11 seeds are nearly four times worse at 7-47 (.130). Sixth-seeded upset victims tend to be high scoring and riding false momentum. Six seeds that average more than 74.7 points a game and have won at least seven of their past 10 games are just 13-15 (.464) in Round 1; the rest are 56-16 (.778). Last year, no 11 seeds had the Cinderella stats to spring an upset and no six seed showed victim vulnerabilities. Still, West Virginia did bow out to Dayton.

Recent upsets: Dayton over West Virginia, 2009. Kansas State over USC, 2008. Virginia Commonwealth over Duke, 2007. Winthrop over Notre Dame, 2007. George Mason over Michigan State, 2006. Wisconsin-Milwaukee over Oklahoma, 2006.

7 vs. 10 | 61-39, .625 | Higher seed scoring margin: +2.7 ppg
As close as these seeds are, it's surprising that seven seeds have been so dominant in this matchup. They win nearly over 60 percent of the time, but they did take it on the chin in 2009, losing three of four matchups against their 10th-seeded opponents. Worse than that, last year marked the first tourney since 2003 that a seven seed didn't surprise a two seed in the second round. More on that later.

Upset watch: Tenth-seeded teams that: 1) get over 3 percent more points than their opponents, 2) get at least 30 percent of their points from guards, 3) have gone to the dance fewer than three years in a row and 4) are led by coaches who've made fewer than six tourney trips are 23-20 (.535); all other 10 seeds are 16-41 (.281). The most victimized seven seeds lack offensive punch and backcourt scoring. Squads that score no more than 76 points a game and get less than 56 percent of their points from guards are just 13-20 (.394); the rest of the seven seeds are 48-19 (.766). Last year, every matchup contained either a 10 seed Cinderella or a seven seed victim. Michigan had the numbers to knock off Clemson -- and did. Boston College, California and Texas all fit the seven-seed victim mold, but the Longhorns managed to avoid a mini-upset.

Recent upsets: Michigan over Clemson, 2009. USC over Boston College, 2009. Maryland over California, 2009. Davidson over Gonzaga, 2008. Alabama over Marquette, 2006. North Carolina State over California, 2006.

8 vs. 9 | 46-54, .460 | Higher seed scoring margin: +0.1 ppg
The 8v9 matchup is the closest thing to a pick-'em contest in the opening round. While eight seeds hold a razor-thin scoring margin of 0.1 points per game, it's the nine seeds that hold the upper hand in the win/loss column. It's somewhat of a surprise that this isn't more of a 50/50 matchup -- and that the lower seed prevails more often. But don't be tempted by the record into giving nine seeds too much credit; they have a horrendous 3-51 record against top seeds in Round 2. Of course, nobody's going to predict either of these seeds to knock off a top seed ... so the value of correctly predicting this matchup is usually restricted to four points in the first round.

Toss-up tips: The key performance indicator for this matchup is team experience and scoring margin. Eight seeds from power or mid-major conferences that have been to the tourney the previous year with a coach who has gone to the dance and a margin percentage of over 7 percent are 23-12 (.657); all other eight seeds are 23-42 (.354). On the other hand, nine seeds with fewer than three consecutive tourney trips that beat their opponents by an average of fewer than six points are 4-13 (.235); the rest are 50-33 (.602). Last year, no nine seeds met the criteria, and BYU was the only eight seed to fit the characteristics of a winner. Alas, the Cougars bowed out to Texas A&M.

Last year's matchups: Siena (9) over Ohio State (8). Texas A&M (9) over BYU (8). Oklahoma State (8) over Tennessee (9). LSU (8) over Butler (9).

ROUND TWO

1|8|9|16 Bracket

Round 2 is made up of four gateway matchups to the Sweet 16. The 1|8|9|16 bracket is easily the most predictable. Top seeds advance an astounding 88 percent of the time. Eight and nine seeds pull off upsets nearly every other tourney. They did it twice in 2004 (thanks to eight seed Alabama and their in-state brethren, nine seed UAB) -- and haven't done it since. While this might suggest they're due, it's really not worth picking them in your bracket. If you're feeling rebellious, though, you're much better off picking an eight- than a nine-seed upset.

1 vs. 8 | 37-9, .804 | Higher seed scoring margin: +9.3 ppg
Unlike nine seeds, eight seeds offer up some resistance against top seeds in Round 2. In eight of the past 25 tourneys, at least one eight seed has made it to the Sweet 16 (two made it in 2000 -- North Carolina and Wisconsin). What are the characteristics of these eighth-seeded giant killers? They're experienced, having gone to the tourney the previous year, and they're battle-tested, with an average victory margin less than six points. Eight seeds satisfying these conditions are 7-10 (.412); all the rest are 2-27 (.069). The most likely one-seed victims are either offensively challenged or inexperienced. Top seeds that either didn't go the previous year's dance or that score fewer than 80 points a game with an average margin of less than 14 points are just 5-6 (.455) while their counterparts are 32-3 (.914).

Recent upsets: Alabama over Stanford, 2004. UCLA over Cincinnati, 2002. North Carolina over Stanford, 2000. Wisconsin over Arizona, 2000.

1 vs. 9 | 51-3, .944 | Higher seed scoring margin: +14.1 ppg
Incredibly, nine seeds are about as likely to upset top seeds in Round 2 as 15 seeds are liable to knock off two seeds in Round 1. It would be a severe bracket brain cramp to advance a nine seed to the Sweet 16. But if you happen to get in a debate over which nine seed is most likely to knock off a top seed, pick a team that didn't go to the tourney the year before, scores more than 73 points and wins by more than seven points a game -- and has won no more than seven of their past ten games. These ninth-seeded squads are 3-5; the rest are a big, fat 0-46. That includes the two nine seeds that lost year -- Siena and Texas A&M.

Upset history: Alabama-Birmingham over Kentucky, 2004. Boston College over North Carolina, 1994. UTEP over Kansas, 1992.

4|5|12|13 Bracket

This is the most difficult of the Sweet 16 gateway brackets to predict. No other bracket has three seeds so evenly matched. Only .045 separates the winning records of the four, five and 12 seeds in Round 2 -- and four seeds actually have a lower winning percentage than five seeds. True, the pressure of getting this mini-bracket right is relieved by the fact that the winner plays the top seed in Round 3 -- and consistently takes it on the chin, losing 83 percent of the time. However, you only have to go back four tourneys to find a fourth-seeded team that reached the Final Four from this bracket -- John Brady's LSU squad.

4 vs. 5 | 27-26, .509 | Higher seed scoring margin: +0.6 ppg
While four seeds have held a slight advantage in this matchup since 1985, five seeds have actually won 11 of the past 14 games, including Purdue's victory over Washington last year. Four seeds with an experienced tourney coach that went to the previous dance, have a margin percent less than 20 percent and get more than 18 percent of their scoring from the bench are 13-5 (.722); all other four seeds are 14-21 (.400). Meanwhile, five seeds with an experienced tourney coach that score more than 77 points a game and haven't lost two or more consecutive games entering the dance are 14-6 (.700); all others are 12-21 (.364). Last year, neither Washington nor Purdue met the four and five seed criteria.

Recent five-seed wins: Purdue (5) over Washington (4), 2009. Michigan State (5) over Pittsburgh, 2008. Butler (5) over Maryland, 2007. USC (5) over Texas (4), 2007. Tennessee (5) over Virginia (4), 2007.

4 vs. 12 | 16-10, .615 | Higher seed scoring margin: +4.0 ppg
This matchup is closer than the disparity in seed positions indicates. The most reliable four seeds score more than 72 points a game, have won at least five of their last 10 pre-tourney games and have coaches with more than three tourney trips. They're 16-5 (.762) while other four seeds are 0-5. The most surprising 12 seeds are either from Big Six or mid-major conferences (no smalls) and have won no more than three games in a row. They're 8-6 (.571); other 12 seeds are 2-8 (.200). Last year, Gonzaga had the signs of a fourth-seeded victor and won. Meanwhile, Wisconsin had the right statistical stuff to spring a 4 vs. 12 upset but lost.

Recent upsets: UW-Milwaukee over Boston College, 2005. Butler over Louisville, 2003. Missouri over Ohio State. Southwest Missouri State over Tennessee, 1999.

5 vs. 13 | 10-3, .769 | Higher seed scoring margin: +7.0 ppg
Unlike four seeds, five seeds have little trouble in their Cinderella mismatch against 13 seeds -- despite the fact that the most recent of these matchups, the 2006 Pittsburgh-Bradley game, went to the underdog. Still, there's little reason to pick against five seeds -- particularly if their backcourt shoulders more than 38 percent of the scoring load. Guard-dominant five seeds are 8-1, while more frontcourt-oriented fifth-seeded teams have struggled (2-2, .500). If you get a wild hair and feel like picking an upset in this matchup, go for 13 seeds that get less than 48 percent of their points from guards, plus score more than 70 and give up less than 66 points a game. These 13th-seeded spoilers are 3-0; all other 13 seeds are 0-10.

Upset history: Bradley over Pittsburgh, 2006, Oklahoma over Charlotte, 1999. Richmond over Georgia Tech, 1988.

12 vs. 13 | 7-1, .875 | Higher seed scoring margin: +7.9 ppg
The long shot seeds in this bracket have squared off against each other more often than any other longshot pairing (9v16, 10v15, 11v14) in the second round. Surprisingly, 12 seeds treat 13 seeds like pushovers, prevailing 88 percent of the time. Last year, twelfth-seeded Arizona cruised past Cleveland State. In 2008, for the first time in the 25-year modern era, there were two 12v13 matchups -- Villanova versus Siena and Western Kentucky versus San Diego. In both cases, the 12 seed prevailed. What did the one 13-seed victor have that the victims didn't? Team experience. Valparaiso had been to the tourney three straight years when it knocked off Florida State in 1998. The seven 13-seed losers had fewer than three appearances -- with only Indiana State having gone to the previous year's tourney.

Recent Matchups: Arizona (12) over Cleveland State (13), 2009. Villanova (12) over Siena (13), 2008. Western Kentucky (12) over San Diego (13), 2008. Gonzaga (12) beat Indiana State (13), 2001. Valparaiso (13) beat Florida State (12), 1998. George Washington (12) over Southern University (13), 1993.

3|6|11|14 Bracket

The 3|6|11|14 bracket is perhaps the most difficult pairing to figure out in the second round. That's because sixth-seeded teams are such surprising performers. While fewer six seeds make it to Round 2 than three seeds (69 to 85), their winning percentage isn't too far behind that of higher-seeded rival in the second round. In fact, they win with nearly the same regularity as four or five seeds. Eleventh-seeded teams aren't slouches either. In fact, odds are that the four teams advancing from this bracket pairing will comprise nearly as many lower seeds as three seeds. One other reason this bracket is so difficult: The stakes for getting it right are higher. More than twice as many teams (41) will advance to the Elite Eight as from the 4|5|12|13 bracket (20 teams). Last year, three seeds Missouri and Villanova both reached the Elite Eight -- with the Wildcats reaching the Final Four.

3 vs. 6 | 32-25, .561 | Higher seed scoring margin: +2.7 ppg
This used to be one of the most hotly contested matchups in the second round -- four tourneys ago. After the 2005 dance, the two seeds had split their 48 game. Since then, three seeds have won eight of nine. Despite this surge, the 3 vs. 6 tilt deserves more scrutiny from bracket pool players. I usually give one and two seeds automatic passes to the Sweet 16. And I'm less concerned about the 4 vs. 5 games since the winners are served up to top seeds in Round 3. That leaves this mini-bracket as the most important one to get right in Round 2.

If you're looking to pick a six seed in this matchup, consider teams with coaches who have: 1) been to the dance between one and five times, 2) won between five and nine of their past 10 games -- but have fewer than seven wins in a row, and 3) rely on guards for less than 65 percent of their points. Six seeds with these three attributes are 16-6 (.727); the rest are 9-26 (.257). Last year, none of the six seeds had these qualities -- and all lost. Meanwhile, less experienced three seeds hold serve better than tourney-grizzled squads. Third-seeded teams who've been to the dance fewer than six times in a row are 27-16 (.628); the rest are 6-11 (.353). Last year, this rule worked perfectly. Villanova, Missouri and Syracuse all met the experience condition -- and all won.

Recent upsets: Vanderbilt over Washington State, 2007. Texas Tech over Gonzaga, 2005. Utah over Oklahoma, 2005. Vanderbilt over North Carolina State, 2004.

3 vs. 11 | 20-8, .714 | Higher seed scoring margin: +8.5 ppg
It isn't exactly a "gimme" when three seeds square off against 11 seeds in Round 2. Just ask North Carolina, which was victimized by George Mason in 2006. The best performance indicators to explain why 11 seeds spring upsets are pre-tourney momentum and offensive output. Eleventh-seeded squads that have won between five and eight of their past 10 contests while averaging 68 points per game are 8-10 in this matchup (.444); all other 11 seeds are perfectly inept at 0-10. The tell-tale sign of faltering three seeds is team experience. Third-seeded squads that either didn't go to the tourney the previous year or are tourney fixtures (more than six consecutive appearances) are 12-0; all others are 8-8. Last year, third-seeded Kansas was a tourney veteran and easily handled 11 seed Dayton, which averaged fewer than 68 points a game.

Recent upsets: George Mason over North Carolina, 2006. Southern Illinois over Georgia, 2002. Temple over Florida, 2001. Loyola-Marymount over Michigan, 1990.

6 vs. 14 | 10-2, .833 | Higher seed scoring margin: +5.1 ppg
If you went out on a bracket limb and advanced a 14 seed into the second round, you'd be smart to eliminate it in Round 2. Then again, logic would've dictated that you never advance a 14 seed in the first place. So if you're still feeling reckless with this matchup, take the 14 seeds that beat their opponents by more than 12 points per game. They're 2-0 against six-seeders; the rest of the 14 seeds are 0-10.

Upset history: Tennessee-Chattanooga over Illinois, 1997. Cleveland State over St. Joseph's, 1986.

11 vs. 14 | 3-0, 1.000 | Higher seed scoring margin: +14.7 ppg
Anyone who's contemplating an 11 vs. 14 matchup in their bracket probably isn't reading this article to begin with -- and doesn't care that 11 seeds have never lost to 14 seeds in the second round. Washington beat Richmond in 1998, Connecticut handled Xavier in 1991 and Minnesota stopped Siena in 1989.

2|7|10|15 Bracket

By all rights, the 2|7|10|15 bracket should be a no-brainer -- and most bracket pool players pick it that way, giving two seeds an automatic pass to the Sweet 16. But second-seeded squads aren't nearly as reliable as top seeds in advancing beyond the second round. On average, 1½ seven or 10 seeds per tourney will take the place of two seeds. Last year marked just the third time in the 25-year, 64-team tourney era that all four two seeds reached the Sweet 16. You can go the safe route, cross your fingers and advance all the two seeds. Or you can be a rebel, observe the tell-tale signs of seven and 10 seed victors, and advance a Cinderella. The bigger the pool you're in, the more likely it is that the winner will go against the grain.

2 vs. 7 | 43-17, .717 | Higher seed scoring margin: +5.7 ppg
Despite being the closer competitor by seed position, seven seeds are surprisingly more prone to getting beaten by second-seeders than 10 seeds are. Last year, there was only one 2 vs. 7 game, and second-seeded Duke edged Texas. Seventh-seeders that offer the stiffest resistance are tourney tested but not fixtures, having been to the dance two to eight consecutive years; they're led by experienced tournament coaches; and they beat their opponents by more than six points a game. These seven seeds are a respectable 9-9 (.500); others are 8-34 (.190). Texas didn't meet this criteria last year, as they'd gone to the dance 11 consecutive years.

Which second-seeders are most likely to tank in this matchup? Steer clear of two seeds whose coach has made no more than one Elite Eight appearance. They're just 17-14 (.548) against seven seeds; teams with more successful tourney coaches are 26-3 (.897). Last year, Duke's Coach K put the Blue Devils firmly in the latter category.

Recent upsets: West Virginia over Duke, 2008. UNLV over Wisconsin, 2007. Wichita State over Tennessee, 2006. Georgetown over Ohio State, 2006.

2 vs. 10 | 21-15, .583 | Higher seed scoring margin: +6.1 ppg
Amazingly, 10 seeds beat two seeds at more than a 40 percent rate in the second round. Of course, the odds of a 10 seed winning its first two games are still just 18 percent, so it's not worth getting too excited about their propensity to topple two seeds. That said, you only need to look to 2008's Cinderella run by Davidson to know that 10 seeds can be dangerous. Then again, there were three 2 vs. 10 games last year (for the second time in 25 years; amazingly, 1999 had four), and the higher seed prevailed in every game.

Which 10 seeds have the best odds of reaching the Sweet 16? Look for those that score more than 72 points and beat opponents by more than five points a game. These teams are 12-10 (.545), while other 10 seeds are just 3-11 (.214). Last year, Memphis beat Maryland, Oklahoma beat Michigan and Michigan State beat USC -- and none of the underdogs fulfilled the upset criteria.

Recent upsets: Davidson over Georgetown, 2008. North Carolina State over Connecticut, 2005. Nevada over Gonzaga, 2004. Auburn over Wake Forest, 2003.

7 vs. 15 | 1-0, 1.000 | Higher seed scoring margin: +11.0 ppg
A seven seed has only played a 15 seed once in Round 2. In 1993, seventh-seeded Temple beat 15 seed Santa Clara.

10 vs. 15 | 3-0. 1.000 | Higher seed scoring margin: +11.0 ppg
Tenth-seeders have pushed their Round 2 record almost to .500 by beating 15 seeds in all three of their matchups. Georgetown beat Hampton in 2001, Texas beat Coppin State in 1997, and Temple handled Richmond in 1991.

SWEET 16

Top Seed Bracket

You wouldn't go too far wrong to advance top seeds to the Elite Eight. More top seeds advance to the quarterfinals than two seeds get to the Sweet 16 -- or five seeds win in Round 1! Heck, more first-seeders get to the Elite Eight (73) than two and three seeds combined (71). More than 70 percent of top seeds -- almost three per tourney -- win their first three games. The only other seeds worth considering in this bracket are four, five and eight seeds. They've graduated 25 teams -- exactly one per tourney -- to the fourth round. The three lower seeds have only advanced two teams.

1 vs. 4 | 26-10, .722 | Higher seed scoring margin: +6.6 ppg
On the top-seed side of the Sweet 16 bracket, the most frequent matchup pits the two highest seeds against each other. One and four seeds are involved in 36 percent of the games, with top seeds winning nearly three out of four. That might sound like a lock, but it's the worst performance by a top seed in any of its matchups over the first three rounds. Four seeds thrive when they beat their foes by more than 10 points a game. They're 7-6 (.538); less dominant four seeds are 3-18 (.143). Last year's only 1 vs. 4 matchup followed this rule -- but just barely: fourth-seeded Washington State beat opponents by an average of exactly 10 points per game -- but not more. And they wound up losing to top-seeded North Carolina.

Recent four seed victories: LSU (4) over Duke (1), 2006. Villanova (1) over Boston College (4), 2006. Louisville (4) over Washington (1), 2005. Ohio State (4) over and Auburn (1), 1999. Arizona (4) over Kansas (1), 1997.

1 vs. 5 | 28-5, .848 | Higher seed scoring margin: +7.8 ppg
Given their proximity, you'd think that five seeds would do nearly as well against top seeds as four seeds. In fact, they're easy marks for the big guns, pulling upsets less often than eight seeds do against top seeds in the second round. What distinguishes the five top seeds that got upset? Interestingly, it's the tourney fixtures that tend to fare worse. Top seeds that have been to the tourney at least five consecutive times are just 14-4 (.778); the less experienced top seeds are 14-1 (.933). Here's another sign of a top-seed victim: their name is "Duke." The last three one seeds to get upset by five seeds were the Blue Devils -- in 2000, 2002 and 2005. On the dark horse side of the matchup, fifth-seeded Cinderellas get more than 22 percent of their scoring from the bench and are from Big Six conferences. These teams are 4-9 (.308); their counterparts are 1-19 (.050). Last year, UConn wasn't saddled with the experience stigma of an upset victim, and Purdue didn't have the bench strength of a Cinderella. The matchup went to form, with the Huskies beating the Boilermakers 72-60.

Upset history: Michigan State over Duke, 2005. Indiana over Duke, 2002. Florida over Duke, 2000. Mississippi State over Connecticut, 1996. Virginia over Oklahoma, 1989.

1 vs. 12 | 16-0, 1.000 | Higher seed scoring margin: +14.6 ppg
Given how dominant top seeds are in the first three rounds, it's not surprising that they're a perfect 16-0 against 12 seeds. It is a little eye-opening, however, that they handle their underdog opponents so easily. Top seeds have beaten 12 seeds by an average of 14.6 points, with only four of the 16 games being settled by single digits. Ball State came the closest to springing an upset in 1990 when the Cardinals lost to UNLV 69-67. 2008 was the only dance in the 25-year modern tourney era that featured two 1 vs. 12 games. Both of them were reasonably competitive games: UCLA beat Western Kentucky 88-78, and Kansas slipped by Villanova 72-57. Last year, Louisville absolutely pulverized 12 seed Arizona, 103-64.

Recent matchups: Louisville over Arizona, 2009. Kansas over Villanova, 2008. UCLA over Western Kentucky, 2008. Illinois over UW-Milwaukee, 2005. Oklahoma over Butler, 2003. Michigan State over Gonzaga, 2001.

1 vs. 13 | 3-0, 1.000 | Higher seed scoring margin: +15.3 ppg
In the three Sweet 16 matchups between these seeds, the top seed has held serve against its long shot opponent. Memphis rolled over surprising Bradley in 2006, Michigan State took care of Oklahoma in 1999, and Temple dispatched Dick Tarrant's pesky Richmond Spiders in 1988.

4 vs. 8 | 2-3, .400 | Higher seed scoring margin: +0.4 ppg
Once every five years or so, a four and eight seed go head-to-head in the Sweet 16. The Cinderella eighth-seeders are tough teams that keep the score low and close, averaging fewer than 80 points a game and winning by no more than eight. Teams with these qualities are 3-0. The other two eight seeds have fallen to their fourth-seeded opponents.

Matchup history: North Carolina (8) over Tennessee (4), 2000. Wisconsin (8) over LSU (4), 2000. Syracuse (4) over Georgia (8), 1996. Arkansas (4) over North Carolina (8), 1990. Auburn (8) over UNLV (4), 1986.

4 vs. 9 | 2-0, .000 | Higher seed scoring margin: +14.0 ppg
The only two times these seeds have met in the Sweet 16, the favored four seeds have prevailed. Bill Self's Kansas Jayhawks beat Mike Anderson's UAB Blazers in 2004. And Bob Huggins Cinci Bearcats beat Clem Haskins' UTEP Miners in 1992.

5 vs. 8 | 0-2, .000 | Higher seed scoring margin: -6.0 pgg
Eight seeds have upset five seeds both times that they've played against each other. Mike Gottfried and Alabama were the most recent school to do the trick, toppling Jim Boeheim's defending champion Orangemen in 2004. And the first upset came in the very first year of the modern 64-team era, when Rollie Massimino masterminded an upset of Lefty Dreisell's Maryland Terps on his way to the 1985 championship.

5 vs. 9 | 0-1, .000 | Higher seed scoring margin: -9.0 ppg
The 1994 Boston College squad, coached by Jim O'Brien, has the distinction of being the only nine seed to reach the Elite Eight. They achieved the feat by knocking off Bobby Knight's fifth-seeded Hoosiers.

8 vs. 12 | 0-1, .000 | Higher seed scoring margin: -9.0 ppg
In 2002, Quin Snyder's Missouri Tigers upset Steve Lavin's UCLA Bruins in the only 8 vs. 12 matchup of the modern tourney era.

8 vs. 13 | 1-0, 1.000 | Higher seed scoring margin: +6.0 ppg
The only matchup pitting these seeds against each other occurred in 1998 when Jim Harrick's Rhode Island Rams beat Cinderella Valparaiso, led by coach Homer Drew and his son, guard Bryce.

Second Seed Bracket

On the two-seed side of the Sweet 16 bracket, the competition is much more balanced than the one-seed side, where top seeds advance 73 percent of the time. While two seeds are the most common winners, claiming 46 percent of the Elite Eight positions, the likelihood is that some other seed will advance. Three and six seeds prevail in 37 percent of matchups. In the other bracket, the four and five seeds only get to the quarterfinals 19 percent of the time. Even the seven, 10 and 11 seeds get into the act, advancing 17 teams -- more than twice as many as the eight, nine and 12 seeds in the other bracket. On the other hand, two seeds are the only seed on this side of the Sweet 16 bracket with a winning record.

2 vs. 3 | 21-12, .636 | Higher seed scoring margin: +2.3 ppg
Of all the matchups with a single seed position difference in the first three rounds (8 vs. 9, 4 vs. 5, 6 vs. 7 and 12 vs. 13), this one is amazingly the second-most lopsided, behind only 12 vs. 13 (7-1). It gets more lopsided if you concentrate only on two seeds that get imbalanced scoring -- more than 60 percent of their points from either the backcourt or frontcourt. These squads are 15-4 (.789); the more balanced scoring two seeds are just 6-8 (.429). If your heart's set on picking a three seed, go with one whose coach isn't a rookie to the tourney but has fewer than 10 trips. These three seeds are 9-8 (.529), while their counterparts are 3-13 (.188). Last year, this three-seed rule perfectly identified the three seeds that won (Villanova and Missouri) while avoiding the threes that lost (Kansas and Syracuse). Two two-seed imbalanced scoring rules caused a split, accurately identifying Michigan State as a victor but missing on the Memphis pick. One interesting point: 2009 marked the only dance of the 25-year, 64-team era in which all four regions had a 2 vs. 3 matchup.

Recent matchups: Michigan State (2) over Kansas (3), 2009. Oklahoma (2) over Syracuse (3), 2009. Villanova (3) over Duke (2), 2009. Missouri (3) over Memphis (2), 2009. Texas (2) over Stanford (3), 2008. Louisville (3) over Tennessee (3), 2008.

2 vs. 6 | 18-5, .783 | Higher seed scoring margin: +5.6 ppg
This matchup happens almost as frequently as 2 vs. 3 -- a testament to the resilience of six seeds in the first two rounds. Unfortunately, that resiliency doesn't seem to help them against two seeds. In fact, 16 of the past 17 2 vs. 6 games have gone to the favored seed, with USC's 2001 win over Kentucky the only upset. The six-seeders that tend to win are inexperienced schools (less than three consecutive tourney trips) with solid frontcourts (more than 45 percent of their scoring from forwards and centers). Teams with these two qualities are 5-7 (.417); the rest of the six seeds are 0-11.

Upset history: USC over Kentucky, 2001. Michigan over Oklahoma State, 1992. Minnesota over Syracuse, 1990. Villanova over Kentucky, 1988. Providence over Alabama, 1987.

2 vs. 11 | 7-1, .875 | Higher seed scoring margin: +6.4 ppg
This matchup comes around about once every three years -- and is nearly always won by the two seed. The only 11 seed victory came way back in 1986, the second year of the modern tourney era, when Dale Brown's LSU Tigers knocked off Georgia Tech. What did LSU have that the other 11 seeds didn't? Tourney experience. They're the only 11 seed that had been to the tourney more than two years in a row. Want a weird little factoid to impress your buddies (and who doesn't)? UConn has been in the past three 2 vs. 11 matchups, twice as a two seed and once as the 11 seed.

Recent matchups: Connecticut (2) over Southern Illinois (11), 2002. Connecticut (2) over Washington (11), 1998. Duke (2) over Connecticut (11), 1991. Duke (2) over Minnesota (11), 1989.

3 vs. 7 | 6-2, .750 | Higher seed scoring margin: +1.4 ppg
Four of the eight 3 vs. 7 matchups happened in the first decade of the tourney. The other four have happened in the past six dances. Last year, third-seeded Xavier edged seven seed West Virginia in overtime; in 2007, three seed UCLA beat UNLV; in 2006, three seed Florida beat Georgetown; and in 2004, Xavier knocked off three seed Texas. What separates the third-seeded winners from losers? Pre-tourney momentum. Three seeds that come to the tourney neither too hot nor too cold -- that is, winning either seven or eight of their past ten games -- are 6-0. The other two third-seeders got upset.

Recent matchups: Xavier (3) over West Virginia (7), 2008. Oregon (3) over UNLV (7), 2007. Florida (3) over Georgetown (7), 2006. Xavier (7) over Texas (3), 2004.

3 vs. 10 | 7-4, .636 | Higher seed scoring margin: +1.4 ppg
Nearly every other tourney pits a three seed against an underdog 10 seed. The best guidance to the outcome of this matchup hinges on the team experience of the Cinderella. The four 10th-seeders that were making a return trip to the dance all won -- that includes Davidson last year; the seven that didn't go to the previous tourney all lost.

Upset history: Davidson over Wisconsin, 2008. Kent State over Pittsburgh, 2002. Temple over Oklahoma State, 1991. LSU over DePaul, 1987.

6 vs. 7 | 3-3, .500 | Higher seed scoring margin: +2.5 ppg
This matchup has been a tale of two eras. Six seeds won the first three games and seven seeds won the next three. One dynamic has remained fairly consistent through all six: The team that allows the fewest points per game has a solid 5-1 edge.

Recent matchups: West Virginia (7) over Texas Tech (6), 2005. Michigan State (7) over Maryland (6), 2003. Tulsa (7) over Miami (Fla.) (6), 2000. Memphis State (6) over Georgia Tech (7), 1992.

6 vs. 10 | 4-2, .667 | Higher seed scoring margin: +6.8 ppg
Since 2000, six seeds have asserted their dominance in what was once an even matchup. The tell-tale sign of a sixth-seeded winner is team experience. Schools that have been to the tourney more than five consecutive years are 4-0; the other two schools are 0-2. The mark of a 10th-seeded victor is scoring punch. Both 10 seeds that scored more than 78 points a game were 2-0, while their more offensively challenged counterparts were winless.

Recent Matchups: Wisconsin (6) over North Carolina State (10), 2005. Purdue (6) over Gonzaga (10), 2000. Temple (6) over Purdue (10), 1999. Gonzaga (10) over Florida (6), 1999.

7 vs. 11 | 0-3, .000 | Higher seed scoring margin: -7.3 ppg
Seven seeds have had a rough time with 11 seeds in the Sweet 16. They've lost all three times the two seeds have gone head to head -- in 1990 when offensive juggernaut Loyola-Marymount upset Alabama, in 2001, when Temple beat Penn State and in 2006, when George Mason upset Wichita State.

7 vs. 14 | 1-0, 1.000 | Higher seed scoring margin: +1.0 ppg
The only 7 vs. 14 matchup of the modern tourney era came just one year after the field expanded to 64 teams. Navy, led by David Robinson, held off Cleveland State in 1986.

10 vs. 14 | 1-0, 1.000 | Higher seed scoring margin: +6.0 ppg
In the only game pitting these two low seeds against each other, Pete Gillen's 10th-seeded Providence Friars avoided an upset at the hands of Tennessee-Chattanooga in 1997.

ELITE EIGHT

A funny thing happens on the way to the Elite Eight. The matchup the brackets were designed to yield -- 1 vs. 2 -- happens only 35 percent of the time. The second most likely matchup, 1 vs. 3, happens in just 18 percent of the regions. Top seeds do their part, appearing in 73 percent of the quarterfinal games. It's the other side of the bracket that's splintered. As for which seed will advance in this round, one seeds get to the Final Four as many times (44) as two, three and four seeds combined. And these top four seeds account for 88 of the 100 Final Four teams. So when you're slotting teams into your semifinal brackets, you wouldn't be too far wrong to pick two top seeds and two of the next three seeds. Of course, the big questions are: Which of those top-seeded teams should you choose? And which two, three and four seeds should you have join them in the Final Four? In 2008, the answers to these questions were simple: Just advance all the top seeds and forget about every other seed. For the first time in 25 years, the "all-top-seed" strategy yielded perfect results. But lightning didn't strike twice in 2009; two top seeds (North Carolina and UConn) were joined by a two seed (Michigan State) and three seed (Villanova). What's the right approach this year? These matchup breakdowns should provide guidance.

1 vs. 2 | 18-17, .514 | Higher seed scoring margin: +1.4 ppg
More one seeds have their tourney run ended by two seeds in the Elite Eight than by any other opponent in any other round. Considering that only 12 top seeds lose in Round 2 and 15 in the Sweet 16, this matchup is somewhat of a Waterloo for top-seeded teams. Still, they do own a winning record against two seeds -- performing the best against inexperienced teams with coaches who lack tourney seasoning. Second-seeders with fewer than four consecutive tourney trips or coaches who've been to the dance fewer than four times are 3-14 (.176); the more tourney-tested two seeds are 14-4 (.778). Be careful not to rely too much on this rule, however. The three second-seeders to buck the trend did so within the past four tourneys. In 2007, both UCLA and Georgetown toppled top seeds, despite having fewer than four consecutive bids. And in 2006, UCLA also accomplished the feat. Last year, however, the two-seed inexperience exclusion rules got both matchups right. Two-seed tourney fixture Michigan State knocked off Louisville, while less experienced Oklahoma fell to the Tar Heels.

Recent matchups: North Carolina (1) over Oklahoma (2), 2009. Michigan State (2) over Louisville (1), 2009. Memphis (1) over Texas (2), 2008. Ohio State (1) over Memphis (2), 2007. Georgetown (2) over North Carolina (1), 2007. UCLA (2) over Kansas (1), 2007. UCLA (2) over Memphis (1), 2006.

1 vs. 3 | 10-8, .556 | Higher seed scoring margin: +1.3 ppg
Here's more proof that there's no difference between the top three seeds in the Elite Eight: like two seeds, three-seeders nearly break even with top seeds. The one seeds that lose in this matchup tend to be sputtering heading into the tourney. Top seeds that have won fewer than nine of their past 10 pre-tournament games are 3-6; those that have notched nine or 10 wins are 7-2. The tell-tale sign of a triumphant three seed is coaching experience. Three seeds with coaches who've been to the dance more than six times are 6-3; those with less experienced coaches are 2-7. Last year, this rule correctly picked three seed Villanova to win and Missouri to lose. On the other hand, the top-seed momentum rule got UConn right and Pitt wrong.

Recent matchups: Connecticut (1) over Missouri (3), 2009. Villanova (3) over Pittsburgh (1), 2009. Florida (1) over Oregon (3), 2007. Florida (3) over Villanova (1), 2006. Illinois (1) over Arizona (3), 2005. Marquette (3) over Kentucky (1), 2003. Syracuse (3) over Oklahoma (1), 2003.

1 vs. 6 | 6-2, .750 | Higher seed scoring margin: +8.5 ppg
When they're not struggling with two and three seeds, top seeds are 15-4 against the rest of the field -- 9-2 if you back out their performance against six seeds. In this matchup, the difference between a top-seeded winner and loser is frontcourt scoring. The six top-seeded victors got at least 50 percent of their points from forwards and centers; the two losers leaned on their backcourt for more than 60 percent of their scoring. Both the sixth-seeded squads that won (Michigan in 1992 and Providence in 1987) were high-scoring teams, averaging at least 78 points a game. The six seed victims all scored fewer than 78 points per game.

Upset history: Michigan over Ohio State, 1992. Providence over Georgetown, 1987.

1 vs. 7 | 4-0, 1.000 | Higher seed scoring margin: +9.0 ppg
Top seeds have no trouble with seven seeds, but this matchup may be a curse for the favorites. None of the top seeds in this showdown have gone on to win the tourney. The seven seed that came closest to springing an upset was Xavier, which lost by just three points to Duke in 2004.

Matchup history: Duke over Xavier, 2004. Texas over Michigan State, 2003. Michigan over Temple, 1993. Duke over Navy, 1986.

1 vs. 10 | 4-0, 1.000 | Higher seed scoring margin: +3.0 ppg
Top seeds have ended the Cinderella stories of four 10 seeds in the Elite Eight. The games have served as a good tuneup for the top seeds; three of the four winners have gone on to win the tourney. Indiana did it first in 1987. Then UConn used a 10 seed as a tuneup for the 1999 championship. And in 2008, Kansas did the same thing with Davidson.

Matchup history: Kansas over Davidson, 2008. Connecticut over Gonzaga, 1999. North Carolina over Temple, 1991. Indiana over Louisiana State, 1987.

1 vs. 11 | 2-2, .500 | Higher seed scoring margin: +8.3 ppg
Top seeds have more trouble in the Elite Eight with 11th-seeded long shots than they do with two seeds. That's because there have been only four matchups -- and the underdog seed has split the series. LSU's victory over Kentucky in 1986 stood as the biggest late-round upset in the 64-team era -- until George Mason duplicated the feat in 2006 by knocking off UConn. What did the two Cinderellas have that the two losing 11 seeds didn't (not that you'd ever pick an 11 seed to advance this far)? Both George Mason and LSU had lost one game before entering the tourney; Temple and Loyola-Marymount had winning streaks coming into the dance.

Matchup history: George Mason (11) over Connecticut (1), 2006. Michigan State (1) over Temple (11), 2001. UNLV (1) over Loyola-Marymount (11), 1990. Louisiana State (11) over Kentucky (1), 1986.

2 vs. 4 | 2-3, .400 | Higher seed scoring margin: -0.6 ppg
Once every four or five years, a two seed plays a four seed in the Elite Eight -- like Texas did in 2006 when it squared off against fourth-seeded LSU. The outcome of the game usually hinges on pre-tourney momentum. The team that's won more of their past 10 games before entering the dance is 4-1, with only fourth-seeded Oklahoma State bucking the trend in 1995 when it beat a hotter Massachusetts squad. If you put more stock in recent tourney results, however, you might as well just pick the four seed in this matchup. They've won the past three games against two seeds in the Elite Eight.

Matchup history: LSU (4) over Texas (2), 2006. Syracuse (4) over Kansas (2), 1996. Oklahoma State (4) over Massachusetts (2), 1995. Arkansas (2) over Virginia (4), 1995. Duke (2) over St. John's (4), 1991.

2 vs. 5 | 0-2, .000 | Higher seed scoring margin: -8.0 ppg
As close as these seeds are, you'd think there would be more than two games over the last 21 years. You'd also think that two seeds would do better. Five seeds have won both matchups, most recently in 2005 when Michigan State upset Kentucky. The other game occurred in 1996 when Mississippi State beat Cincinnati.

2 vs. 8 | 2-1, .667 | Higher seed scoring margin: +4.0 ppg
Over the past 23 years, a two seed has squared off against an eight seed only once, when Connecticut held off Alabama in 2004. The other two 2 vs. 8 matchups happened in the first two years of the modern tourney era. Villanova is the only eighth-seeded squad to come out on top against a two, knocking off North Carolina on its improbable way to the 1985 championship. What set Villanova apart from the other two eight seeds was that their coach, Rollie Massimino, had been to the Elite Eight before.

Matchup history: Connecticut (2) over Alabama (8), 2004. Louisville (2) over Auburn (8), 1986. Villanova (8) over North Carolina (2), 1985.

2 vs. 12 | 1-0, 1.000 | Higher seed scoring margin: +6.0 ppg
In 2002, second-seeded Oklahoma put an end to the longest tourney run by a 12 seed when the Sooners beat Big 12 rival Missouri.

3 vs. 4 | 2-1, .667 | Higher seed scoring margin: +9.3 ppg
On those rare occasions when a three seed goes up against a four seed in the Elite Eight, the older team in terms of class composition has won each time. All three winners -- Georgia Tech, Ohio State, and Seton Hall -- had more juniors and seniors in their starting lineup than their opponents.

Matchup history: Georgia Tech (3) over Kansas (4), 2004. Ohio State (4) over St. John's (3), 1999. Seton Hall (3) over UNLV (4), 1989.

3 vs. 5 | 1-1, .500 | Higher seed scoring margin: +12.5 ppg
Three and five seeds have split their two games against each other in the quarterfinals. In 1989, three seed Michigan beat fifth-seeded Virginia on its way to the championship. In 2000, five seed Florida beat three seed Oklahoma State, advanced to the finals, then lost to Michigan State.

3 vs. 8 | 1-0, 1.000 | Higher seed scoring margin: +2.0 ppg
Three and eight seeds have met just once in the Elite Eight. In 1998, third-seeded Stanford ended Rhode Island's long-shot run.

3 vs. 9 | 1-0, 1.000 | Higher seed scoring margin: +8.0 ppg
The only Elite Eight matchup involving a nine seed occurred more than a decade ago, in 1994, when third-seeded Florida beat Boston College.

4 vs. 6 | 2-1, .667 | Higher seed scoring margin: +6.7 ppg
Once they slip by top seeds in the Sweet 16, four seeds are a pretty resilient bunch. In addition to going 4-4 against two and three seeds, they're a solid 5-1 against lower seeded opponents. The only lower seed to win was Kansas in 1988 when the sixth-seeded Jayhawks and Danny Manning beat their rival, fourth-seeded Kansas State and Mitch Richmond. In the other two matchups, four seed Georgia Tech vs. six seed Minnesota in 1990 and four seed Cincinnati vs. six seed Memphis State in 1992, the higher seed held sway. In all three 4 vs. 6 matchups, the team that allowed the fewest points per game prevailed.

4 vs. 7 | 1-0, 1.000 | Higher seed scoring margin: +8.0 ppg
The 2005 tourney saw the only 4 vs. 7 quarterfinal matchup in the 64-team era. Fourth-seeded Louisville (actually a one or two seed in disguise) burst seven seed West Virginia's bubble.

4 vs. 10 | 2-0, 1.000 | Higher seed scoring margin: +3.5 ppg
It has been 10 tourneys since a four and 10 seed have squared off in the quarterfinals. In 1997, fourth-seeded Arizona knocked off Pete Gillen's 10th-seeded Providence Friars. Then the Wildcats went on to win their only championship of the modern era. In the other 4 vs. 10 matchup seven years earlier, four seed Arkansas beat 10 seed Texas.

5 vs. 10 | 1-0, 1.000 | Higher seed scoring margin: +12.0 ppg
A five seed has played a 10 seed only once in the Elite Eight. In 2002, fifth-seeded Indiana ended 10th-seeded Kent State's Cinderella run and eventually lost to Maryland in the championship game.

6 vs. 8 | 0-1, .000 | Higher seed scoring margin: -4.0 ppg
Here's another matchup that has only happened once. In 2000, eight seed Wisconsin upended Big Ten rival Purdue.

7 vs. 8 | 0-1, .000 | Higher seed scoring margin: -4.0 ppg
In the only Elite Eight matchup between these two middle seeds, eighth-seeded North Carolina beat seven seed Tulsa in 2000 -- the same year eight seed Wisconsin beat Purdue in the only 6 vs. 8 matchup. Eerie.

FINAL FOUR

The semifinals mark the point in the tourney where seeding offers virtually no guidance to the outcomes. For one thing, 12 of the 50 Final Four games in the modern era have involved like-seeded opponents. Second, of the 38 remaining games, the higher seed is just 23-15. In matchups where the gulf between opponents is one or two seeds, the higher seed is just 13-11; in games where the difference in seed position between opponents is more than two, the higher seed holds a solid 10-4 record. The keys to predicting the like-seeded matchups with 90 percent proficiency are conference affiliation, playing location and winning record. In general, Big Six conference teams with lesser records playing closer to their campus prevail. For tossup games, where the seed difference between teams is one or two positions, frontcourt scoring accurately predicts the outcome in 16 of the 24 matchups.

1 vs. 1 | 11-11, .500 | Average point spread: +8.8 ppg
Of the 50 semifinal games played in the modern tourney, only 11 have pitted top seeds against each other. That might be reason to avoid penciling too many top seeds into your Final Four. Then again, just two tourneys ago, both semifinal games featured top-seed matchups. If you had observed the following rules in order, you would've picked all 11 of these matchups right: 1) take any top seed whose average points scored are at least 15 higher than its opponent, 2) take any top seed with at least a 15-game winning streak, 3) take the top seed playing significantly closer to its campus (a gulf of at least 150 miles), and 4) take the top seed with the better guards.

Recent matchups: Memphis over UCLA, 2008. Kansas over North Carolina, 2008. Maryland over Kansas, 2002. Duke over Michigan State, 1999. Kentucky over Minnesota, 1997.

1 vs. 2 | 5-4, .556 | Higher seed scoring margin: +0.6 ppg
Top seeds hold a one-game edge over two seeds in the Final Four. The key performance indicators in the nine matchups have been coaching experience, momentum and frontcourt strength. Avoid any team with a coach who either hadn't been to the Elite Eight or only gotten there once. Pick against any team coming to the dance with two or more pre-tourney losses in a row. Then, pick the team that gets the higher percentage of scoring from its frontcourt. By following these guidelines, you'd pick eight of the nine games correctly. The only exception occurred in 1991, when Duke ruined UNLV's perfect season en route to the championship. The Blue Devils relied more on their backcourt for points than the Runnin' Rebels. As for last year, the Huskies had lost twice in a row before entering the dance, so MSU would've been the team to pick -- and the Spartans upheld the rules.

Recent matchups: Michigan State (2) over Connecticut (1), 2009. Florida (1) over UCLA (2), 2007. Ohio State (1) over Georgetown (2), 2007. Connecticut (2) over Duke (1), 2004. Arizona (2) over Michigan State (1), 2001.

1 vs. 3 | 2-4, .333 | Higher seed scoring margin: -1.2 ppg
The toughest matchup for top seeds in the entire tournament is in the semifinals when they face three seeds. They've only won twice in six tries, when Duke broke the curse in 2001, upending third-seeded Maryland -- and last year, when North Carolina beat Villanova. The key to this matchup, as with the 1 vs. 2, is frontcourt scoring. The team that relies on forwards and centers for the higher percentage of its scoring load is a perfect 6-0. Last year, the Tar Heels got nearly half their points from the frontline, while Villanova only got a third.

Matchup history: North Carolina (1) over Villanova (3), 2009. Syracuse (3) over Texas (1), 2003. Duke (1) over Maryland (3), 2001. Utah (3) over North Carolina (1), 1998. North Carolina (3) over Kansas (1), 1991. Michigan (3) over Illinois (1), 1989.

1 vs. 4 | 4-1, .800 | Higher seed scoring margin: +7.0 ppg
What a difference one seed makes. Whereas top seeds struggle against three seeds, they have little trouble with fours, winning all but one of the five matchups. The only fourth-seeded squad to rain on a top seed's parade was Arizona, which parlayed a 1997 Final Four upset over North Carolina into a national championship. What did the Wildcats have that the other four seeds lacked? An explosive offense. Arizona was the only four seed in this matchup that actually averaged seven points per game more than its opponent.

Matchup history: Illinois (1) over Louisville (4), 2005. Connecticut (1) over Ohio State (4), 1999. Arizona (4) over North Carolina (1), 1997. UCLA (1) over Oklahoma State (4), 1995. UNLV (1) over Georgia Tech (4), 1990.

1 vs. 5 | 1-0, 1.000 | Higher seed scoring margin: +16.0 ppg
The North Carolina-Michigan State Final Four game in 2005 marked the first time that a five seed played a top seed for the right to advance to the championship. The Tar Heels stomped the Spartans 87-71.

1 vs. 8 | 1-0, 1.000 | Higher seed scoring margin: +12.0 ppg
Michigan State was also involved in the only 1 vs. 8 semifinal. The top-seeded Spartans beat Big Ten rival Wisconsin on their way to the 2000 championship.

2 vs. 2 | 1-1, .500 | Average point spread: +7.0 ppg
Amazingly, two seeds have squared off against each other in the Final Four only once in the modern tourney era. It happened in 1995, when Arkansas beat North Carolina. Just as with the 1 vs. 1 matchup, proximity was a reliable guide in predicting the outcome of this like-seeded game. Arkansas was playing closer to home than North Carolina and prevailed by seven points.

2 vs. 3 | 3-2, .600 | Higher seed scoring margin: +2.8 ppg
Surprisingly, the 2 vs. 3 matchup occurs almost as often in the Final Four as a 1 vs. 2 game. Two seeds prevail in the best-of-5 series, but by the slimmest of margins. If you took the two seed in every situation except when the three seed was from the ACC or Big East, you'd be a perfect 5-0 in predicting outcomes.

Matchup history: Georgia Tech (3) over Oklahoma State (2), 2004. Kansas (2) over Marquette (3), 2003. Kentucky (2) over Stanford (3), 1998. Duke (2) over Florida (3), 1994. Seton Hall (3) over Duke (2), 1989.

2 vs. 4 | 1-0, 1.000 | Higher seed scoring margin: +14.0 ppg
Amazingly, the 2006 UCLA/LSU clash marked the only time that a two seed has faced a four in the Final Four. The second-seeded Bruins had no trouble dispatching the Tigers 59-45.

2 vs. 5 | 0-1, .000 | Higher seed scoring margin: -9.0 ppg
In 2002, Indiana sprung a mild upset when the fifth-seeded Hoosiers upended Oklahoma in the only 2 vs. 5 matchup of the 64-team era.

2 vs. 6 | 1-1, .500 | Higher seed scoring margin: +3.5 ppg
These seeds haven't played each other in the Final Four since 1988, when Kansas and Danny Manning upset Duke on its way to Larry Brown's only NCAA championship. The year before that, two seed Syracuse staved off Providence.

2 vs. 8 | 0-1, .000 | Higher seed scoring margin: -7.0 ppg
The first four years of the modern tourney era saw some of the Final Four's funkiest matchups. Here's another one: in 1985, eight seed Villanova beat two seed Memphis State before its date with destiny against Georgetown.

2 vs. 11 | 1-0, 1.000 | Higher seed scoring margin: +11.0 ppg
One year after the improbable Memphis State/Villanova matchup, 11 seed LSU lost to two seed Louisville, which went on to beat Duke in the 1986 final.

3 vs. 4 | 1-0, 1.000 | Higher seed scoring margin: +14.0 ppg
In 1990, just six years into the 64-team era, the only 3 vs. 4 semifinal saw three seed Duke holding off Arkansas. Duke's reward for the victory was the privilege of getting steamrolled by UNLV 103-73 in the finals.

3 vs. 11 | 1-0, 1.000 | Higher seed scoring margin: +15.0 ppg
If you thought it was amazing that 2006's 2 vs. 4 matchup was the only time the two seeds squared off in the modern era, how's this for uncanny? The other matchup in the 2006 Final Four was also a first. In the only 3 vs. 11 semi-final matchup of the 64-team era, third-seeded Florida trounced 11 seed George Mason 73-58.

4 vs. 5 | 1-0, 1.000 | Higher seed scoring margin: +8.0 ppg
In 1996, four seed Syracuse beat Mississippi State in the modern tourney's only 4 vs. 5 semifinal matchup. The Orangemen lost to Kentucky in the finals.

4 vs. 6 | 0-1, .000 | Higher seed scoring margin: -4.0 ppg
In 1996, Michigan's Fab Five -- a six seed that could've been a two seed -- beat fourth-seeded Cincinnati. The Wolverines then got trounced by Duke 71-51 in the finals.

5 vs. 8 | 1-0, 1.000 | Higher seed scoring margin: +12.0 ppg
2000 marked the only year of the modern tourney era that had two eight seeds in the Final Four. In addition to the 1 vs. 8 matchup between Michigan State and Wisconsin, fifth-seeded Florida squared off against eight seed North Carolina and won by a dozen. The Gators wound up losing to MSU in the finals.

CHAMPIONSHIP

As little an impact as seeding had on Final Four outcomes, you'd think it wouldn't make a difference in the finals. Not so. Of the 19 championship games involving teams with different seeds, the higher seed has won 13. Since 1990, higher seeds are 12-3 against their lower seeded opponents. Another surprising fact about the finals is that the matchup the brackets were intended to yield -- a 1 vs. 1 showdown -- has happened only five times in 25 years. Then again, it's occurred in two of the past three years, with Kansas beating Memphis in 2008 and Florida downing Ohio State in 2007. The only other like-seeded matchup involved three seeds Michigan and Seton Hall in 1989. The Wolverines' frontcourt scoring was a key to their victory -- as it was in four of the other five like-seeded showdowns. The bottom line: If you went with the higher seed in championship games and the team with the better front line in like-seeded battles, your prediction rate would be 72 percent (18-7).

1 vs. 1 | 5-5, .500 | Average point spread: +6.0 ppg
About one in five tournaments features two heavyweight top seeds going toe to toe. Since it happened in 2008 with Kansas/Memphis and the year before with Florida/Ohio State, the law of averages says it won't happen for a few seasons. Which statistic points to the top-seeded victor? Pre-tourney winning streak. In every instance but one, the team that had the shorter streak going into the tourney prevailed. That one exception was 2008, when both Memphis and Kansas came to the dance with seven consecutive wins. So when in doubt, pick the team with the better frontcourt; the deeper you go in the dance, the more that big men make a difference.

1 vs. 2 | 5-1, .833 | Higher seed scoring margin: +6.7 ppg
Strange that seeding should have such an impact between two close seeds in the finals. Top seeds treat two seeds like one of those lowly seeds they face in the early rounds, beating them 83 percent of the time. The only two seed to buck the trend was Louisville, which beat Duke in the 1986 championship game. Actually, the more reliable performance indicators in this matchup are coaching experience and scoring margin. Teams with coaches who haven't been to the Elite Eight in a previous year are 0-2. In the remaining four matchups, the team with the higher scoring margin has always prevailed ... like last year, when the Tar Heels (17.4 ppg margin) beat the Spartans (9.0 ppg margin).

Matchup history: North Carolina (1) over Michigan State (2), 2009. Duke (1) over Arizona (2), 2001. UCLA (1) over Arkansas (2), 1995. Arkansas (1) over Duke (2), 1994. Indiana (1) over Syracuse (2), 1987. Louisville (2) over Duke (1), 1986.

1 vs. 3 | 1-0, 1.000 | Higher seed scoring margin: +30.0 ppg
You'd think a 1 vs. 3 championship game would've happened more than once. For Duke fans, once might be enough, considering how soundly UNLV throttled the Blue Devils 103-73 in the 1990 finals. The Blue Devils would get their revenge the following year, when they ruined UNLV's perfect season in the 1991 semifinals.

1 vs. 4 | 1-1, .500 | Higher seed scoring margin: +2.0 ppg
The two 1 vs. 4 championship games happened in successive years -- and they both involved top-seeded Kentucky. In 1996, Kentucky took care of Syracuse, but the following year, the Wildcats were upset in overtime by fourth-seeded Arizona.

1 vs. 5 | 2-0, 1.000 | Higher seed scoring margin: +9.5 ppg
The two 1 vs. 5 finals matchups came two years apart. Michigan State beat Florida in 2000, and Maryland handled Indiana in 2002.

1 vs. 6 | 1-1, .500 | Higher seed scoring margin: +3.0 ppg
If it weren't for Villanova's upset of Georgetown in 1985, Kansas could lay claim to springing the biggest championship upset of the modern tourney era. The Jayhawks beat top seed Oklahoma in 1988. Four years later, Michigan's Fab Five tried to duplicate the feat but was thumped by Duke.

1 vs. 8 | 0-1, .000 | Higher seed scoring margin: -2.0 ppg
In the first year of the modern tourney era, the championship game saw its most unlikely matchup -- and most surprising outcome. Eighth-seeded upstart Villanova toppled overwhelming favorite Georgetown 66-64, playing a near flawless game that included 90 percent shooting in the second half.

2 vs. 3 | 3-2, .600 | Higher seed scoring margin: +1.2 ppg
Florida's 2006 win over second-seeded UCLA tightened this matchup, which, along with 1 vs. 2 final games, is the most popular championship seed pairing. The first three seed to beat a two seed in the finals was Syracuse, which knocked off Kansas in 2003. (Syracuse also holds the distinction of being one of only two champions that hadn't gone to the previous year's tourney; the other was Louisville in 1986.) The key to this matchup is scoring balance. The squad with the smallest percentage gap between its frontcourt and backcourt scoring has won all five games.

Matchup history: Florida (3) over UCLA (2), 2006. Connecticut (2) over Georgia Tech (3), 2004. Syracuse (3) over Kansas (2), 2003. Kentucky (2) over Utah (3), 1998. Duke (2) over Kansas (3), 1991.

3 vs. 3 | 1-1, .500 | Average point spread: +1.0 ppg
The only other like-seeded finals matchup of the 64-team era besides the five 1 vs. 1 games saw Michigan squeak by Seton Hall in overtime in 1989.

Peter Tiernan has been using stats to analyze March Madness for 20 years. His insights into the NCAA basketball tournament can help you build a better bracket. E-mail him at bracketscience@comcast.net or visit bracketscience.com.