Thursday, March 14, 2019

Yes, Quad 1a IS Real and No, NC State isn't a Tourney Team Today

The NC State/Clemson game yesterday left fans of both teams a little salty.  Clemson fans saw their team blow an 18 point lead and lose on a couple of last second free throws on a foul call that Tiger fans weren't particularly thrilled about.  NC State fans were feeling pretty good about their Wolfpack for a couple minutes until Joe Lunardi took to the airwaves and basically said that despite the win, NC State didn't look like a tournament team to him, and that they would be very lucky to make it in when all is said and done.  Seth Davis tweeted not long after that he felt NC State still needed to win two more games to earn an At-Large bid.  NC State fans thought they just earned a tournament bid ... but some Analyst guys started crapping on that parade.  I can't say I blame them for being a bit defensive.

Now, I'm no fan of Lunardi, but in this case I agree with him, and at least partially with Davis.   I don't have NC State in even with yesterday's win.  Should they upset Virginia today it likely would be enough to move them in, but it would be close, particularly if teams ahead of them also win and the bubble shrinks due to bid thieves stealing some of the current at-large bids being competed for by the likes of NC State.  They very well may need to win more than just today's game.  Time will tell on that, and frankly, its not worth worrying about until they actually beat Virginia.  If they do, it will be the first team in the Quad 1a grouping they have beaten all year.  I stated my agreement publicly and gave my reasoning.  A big part of that argument is pinned to that winless record against the Quad 1a grouping despite 8 opportunities to date to win one.  A number of people challenged me that such a grouping even exists.

Many of you may be thinking the same thing.  No one on TV is talking about Quad 1a.  They never show that record on TV when showing team resumes.  It sounds like something someone made up, and it was suggested I did just that.  But I assure you the grouping is real, it is data that is being provided every day to the NCAA Selection Committee by the NCAA itself, and most importantly, it was data specifically requested by the Selection Committee itself last summer.  No one on TV may be talking about it, but you can bet it is in the Selection meetings, which began yesterday.

What is Quad 1a?

Most who follow college basketball have become familiar with Quad records.  The NCAA started using them last year to help group each team's wins an losses into common 'levels of difficulty' for lack of a better term.  The better the team, the better of the quality of victory.  Quality was determined by the opponent, and where you played them.  Last season the quality of the team was determined by the Ratings Percentage Index, or RPI.  The RPI has been scrapped and replaced by a new ranking system called the NCAA Evaluation Tool, or NET.    NET rankings are used to define the Quads.  Here is how Quads are set up:


A game played against a NET ranked team ranked 70th at their gym, would count as a Quad 1 win.  However, beat that same team on your home floor or at a Neutral Site, and it only counts as a Quad 2 win.  Not all wins against the same team are created equal in most cases.  Its all about where you play them.

As a macro tool for getting a high level view of how a team performed against various qualities of competition it serves the purpose.  It aids the selection committee in doing an initial scrub of all the teams to try and identify those that have fared better or worse than other teams.  It's really pretty simple.  If I'm looking at two teams who both have 20-10 records, how do I sort who has performed better than the other.  A good start would be to look at their records against the best competition.  A team that had 8 Quad 1 wins among the 20 total they have would initially be sorted higher than a 20 win team with 1 Quad win.

The problem comes in when those two teams both have 8 Quad 1 wins.  Are all Quad 1 wins created equal?  The logical answer is, of course not, unless of course you believe that beating Virginia, ranked 1 is truly as impressive as beating Penn State (49) on a Neutral Court or Liberty (60) on their home floor.  The NCAA, in response to this quandary, started providing a break down of wins in the Quad 1 and Quad 2 groupings into 2 tiers per Quad, commonly referred to as Quad 1a, Quad 1b, Quad 2a and Quad 2b.  The data is designed to help the committee see which Quad 1  or Quad 2 games were against the top half of the Quad or bottom half of that Quad.  Here is how those groupings are further broken down within their respective Quads.

Using this further slicing of a data, a Game won against a team ranked 24th would be a Quad 1b win, but a Quad 1a win on a nuetral or road court.  They are both Quad 1 wins at the Macro level, but the win on a nuetral or road court is ranked higher.  Back to our example of the two teams with 20 wins, and 8 Quad 1 wins each.  If you break them down further into Quad 1a and 1b games and find that of the 8 Quad 1 wins, Team A has, 6 are in the 1a grouping and Team B has just 1 in that same grouping, the committee members likely would conclude that Team A should be ranked higher than B for selection purposes.  Obviously this is a simplistic view, and there are other factors they consider, but it does seem rather obvious to me that this kind of a break down helps the committee do better side by side comparisons of two teams who may otherwise look similar on paper looking at only the Quad 1 totals.

Some may be wondering where the committee gets this data from  ... or even if they truly get it, since none of the talking heads on TV reference it.  The data is provided by the NCAA on what are known as team sheets.  These are used heavily by the committee.  It provides a summarized view of each team, and includes a bunch of information about each team including summarized Quad1 thorugh Quad 4 records.  Those summaries don't include the records for the a and b tiers in Quad 1 or 2.  However, they do list, by Quad, each team that they have played, the NET ranking of the opponent, where the game was played and the outcome.  It is here that the data is provided.  Let's take a look at NC State's team sheet as of this morning (3/14):

Looking at the Quad 1 games column you will see the games broken into two groups.  The criteria for the groups in Quad 1 are circled in red and highlighted in yellow.  For Quad 2, circled in blue and highlighted.  As you can see, the data is clearly provided to give a view to the reader of how the team performed in the top half and bottom half of each Quad.  In this case, NC State is 0-8 against Quad 1a, and 3-0 vs. Quad 1b.  They are 3-0 in each half of the Quad 2 tiers.  

So, how do we know this data is important and not just window dressing?  Because the committee specifically requested it be included this season.  This was not data that was provided last year.  Clearly the committee felt it would be useful and was needed in their process of sorting, selecting and seeding teams.  If just looking at the Quad 1 or 2 records provided all they needed last year for those purposes, there would have been no need to request it.  The only conclusion I can reach is that they asked for it because they intend to use it.  It's overall weight is anyone's guess at this point.  But to pretend it's not there at all is to ignore the obvious.  To pretend people would like at Quad 1 records, but not at what games made up that record when the data is specifically presented to show it, defies logic. 

And how do I know that the committee specifically asked for it?  Because I was told so by David Warlock, the NCAA Director of Media Coordination/Statistics and Media Coordinator for March Madness.  He sits in the room as the committee goes through the selection process and started his 14th year of doing it yesterday.

So yes, there really is Quad1a, 1b, 2a, and 2b data provided to the selection committee by the NCAA.  It was provided at their request.  And frankly, if you don't think they intend to use it, I think you're foolish (or engaging in wishful thinking).

What does this have to do with NC State?

We now return to what prompted this whole discussion in the first place.  Is NC State currently a team that can be expected to be in the NCAA Tournament as an At-Large or not?  Lunardi says yes for now ... Seth Davis says not unless they win a couple more ... and I say no, and I doubt they will get there unless they beat Virginia today ... and then probably, but its not a lock.

Let's do a compare of 4 Teams generally considered to be very close to the bubble cut line by most ... NC State, Ohio State, TCU and Georgetown.  I choose them because they come from 4 different conferences and have easy to compare profiles.  This provides an opportunity to see how the Quad data and NET data, and specifically with a and b data sliced up within it, would or could be used to evaluate the 4 teams.
If you were to look at NET rating and overall record alone, you would likely be inclined to rank them in order left to right.  A deeper dive into the data, I would argue, tells a very different story.   

I tend to think these are the metrics the committee will likely look at first for sorting and selecting.  By my analysis, NC State stacks up worst in almost every measurment compared to the other three.

  • Quad 1 - Worst (Tied)
  • Quad 1a - Worst
  • Quad 1a Road/Neutral - Worst
  • Quad 1-2a - 2nd
  • Quad 1-2a Road/Neutral - Worst
  • Best Win - I would argue worst.  Better ranking but at home. I agree its debatable
  • Best Road/Neutral Win - Worst
  • Quad 3-4 Record - 3rd
  • Worst over all loss of each team, and the worst home loss.
  • Compared to the teams currently in the field or on the bubble, they have the worst record.

Reasonable people can disagree if these metrics are key, and if the conclusion I have reached from evaluating them is correct.  In my book, its not close.  They just aren't an NCAA Tournament team this year.  Historically we know this much, and its been consistent over the years;  
  • The committee puts a high value on Road wins against the best competition.
  • The committee penalizes marginal bubble teams that have played awful non-conference schedules.  It happens every year.  There isn't a worse non-conference schedule in all of the NCAA then the one the Wolfpack played.  As bad as Georgetown's is ... NC State's is over a hundred spots worse.  
  • A well known record the committee looks at in its final bubble scrub is the record of the team versus those in or on the bubble.  I agree its subjective, and fluid, especially right now, so its not a major factor yet.  We also don't know who the Committee considers a bubble team, and my bubble may be different than theirs.  As teams win Tournaments and teams come into or fall out of the field it can change by selection Sunday.  But today, I believe they have a miserable record against those teams.  Even if I give them credit for Clemson (I believe they are no longer under consideration so I didn't count them in the record above) the 4-8 record becomes only a fraction better than OSU.

For the record, the way I have these teams ranked relative to one another is:
  1. Georgetown (IN) - 
  2. TCU (OUT)
  3. Ohio State (OUT)
  4. NC State (OUT)
The only thing that could potentially keep NC State in the discussion if they lose today, in my opinion, is their NET rating.  Its's possible the committee could ignore the actual results of the games every team has played and value the NET rating above the actual results of games.  If that happens, all bets are off.  But the NCAA has stated that the NET Ranking is sorting tool.  The Quads are a selection tool.  That statement alone tells me that game outcomes will trump a rating, as it should. 

No comments:

Post a Comment