Signs of high-seed tourney pretenders

Over the weekend, I posted a blog about the 15 factors that correlated to overachievement in the tourney. Today, instead of identifying the types of teams that could make surprise runs in the dance, we’re going to look at the characteristics that disqualify high seeds from reaching the Final Four. I’m restricting this analysis to the one, two and three seeds because they’ve made up nearly 80% of the semi-finalists since the tourney expanded to 64 teams in 1985. Of the 112 Final Four teams of the last 28 years, 85 of them (79%) have been among the top three seeds. Not only that, but these seeds have accounted for 25 of the 28 champions.

If you separate the teams that make the Final Four from those that don’t at each of the top three seed positions, you’ll find that the pretenders tend to share attributes that the contenders don’t possess. Let’s look at the one, two and three seeds individually and identify the statistical signs that point to a bracket breakdown before the Final Four.

Eight signs of one seed pretenders

Top seeds are by far the most successful teams in the tourney. More than 40% of them—46 of 112—reach the Final Four. Of course, that means there are 66 one seeds that fall by the wayside before the last weekend of the dance. Beware of top seeds with any one of these telltale pretender attributes:

  • A scoring margin of 10.5 points or fewer per game
  • Imbalanced scoring, with more than 73 percent of points from either the backcourt or frontcourt (forwards and centers)
  • A strength of schedule ranked higher than 80
  • A “snake-bit” coach (more than four tourney trips without an Elite Eight run)
  • A coach with fewer than seven tourney appearances leading a team either without an All-American or that didn’t go to the previous dance.
  • A combined rebounding and turnover margin worse than five possessions a game
  • An assist/field goal ratio higher than 66% (signaling the lack of a “go-to” guy)
  • A Pythag ranking worse than fifth, suggesting an over-seeded team

Only one of 37 top seeds possessing any of these attributes has managed to reach the Final Four. The lone exception was Texas in 2003. Coach Rick Barnes had been to the dance 11 times without making an Elite Eight run. Top seeds that aren’t hampered by any of these weaknesses are 45 for 75 in reaching the Final Four. That works out to a 60% success rate—and these teams also own an impressive +.532 PASE. Compare that to the pretenders, who fall short of seed expectations by more than a game per tourney (-1.078 PASE).

A 60% success rate in reaching the Final Four is 46% better than the typical 41% advancement rate of top seeds. Perhaps more importantly, these exclusion factors identify 45 of the 46 actual contenders.

Ten signs of a two seed pretender

Since 1985, 25 of the 112 two seeds have made it to the semi-finals. That works out to a 22% Final Four advancement rate. How can you separate the second-seeded contenders from the pretenders? Cast a cold eye on any two seed with one or more of these ten characteristics:

  • A coach who’s been to the tourney less than four times
  • A winning percentage less than .765 or more than .900
  • Less than six wins in the last 10 pre-tourney games
  • A scoring defense allowing less than 56.9 points a game
  • A scoring margin less than 8.8 or more than 18.4 points a game
  • A strength of schedule ranked softer than 40
  • A rebounding margin of less than two boards per game
  • A combined rebounding/turnover margin of three possessions or fewer per game
  • An assist-to-field goal percentage of 51 percent or lower
  • More than one freshman starter or three senior starters

Two seeds that possess any of these qualities are an astounding 0 for 62 in their quest to reach the Final Four and their PASE is a woeful -.758. On the other hand, two seeds without any of these warning signs are 25 for 50 in getting to the last weekend of the dance. And they overachieve at a rate of almost a game per tourney (+.940 PASE). This 50% Final Four advancement rate is more than twice as good as the typical 22% success rate of second-seeded squads in reaching the semi-finals.

Fourteen signs of three seed pretenders

Only 14 of 112 three seeds have reached the Final Four in the 64-team bracket era. That’s just a 12.5% success rate. How can you spot one of the 98 three-seed pretenders? Beware of any team with one of these 14 attributes:

  • A Mid-Major without an All-American
  • A rookie tourney coach leading a team that didn’t go to the last dance
  • A snake-bit coach with more than four dances and no Elite Eight runs leading a team with fewer than four straight bids
  • Less than six or more than eight wins in the last 10 pre-tourney games
  • A pre-dance losing streak of two or more
  • A scoring average lower than 73 points per game
  • A points-against average lower than 64 points per game
  • A scoring margin less than seven points per game
  • A starting unit that gets less than 72 percent of the points
  • A backcourt that gets more than 72 percent of the points
  • A rebounding margin of less than three boards a game
  • A turnover margin favoring the opponent by more than one turnover
  • A strength of schedule ranked softer than 66
  • Fewer than 29 percent of field goals attempted beyond the three-point line

Three seeds with any of these qualities are a dismal one for 73 in reaching the Final Four. The only team to defy the odds was Rick Majerus’ Utah Utes in 1998. Meanwhile, three seeds without any disqualifiers are 13 for 39 in getting to the final weekend of the dance. That’s a 1-in-3 advancement rate—nearly three times better than the overall 1-in-8 rate for three seeds.

Separating high-seed contenders from pretenders: the overall impact

All tolled, the top three seeds in the tourney have had a 25% success rate in reaching the Final Four (85 for 336, 25.3%). But teams with any of the pretender characteristics cited above are just a measly two for 172—1.2% proficient—in getting to the semi-finals. If you eliminated these teams from consideration in your bracket pondering, the rest of the 1-3 seeds are 83 of 164 in hanging around to the last weekend of the dance. That’s better than a 50/50 success rate—and twice as good as the typical rate of the top three seeds.

For the last five years, I’ve filled out a bracket model that starts with this seed-by-seed “contender/pretender” method for identifying Final Four candidates. That model contemplates the top eight seeds. After identifying the Final Four candidates, I then use a second analysis to narrow down on my ultimate semi-finalists. The model has performed in the 85th percentile of the tourney challenge contests run by major sports sites. Some years, like 2009 (97th percentile) and 2010 (98th percentile), it would’ve been in the running to win an average-sized pool. Other years, like last year (62nd percentile), it wouldn’t have done so well.

After Selection Sunday, I’ll post the bracket that this methodology yields, along with two other stats-based bracket models. I’m never one to crutch exclusively on what the numbers say come tourney time—largely because even the best approach to picking your bracket is a low-probability proposition. But it’s always good to know what the historical data says about how the bracket will play out, if for no other reason than to be aware of when you’re taking a big risk.

This entry was posted in General News, Tourney Trends. Bookmark the permalink.

4 Responses to Signs of high-seed tourney pretenders

  1. Mike L. says:

    Great work, Pete!!! I was wondering if you could clarify a few of these for me?

    For 2-seeds, you said “More than one freshman starter or three senior starters”. Does the latter part mean “More than three senior starters” or does it mean “Exactly three senior starters”.

    For 3 seeds, you said “A rookie tourney coach leading a team that didn’t go to the last dance”. Does ‘rookie tourney coach’ mean a coach that has never ever made a tourney appearance or does it mean a coach that has not made a tourney appearance with his current team?

    Also, I know you do a ton of statistical analysis. Do you ever venture into meta-analysis, or more specifically, looking for macro-events that alter micro-factors. For example, the NCAA recently said they are looking into ways to increase scoring. Some of your metrics involve scoring avgs, so if the NCAA finds ways to artificially boost scoring, there will be a lot more teams meeting the criteria on the offensive side but less meeting the criteria on the defensive side. I guess what I’m asking is do you look for conditions where the rules/criteria may no longer apply. I’m a big believer that the one-and-done rule has significantly changed the MCBB landscape, I just haven’t found a way to quantify it yet. Again, great work and I look forward to hearing from you.

    • ptiernan says:

      It means more than three senior starters. I took a grammatical shortcut. Thanks for pointing out. I’ll clarify.

      On the 3 seeds, it means a coach making his first tourney appearance, regardless of coach.

      I wrote a blog on the one-and-done rule…I think it had to do with team age. The short answer to your meta-analysis question is: no. That said, I recognize that scoring has gone down steadily since its height in 1989 when the average tourney team scored 83 points a game. Now they score about 73. One year, I did my PASE analysis based on “points below and average average scoring for the field.” That’s probably a better way to do it–or possession-based data.

      I also noted that the 2002 change to all Big Six conferences having their own tourney in advance of the dance had a significant impact on the importance of momentum. Check out that blog in the tourney trends category.

      Thanks for writing me, Mike!

      Pete

  2. Scott says:

    I could be reading this wrong, but in the 3 seed section you say one of the criteria is: ‘A points-against average lower than 64 points per game’. Do you mean higher than? The same can be said for the scoring defense bullet point in the 2 seed section. The other way around didn’t make sense to me.

    Otherwise, great stuff! I’ve been reading all of your articles for the past 4 hours!!

Leave a Reply

Your email address will not be published. Required fields are marked *

* Copy This Password *

* Type Or Paste Password Here *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>