Scoring offense, seeding are key NCAA indicators

Thanks to ESPN's Bracketology site, there's a wealth of information to help you decide which teams could advance in your bracket.

The problem is, there's almost too much information … and not enough guidance on which statistics really matter. In determining the outcome of a second-round
matchup against, say, a fourth-seeded Big Ten squad and a fifth-seeded ACC team, what are the key considerations? Does conference affiliation give a clue to who will win? What about scoring offense, coaching or team experience? And how about pre-tourney momentum, guard play, or where the game is played?

Clearly, seeding gives the best clue to a team's fate. Squads seeded one, two or three have won 18 of the last 21 tourneys, and no team below an eight seed has won in the 64-team era. Beyond seeding, however, the relative value of other factors is unclear -- at least until now. We did a statistical analysis quantifying whether teams with 16 key attributes exceeded or fell short of the win totals that their seeding would dictate.

We'll list the top 10 overall contributors to overachievement in the tourney and the bottom five factors for overall underachievement. Then we'll examine the key characteristics of over- and underperformance for each of the seed classes -- favorites (one and two seeds), contenders (three through six seeds), toss-ups (seven through ten seeds), longshots (11 through 14 seeds) and pushovers (15 and 16 seeds).

Top 10 overall contributors to overachievement
We compared the number of games that teams possessing certain characteristics won at each seed position to the number of wins the average team won with that seed. The difference between actual and expected wins divided by the number of teams possessing the given characteristic allows us to compare each characteristic's performance against seed expectations (PASE). Here are the top 10 overall contributors to overachievement in the tourney:

Besides seeding, the factor that contributes most to overperformance in the tourney is having a star player. The 180 teams with preseason All-Americans had a PASE of 0.193, which says on average, they won about a fifth of a game more than their seeding would dictate.

To put it another way, about one in five star-led teams wins a game more than the average team at any given seed. The trick with this factor is identifying which teams have preseason All-Americans. Over the last 21 years, the Associated Press has published its preseason selections -- and those are the selections on which these stats are based.

However, the AP used to list its first and second teams. However, in recent years, they're only publicizing their first team -- and occasionally mentioning a sixth man. This year, for instance, the AP's preseason All-American list includes J.J. Redick (Duke), Shelden Williams (Duke), Dee Brown (Illinois), Craig Smith (Boston College) and Adam Morrison (Gonzaga). Rudy Gay of Connecticut was cited as the sixth-highest vote getter. So just five teams would satisfy the "preseason All-American" characteristic. But in past years, 10 teams qualified as having a star.

Certainly, there are a host of teams you might also want to put in the star-led category -- Arizona (Hassan Adams), Arkansas (Ronnie Brewer), California (Leon Powe), Indiana (Marco Killingsworth), Louisville (Taquan Dean), Michigan State (Paul Davis or Maurice Ager), Ohio State (Terence Dials), Oklahoma (Taj Gray), Stanford
(Chris Hernandez), and Texas (Daniel Gibson) to name just a few. Just be careful in how liberal you get with assigning star status to a squad, because the statistics were not calculated on the basis of conjecture.

Fortunately, most of the other characteristics on the overachievement list are straightforward.

The second highest contributor to tourney overperformance is team experience. The 252 schools that have been to the tournament more than four straight times have a PASE of 0.154. That translates to about three tourney-tested teams in 20 winning one more game than seeding expectations. The third best contributor to overachievement is offensive output. The 385 teams that average scoring three points or better per game than the average scoring output of the tourney field have a PASE of 0.113; one in nine of these offensive-minded squads outperform expectations.

Winning percentage is the fourth highest overachievement indicator. With a PASE of 0.110, one in nine teams winning at more than a 0.875 rate also advance an extra game more than expected. Rounding out the top five overachievement list is coaching experience. Teams with coaches that have been to the tourney more than 10 times have a PASE of 0.108. All the other characteristics on the overachievement list are below a 0.100 PASE, meaning that fewer than one in 10 teams outperform expectations.

What factors are missing from the overachievement list? Surprisingly, strong bench play, upper-class teams, dominant guard play and the RPI have hardly any impact on a team's ability to exceed seed expectations. Bench play comes the closest, but with a PASE of 0.020, only an average of one in 50 teams getting more than 20 percent of their scoring from the bench win a game more than their seeding would dictate.

Bottom five overall contributors to underachievement
It's one thing for a factor to have virtually no impact on tourney overachievement; it's another when it actually leads to underachievement. Here are the five biggest contributors to underachievement in the tourney:

Not surprisingly, since potent scoring is a leading indicator of overachievement, lack of scoring punch is the top indicator of underperformance. The 696 teams that average fewer points per game than their year's tourney field won 34.4 games fewer than the 700.4 that they should've, yielding a -0.049 PASE. That means about one in 20 of these offensively-challenged teams will fall a game short of seed expectations.

The second leading indicator of underperformance is scoring balance. Oddly enough, teams that get between 40-to-60 percent of their scoring from both their front- and backcourt have a PASE of -0.043. Conversely, more imbalanced teams -- squads getting more than 60 percent of their scoring either from the frontcourt (0.057 PASE) or the backcourt (0.013 PASE) -- surpass seed expectations slightly.

Another contrary finding involves pre-tourney momentum. Despite the pundits' claims that hot teams hold an advantage in the postseason, having more than a one-game winning streak going into the tourney actually leads to underachievement. Six hundred and seventy one teams have come to the dance with two or more straight wins, so they should've won 554 games, but only won 529. That works out to a -0.037 PASE or about one in every 25 teams falling a game short of seed expectations.

The other two characteristics on the underachievement list aren't nearly so surprising. The 338 teams from small conferences are 10 wins short of the 60 their seeding suggests they should win for a -0.029 PASE. Rounding out the bottom five, teams with more than one senior in their starting lineup have a PASE of -0.020. That might seem unusual, but when you consider that few college stars make it to their senior season, the statistic makes sense.

Top performance indicators for favorites (one and two seeds)
One and two seeds -- the favorites in the tourney -- don't need much help advancing through the bracket. Still, there are a number of factors that contribute to favorites overachieving in the tournament. The top three are:
1. More than 60 percent of scoring from the frontcourt (0.409 PASE)
2. Preseason All-American on team (0.313 PASE)
3. More than 80 points scored per game (0.264 PASE)

On average, two out of five one and two seeds that rely on their frontcourt for more than 60% of their scoring win a game more than seed expectations say they should. So much for the theory that guard play is the key to advancing deep in the dance. In fact, favorites getting more than 60% of their scoring from guards actually underperform slightly (-0.051 PASE).

The second biggest contributor to overachievement among favorites, not surprisingly, is having a star player. About one-third of the top two seeds with a star advance an extra game further than their seed dictates. Scoring offense also helps, as over one out of four squads with a scoring average higher than 80 points a game overperform in the tourney.

Which favored seeds are more prone to underachieving in the tourney?

The two biggest indicators that a one or two seed is going to fall short of seed expectations are balanced scoring and difficulty putting up points. Teams that are unable to dominate from either the front- or backcourt, getting between 40-and-60 percent of their points from each, have a PASE of -0.270. That means one in four such teams, on average, fall a game short of expectations. The same goes for favored seeds that score three fewer points per game than the average tourney team. These squads have a -0.256 PASE.

Top performance indicators for contenders (3-6 seeds)
Three, four, five and six seeds, the contender teams in the tourney, regularly advance one or two rounds -- and occasionally sneak up on the favored seeds to make a championship run. When you fill out your bracket, it's good to know which of these squads has what it takes to perform above seed expectations. The top three indicators of a contending overachiever are:
1. Winning percentage greater than 0.850 (0.388 PASE)
2. At least 20% of scoring from bench (0.130 PASE)
3. More than 60% of scoring from frontcourt (0.123 PASE)

Far and away, the key characteristic for overperformance among three through six seeds is winning percentage.

Nearly 40% of the contender teams with a 0.850 winning rate (a 24-4 record or better) win a game more than seeding expectations. The second highest contributor to overachievement is a strong bench that gets at least 20% of its team's points.

However, with a PASE of 0.130, only about one in eight of these teams advances a game beyond their seed expectations. The same is true for teams that rely on their forwards and centers for more than 60% of their scoring. (There's that stat again. It's amazing how important frontcourt scoring is to overachievement among the top six seeds; these are the teams that ultimately will vie for Final Four slots).

How do you spot an underachieving contender seed? Look at pre-tourney momentum and inflated RPI values.

Contenders that come into the tourney with more than a four-game winning streak have a PASE of -0.215, so at least one in five are likely to fall short of their seed's typical win total. Oddly enough, about one in 12 (-0.085) three through six seeds underachieve when they're seeded at or below where their RPI suggests they ought to be. For instance, last year Kansas had an RPI of two going into the tourney.

By all rights, they could've been a top seed. In fact, they were a third seed -- a position where you'd expect teams with RPI values of 9 to 12 to be. If the RPI was a true indicator of the relative quality of teams, teams like Kansas should perform better than seed expectations. In fact, they do the opposite -- just one more indication of the questionable value of the RPI.

Top performance indicators for toss-ups (7-10 seeds)
Seven, eight, nine and 10 seeds have a 50/50 chance of advancing past the first round (because they play each other) and about a one in four chance of getting to the Sweet 16. Here are the top three signs of an overachieving toss-up seed:
1. More than four straight tourney trips (0.182 PASE)
2. Scoring per game better than field average (0.157 PASE)
3. Playing within 500 miles of campus (0.134 PASE)

Nearly one in five toss-up seeds that have been to the tourney more than four consecutive times outperform expectations. One in six of these seeds overachieve when they average more points per game than the tourney field. And about one in seven toss-up squads beat seed expectations when they're playing within a day's trip of their campus.

Ironically, while team experience is the key to overachievement, coaching experience is the biggest sign of an underachiever. Seven through 10 seeds with coaches who've been to the tourney more than 10 times have a PASE of -0.116. Why do one in nine veteran coaches underperform as toss-up seeds? One theory is that they're overseeded. Sometimes, the reputation of a high-profile coach forces his team to be seeded higher than performance warrants.

There's no mystery to the second biggest contributor to toss-up underachievement. Seven through 10 seeds that score less than the average tourney entrant have a PASE of
-0.112 -- further confirmation that scoring offense is a key litmus test of tourney performance across all seed categories.

Top performance indicators for longshots (11-14 seeds)
The Cinderella teams of the tourney -- 11 through 14 seeds -- have less than a one in four chance of springing an opening-round upset, and only 29 of the 336 longshot seeds have won their first two games. Which of these teams are most likely to overachieve? Look for these telltale signs:
1. More than two straight tourney trips (0.283 PASE)
2. Team from Big Six conference (0.212 PASE)
3. Coach with more than 10 tourney appearances (0.118 PASE)

Team experience is the top indicator of a Cinderella. The 49 longshot seeds with more than two consecutive tourney trips should've won only 19 games; they actually won 33. It also helps to be from one of the Big Six conferences -- the ACC, Big East, Big Ten, Big 12, Pac-10 or SEC. The 37 longshot seeds from the Big Six have won 25 games, about eight more than expectations. Finally, unlike toss-up seeds, longshots thrive under veteran coaches. About one in nine 11 through 14-seeded teams with coaches who've been to the tourney more than 10 times outperform seed expectations.

There's only one significant sign of an underachieving longshot: Membership in a small conference. The 159 longshots from small conferences should've won 44 games; they've only sprung 37 upsets, for a -0.158 PASE.

Top performance indicators for pushovers (15 and 16 seeds)
Not that you would ever pick a 15 or 16 seed to advance in your bracket, but the four pushovers that have sprung upsets in the first round do share similar characteristics. The two key signs of an overachieving pushover are pre-tourney momentum and proximity to campus. Only 36 of the 168 pushovers have come into the tourney winning nine or 10 of its last 10 games; all of the four 15 seeds to spring upsets are among them. Two of the 20 pushovers playing within 500 miles of their campus have sprung upsets, and only two of the remaining 148 pushovers accomplished the feat.

Freelance writer Pete Tiernan has been studying the NCAA Tournament for 16 years. E-mail him here.