For many years now, college basketball analysts have filled the airwaves in February with talk of tournament bubbles and resumes, and for the longest time beauty was in the eye of the beholder since the NCAA had no real standardized method for judging teams. Comparing those with gaudy records in weaker leagues to teams with more losses in power conferences is never easy, but a couple of years ago a “quadrant” system was born to add a bit of consistency.
For the uninitiated, this is how quadrants are defined:
Quadrant 1: Home 1-30; Neutral 1-50; Away 1-75.
Quadrant 2: Home 31-75; Neutral 51-100; Away 76-135.
Quadrant 3: Home 76-160; Neutral 101-200; Away 136-240.
Quadrant 4: Home 161-plus; Neutral 201-plus; Away 241-plus.
While this system has a fatal flaw in the way game location is weighted (more on that later), this at least added some clarity to what constitutes a quality win or loss. When the quadrant system was created, teams were ranked based on a wonky Ratings Performance Index, which incorporates three components: the team’s winning percentage, their opponents’ winning percentage, and their opponents’ opponents’ winning percentage.
Needless to say, much is missing from that equation. If game location is as critical as the quadrant system would lead one to believe, then where teams played their toughest opponents should be weighted somehow. This metric is also missing any measure of efficiency or dominance, instead focusing purely on wins and losses. As longtime college basketball statistical analyst Ken Pomeroy noted back in 2013, this is illogical.
However, if you are interested in evaluating a team’s ability and their prospects for future success, you’re better served by looking at more than just how many wins they have or who they beat. You need to know how well they played in those games. If a team needs a buzzer-beater to get by Hawaii, and struggles at home with Gardner-Webb and Auburn, those are warning signs about the prospects for a team’s future success, despite the fact they may have won 12 games in a row to that point.
It would be in any analyst’s best interest to understand this and not parrot the cliché that said team (not naming names!) just “knows how to win”. The teams that truly know how to win are the ones that have a cushion at the end of the game. They aren’t at the mercy of wacky officiating, whimsical replay reviews, or half-court heaves.
This isn’t unlike our criticism of the football playoff committee that willingly ignores all advanced data when choosing the four best teams. To the NCAA’s credit, however, they scrapped RPI last season in favor of a new metric that considers some additional information. Enter the NET rankings.
While no one seems to know the exact formula, the components are listed in order of weighting. As you can see, winning games matters a ton, as wins and winning percentage are accounted for in the first and third components. The first component also incorporates game location and strength of schedule, the second rewards teams for efficient play, the fourth places additional emphasis on game location and the fifth adds a margin of victory element. To be sure, it isn’t a perfect metric - efficiency should ideally have some sort of opponent adjustment, ten points is too low for the margin of victory cap and overtime games should all count as one point wins or losses, for starters - but it is at least an attempt to analyze things a bit deeper.
Most importantly, this is the metric that the NCAA chose. Unfortunately, as in football, they insist on adding human subjectivity into the equation rather than allowing data to be the guide.
It would be logical to simply take the highest NET ranked teams who don’t receive autobids for the at-large NCAA tournament slots. Being that this is the NCAA where logic is consistently ignored, they instead decided to keep the quadrant system and marry it to the NET. This tipped the scales almost completely back in favor of pure wins and losses, and created some strange scenarios. Alabama is one of those cases this season.
Coming into last night’s critical win over LSU, the Tide was #39 according to the NET rating (UPDATE: Alabama moved up to #35 after the LSU win) but nowhere to be found on Joe Lunardi’s latest tournament projection, which includes the 68 team field plus the next eight who just missed the cut. So, a team that is the 39th best according to the NCAA’s own metric is somehow outside the top 76 for the NCAA tournament thanks to the quadrant system and autobids. Meanwhile, #48 Cincinnati, #49 USC, #55 Virginia and #58 Indiana are all in the field as at large selections.
Northern Iowa, which is projected to make the field as an autobid barring a conference tournament upset, a terrible way to determine autobids by the way but a topic for a different day, was four spots ahead of Alabama on the NET at #35 and logically projected as a 9-seed in the tournament. The three teams between UNI and Alabama are all in the field, seeded anywhere from 7 to 10. In fact, Alabama is the highest ranked team that isn’t projected to make the field, yet they aren’t even on the bubble at this time.
Folks, that is patently absurd, not just for Alabama but for several teams in between.
So, why does this kind of thing happen?
Thank the broken quadrant system that puts far too much weight on the game location.
By taking Q1 all the way down to #75 on the road and Q2 down to #135, some ridiculous equivalencies are created. Alabama’s win over LSU is currently a Q1, but if LSU drops a mere four slots then it will be given equal Q2 merit to a win at the College of Charleston or Pepperdine. Even as it stands, a win at Furman or Akron counts as a Q1 win and would thus be an equal resume bullet to a home win over LSU. You’d be kidding yourself not to acknowledge that this was done strategically, to allow teams from smaller conferences to get cheap Q1 and Q2 wins, and grab a few at large berths as a result.
And remember, it doesn’t matter if a win is a laugher or a triple overtime affair. A team can beat a top 30 team by 20 points and the quadrants consider it equal to a one point win at the #75 ranked team. The announcers in last night’s game mentioned that “five or six possessions” have kept Alabama from having three or four more quality wins and being firmly in the field. Would the Tide be a better team if one more shot had fallen for them or not fallen for the opponent in those games? I think not.
Add in the subjectivity and bias that comes with human involvement in the selection process, and you will usually end up with something other than the very best field of teams.
March Madness has always been fun in large part because of the upsets by directional schools, but at this point the NCAA has fallen in love with the Cinderella story and is thus watering down the product. Autobids in a tournament of 68 teams are fine, and as of today at least 18 teams who fall outside the top 68 will receive those. That is plenty of representation for the “little guys” and others have the opportunity to get there on merit.
For those who claim otherwise, San Diego State plays in the Mountain West Conference but has dominated its 2019-20 schedule, winning all 24 games including 20 by double digits, to rightfully earn the current #1 NET ranking. Winning games convincingly should matter, and in the NET ranking it does. The West Coast Conference, anchored by a Gonzaga team that nobody has considered a midmajor for some time, has three teams that have played their way into the top 37. A team that thoroughly dominates a weaker slate will be rewarded by the NET ranking and have a great shot to crack an at large bid in case of a conference tournament upset.
Tweak the formula if necessary, but scrap the committee and the quadrants, and award the at large spots to the highest NET ranked teams. This rewards those that have performed the best consistently relative to schedule strength, and not just at the end of a few close games. The tournament will be better for it.