Welcome to the Lunatic Fringe

It looks like you're new here. If you want to get involved, click one of these buttons!

Want to help support PFN?

In lieu of ads you can help me to offset PFN operating costs by making any of your Amazon.com purchases through this link at no additional cost to you.

Bubble Analysis - 3/10

Vawolf82Vawolf82 Posts: 248PFN Referee
edited March 17 in Wolfpack Sports
INTRODUCTION

When I first started studying the Bubble at SFN, I never wanted to go as far as predicting the field.   I wanted to be able to pick those ACC teams that were clearly in and clearly out and point out the ones that fell in the grey zone between those two extremes.   Overall, I think that the approach used in previous years has served its purpose.   For the most part, I was not really surprised by the Selection Committee's decision on the ACC teams with the exception of the pleasant surprise when State (with TJ Warren) got in the First Four when I had given up all hope.

I changed the approach for this year because of the large number of posts over the Interwebs predicting an easy path to the NCAAT with only one Quadrant 1 win and picking up victories against the bottom six in the ACC.   (This new approach became practical when the NCAA made the data available for download into Excel.)  I think that looking at the whole field made it pretty obvious that State was not secure, even though what I considered obvious wasn't always widely accepted.  

Overall, I like what this new approach has shown.   I hope to be able to take the Selection Committee's final decisions and compare to the various groups that I've broken down the field into and see if there is any new information for us to use in future years.   However, I think it is important to once again stress the simplifying assumption that I have made by ignoring the fact that the team sheets split Q1 and Q2 wins into upper and lower halves.   That distinction is not available in the download data and I'm not about to spend the time going through the team sheets manually.   So once again, we'll see what we see when the Selection Committee finally speaks.

Now onto the data....

ASSUMED CONFERENCE CHAMPS

Take NET rankings from 1-80 and sort by conference.   Assume that the highest rated team in each conference gets the automatic bid and then remove these teams from further analysis.   So here are the assumed conference champs:


If any of these teams has to go into review for an at-large bid, I've highlighted some areas of concern.   One of the most interesting hypotheticals to examine is Nevada.  The other Wolf Pack has not played a particularly challenging schedule and has surprisingly few quality wins despite an impressive NET ranking.   If this scenario comes to pass, it will definitely be interesting to see what the Selection Committee and the talking heads have to say.

EASY AT-LARGE BIDS

I toughened up the standard for this week and took teams with 4+ Q1 wins without particularly disabling deficiencies. 


These selections obviously have some controversy attached.

I tracked at-large bids versus RPI ranking for a number of years at SFN.   There was a very clear drop in frequency at rankings >55.   For the years that I looked, the number selected at 55+ ranged from zero to three.  One of the things that I'm interested in is seeing where the bubble range for NET rankings shakes out to be.   In any event, it is probably unlikely that all five teams at the bottom of this list get at-large bids.  But I left them in because what we're really interested is seeing where State matches up to the field....and right now it is easy to build an argument that these five teams have better resumes than State.  

Lunardi has said that over the last five years, eight teams were selected with losing conference records.   That fact leads me to believe that conference record is somewhat important, but a losing conference record is not an automatic disqualification.   With OU and OSU playing in the two toughest conferences in the nation, I suspect that they will get some slack from the Selection Committee.

Indiana will likely be a controversial decision no matter how the Selection Committee votes.   I really think that Indiana needs two more wins in the B10 tourney, but it's hard to be sure.   Maybe IU and Xavier will give us some data on the minimum overall winning percentage that the Selection Committee finds acceptable.

St Johns and Georgetown have frighteningly similar resumes.   I really don't see how the Selection Committee could only take one of those two.


TOO MANY LOSSES

I just don't see the Selection Committee taking teams with ~.500 overall record no matter how many good wins that they have.   So I'm removing these from further consideration:


TOO FEW QUALITY WINS

The Selection Committee has always had the standard that you have to beat good teams to prove that you deserve an at-large bid.   These teams fall short of that standard:


LOSING CONFERENCE RECORD



I'm going to add TCU (toughest conf) and throw Providence back.


THE GREY ZONE

So my guess is that there is room for somewhere between 7 and 10 teams from those that didn't fit into any other category



The 8/9 game in the ACCT will result in a Q1 win for someone.  I feel like that would be enough to get State into the field, but is probably not enough for Clemson.   Even with a loss, State would have some chance...but probably not a very good one.


FINAL THOUGHTS

According to realtimerpi, State's RPI ranking would be 101 if that metric were still being used.    This RPI ranking would never even be in consideration for an at-large bid.   So as I've pointed out before, Keatts is extremely lucky to have a new system rolled out at the exact same time he decides to schedule teams from various rec leagues.  It remains to be seen how State's ridiculous OOC schedule (ranked last in Division 1) affects State's post-season tournament.

On the other side of the old evaluation techniques, I would feel reasonably confident with four Top 50 wins (with one at neutral location) along with an RPI of <65.   But I'm not really confident that anyone outside the Selection Committee know how the wins in Q1 and Q2 (along with the splits in each category) are going to be used.

So we'll wait and watch the various conference tourneys and then wait with baited breath for the unveiling of the bracket on Selection Sunday.   I've been pressed for time this week so please feel free to add dissenting comments along with the reasoning behind your conclusion.
«134567

Comments

  • WulfpackWulfpack Posts: 929
    I’ll be watching this week for bid stealers. UNCG and Wofford play tonight in the SoCon Finals. Some think Wofford could sneak in even with a loss. 
  • TexpackTexpack Posts: 1,254
    The committee is going to hear it from teams with a NET in the 30's if there are teams with a NET in the 60's & 70's that get into the field ahead of them.  The NET is after all, the creation of the committee with input from the coaches.  I just can't fathom Texas, or TCU getting an at large bid without making some serious noise in the Big 12(-2) Tourney.  

    Seton Hall with a NET of 61, I just don't see it.  Our OOC schedule makes as an awesome target for exclusion.  I think our chances if we lose to Klempson are infinitesimal.  If we beat the Tiggers I think it will be very very hard to keep us out.
  • choppack1choppack1 Posts: 575
    Not sure how those big east teams (st johns, Xavier and Georgetown) all are considered easy in with their bad NET and mediocre conference record.

    The Big East doesn’t have any great teams, an absence of terrible ones. It’s easy to have a healthy record vs quad 1 when you aren’t playing great quad 1 teams. You need look no farther than our record vs Syracuse, Clemson, Auburn, Penn State, Louisville and Wisconsin to see how you can chalk up Ws vs the these “good, not great teams.”

    Obviously, our game vs Clemson is an elimination game. The big question is whether or not we need to beat UVA.
  • Vawolf82Vawolf82 Posts: 248PFN Referee
    Texpack said:
    The committee is going to hear it from teams with a NET in the 30's if there are teams with a NET in the 60's & 70's that get into the field ahead of them. 
    It happened nearly every year with the RPI rankings.   The difference is that it could happen to teams from power conferences now with the NET + quadrants.
  • Vawolf82Vawolf82 Posts: 248PFN Referee
    edited March 11
    choppack1 said:
    Not sure how those big east teams (st johns, Xavier and Georgetown) all are considered easy in with their bad NET and mediocre conference record.

    What's life without a little controversy?  :-)
    Several thoughts or maybe open questions:

    Which is more important, quality wins or NET ranking?

    How committed is the Selection Committee to the concept of quadrant wins?   Will they really treat a win over #75 on the road (a team that might not even make the NIT) as a better win than beating #31 at home (a team that will  likely play in the NCAAT).

    How do you treat the combination of Q1 + Q2 wins?   In other words, can we say that a Q2 is worth half as much as a Q1 win?   Third as much?  etc

    I've went all in on Q1 wins to the exclusion of NET rankings.   I've never been impressed with Pomeroy since I saw that beating up midgets by a bunch inflates your ranking (thanks to Pomeroy consistently over rating Wichita St in the past).   Since the NET, Pomeroy, and Sagarin are all very similar, you can conclude that there is some level of statistical accuracy to this assumption.   i just don't agree with the results in some cases...specifically how they over rank teams with easier schedules.

    The other error that I've likely made is basically ignoring Q2 wins since I don't know how to evaluate them.   Specifically the Q2 wins in the upper half of the quadrant will include a bunch of NCAA teams.   The Selection Committee has often talked about wins versus NCAA teams....which is a metric we can't use because we don't know who made the field.

    The other error that I've made is not looking at the team sheets at all.   Anyone that does will likely be able to separate bubble teams with identical Q1 and Q2 records, where I would consider them essentially equal - like St Johns and Georgetown above.   I just don't have several days to study each team's detailed resume.
  • Vawolf82Vawolf82 Posts: 248PFN Referee
    edited March 11
    In the past, I saw a number of examples of bubble teams where their Top 50 wins was a better predictor than Top 25 wins.   I wonder if we'll see the same thing with Q1 vs Q2?  If splitting the Q1 and Q2 wins up on the team sheets plays a big role in selecting the bubble teams, then this simplistic approach of sorting teams in a spreadsheet won't be worth the effort.
  • Tau837Tau837 Posts: 57
    VaWolf, thanks for continuing these posts. I appreciate all of the work you put into them.

    IMO the top 23 in your Easy group are easy, but the bottom 5 aren't, which is what led you to highlight them. I understand that they fit the criteria that you used to create the Easy group, but IMO those bottom 5 belong in the Grey Zone group. So I see 23 at large bids from that group at that step.

    I'm unclear on why you treated TCU and Providence differently from the Grey Zone group, since 2 teams in your Grey Zone group have losing conference records. That said, I agree that TCU makes it and Providence doesn't.

    So to me the Grey Zone has 19 teams, the 14 you show here plus the bottom 5 from your Easy group. And there are 12 bids for those 19 teams.

    If selections were made today, IMO NC State would get one of those 12 bids. They are not a bottom 7 team within that group.

    I think if State beats Clemson, they are in. If they lose to Clemson, I don't think it is a given they are out, though. IMO it would depend on what happens with the rest of the bubble. This system seems stupid enough that it also actually matters if Penn State hangs on as a Q1 win.
  • Tau837Tau837 Posts: 57
    edited March 11
    Also, having gone through your series of posts this season, I think it has highlighted what a bad system they have designed with the quadrants.

    As you said, should beating #75 on the road be worth more than beating #31 at home? I think the answer is obviously no.

    But even worse is that beating a top 10 team anywhere is treated as equivalent to beating #75 on the road.

    If they want a better system, it should be one without arbitrary cutoffs. But if they are going to use arbitrary cutoffs, they need to tighten up the Q1 cutoffs to something like 20 at home, 25 neutral, 30 road.

    It really should be a system that tries to approximate more closely that Q1 are truly a top group, like top 25, and Q2 is the rest of the teams that represent NCAA bid contenders, like top 75, or something like that. That would make Q1 and Q2 records much more meaningful.

    It feels like the committee is going to end up essentially breaking up Q1 and Q2 in a manner like this, anyway, so they really should tweak the quadrants to track what they will actually use to make bid decisions.
  • Vawolf82Vawolf82 Posts: 248PFN Referee
    Tau837 said:
    I'm unclear on why you treated TCU and Providence differently from the Grey Zone group, since 2 teams in your Grey Zone group have losing conference records. That said, I agree that TCU makes it and Providence doesn't.
    I'm willing to overlook a losing record in the best conference...and the Selection Committee has done so in the past for the B-12 (OU last year).

    On the other hand, Lunardi's stat of selecting 8 teams with losing conference records in the last 5 years, shows that there are limits to what the Committee will overlook.   A losing record in the Big East (that has been reconstituted with teams historically viewed as mid-majors) doesn't earn that slack from me. 
  • choppack1choppack1 Posts: 575
    VaWolf - I noticed that. (And as usual, thanks for assembling a ton of data, that some of us can discuss in detail and be better informed.)

    I really think State is at that crossroad...we don’t have a lot of Q1 wins and our OOC is awful, but BPI, Sagarin, Kenpom and most importantly NET have the Pack as at large bid at this point.

    Conveniently, I side with Tau on the quadrant discussion. We have had limited quad 1 opportunities against the bottom half. Our “easy” Quad 1 game at home was #17 Auburn (our next easiest Quad 1 game was #12 Va Tech.) 

    We had 1 neutral Quad 1 game and we won. 

    Our easiest Quad 1 road game was #22 Louisville. 

    I guess the point is that is spite of us playing 10 Q1 games, most teams in the Top 40 would have a difficulty going 2-8 versus them.

     I don’t think it’s a coincidence that we were 2-0 in the two easiest Q1 games we had. I also don’t think it’s a coincidence that we are undefeated versus teams from 25-50. 

    This Clemson game will be very telling to the committee. I think with a win there, as long as we don’t get humiliated by UVA and there’s no additional carnage, it gives us a Top 30 NET, Kenpom, BPI and Sagarin...I don’t see how you leave a team out when the data shows you should clearly be in.


  • Vawolf82Vawolf82 Posts: 248PFN Referee
    Tau837 said:

    But even worse is that beating a top 10 team anywhere is treated as equivalent to beating #75 on the road.


    Several things on this.

    The situation is still bad, but not quite this bad.   Here is a snip from State's most recent team sheet:


    State's two Q1 wins are both "graded out" into the lower "half" of Q1 (with PSU on the verge of falling into Q2).   So on the surface, beating #1 on the road is exactly the same as beating #40 on the road.   

    However, we know that the Selection Committee spends a lot of time studying these team sheets and I have to believe that beating a Top Q1 team on the road is viewed very favorably in both seeding and Bubble selection.

    Being 6-0 in Q2 has to help State.   Having three Top-Half wins in Q2 (four if PSU falls) also has to help.    The open questions are How much? and Is it enough?
  • tvp1tvp1 Posts: 3
    Here's a way of thinking through the different quadrant rankings I came up with today, just as a way to compare the bubble teams with different types of resumes.  What if we assign points to the different quadrant wins and losses:  A Q1 win is worth 2 points, a Q2 win is worth 1 point, 0 points for Q3 or Q4 wins, a Q1 loss is worth -0.5, a Q2 loss is -1, a Q3 loss -2, and a Q4 loss -3. 

    Using that system, out of VA's gray zone list these teams have a negative score:  Florida, Clemson, Alabama, and Oregon.  TCU also has a negative score.  (State is +2)

    Of course this metric has its flaws - mainly, that it treats all Q1 wins the same and thus probably overvalues the Big East teams who have a lot of lower-half Q1 wins.  But it's not a bad way to sort through the bubble teams in my opinion.  

    Remove those 5 and we only need to take out 2 more to have our field, assuming no more bid thieves.  Here are some teams with serious blemishes:

    Texas - would be the first ever 16 loss at-large, and also would be fewer than 4 wins over .500 (I think only 1 such team has ever made it)

    Xavier - NET over 70, fewer than 4 wins over .500, 6 Q2-Q4 losses, generally unimpressive resume

    Georgetown - NET over 75.  If beating Georgetown isn't considered a Q1 win, even on the road, how do you put them in the field?

    Arizona State - poor NET, 4 Q3 and Q4 losses, 6 Q2-Q4 losses, not a ton of great wins to balance that out.

    Indiana - fewer than 4 wins over .500, only 11-14 against Q1-Q3 (which Jerry Palm has identified as a red flag historically).

    NC State - Awful OOC schedule.

    St. John's - similar to Arizona State; no Q4 losses but they were blown out in a lot of their Q2 and Q3 losses.

    Ohio State - 8-12 conference record, only 18-14 overall, NET over 50, in a free fall right now with a star player out

    Syracuse - Take out the win over a shorthanded Duke team and their resume really isn't that great.




  • tvp1tvp1 Posts: 3
    My best guess as of today on who misses out from this list (again assuming no bid thieves):

    -NC State/Clemson loser is out
    -Indiana/Ohio State loser is out
    -These 4 teams probably need a run to their conference tournament finals to make it, and even that may not be enough:  Texas, Xavier, Alabama, Oregon
    -And then at least 1 of the other teams will flame out early in their conference tournament and miss out:  Georgetown, St. John's, Creighton, Florida, Arizona State, TCU, Temple, Utah State.  If none do, then Belmont could be in trouble, or Clemson if they beat NCSU but lose to UVA, or even Syracuse if they were to lose their opening ACC tournament game.  

    Of course, more teams off of this list will be in jeopardy if any of Wofford, Buffalo, VCU, Washington, or Nevada fail to win their conference's automatic bids.
  • TexpackTexpack Posts: 1,254
    tOSU and Shaka are both bringing back their suspended players for their play in games. Will the Pack follow suit or wait for the Final Four?
  • WulfpackWulfpack Posts: 929
    Wofford evidently is in solidly currently as a 7 seed. They are down four at the break to UNCG. SoCon could be a two bid league. 
  • choppack1choppack1 Posts: 575
    edited March 12
    Watching wofford play tonight is like watching us since January...fall behind and can’t ever get ahead.

    In case you are wondering, their best victory (outside of conference) is at South Carolina. They have OOC losses to Kansas, UnC-Ch, Oklahoma and Mississippi State. 
  • TheAliasTrollTheAliasTroll Posts: 1,484PFN Referee
     VA, thanks for putting these together this season! I think these threads add so much  value to this site.
  • WulfpackWulfpack Posts: 929
    Wofford wins. No bid stolen. 
  • TexpackTexpack Posts: 1,254
    Wofford goes on a 16-0 run to put away UNC-G. The tallest midget just got a little shorter. 
  • Vawolf82Vawolf82 Posts: 248PFN Referee
    edited March 12
    I was thinking today that the Selection Committee has two new toys to play with....NET and the quadrants.   I’m really interested in how the quadrants will be used.  But I suspect that NET will be used much like the RPI was:
    - seeding
    - For bubble teams, the ranking of your best wins is more important than your own ranking.

    Ive seen a number of posters at PP argue that SOS is not or shouldn’t be a big component of the Selection process.   There was even an article at ESPN arguing that no one knows if or how the SOS will be used.   Until I see something different, I’ll go with the assumption that it will be used exactly the way it has always been used.   If the Selection Committee wasn’t going to use it, the SOS rankings wouldn’t be on the team sheets.  
  • GsoPackBackerGsoPackBacker Posts: 323
    I would propose that if one likes all of the conjecture about the bubble, seeding, who's in and who's out, this year is a dream come true.

    Nobody can claim there system is a better predictor as there is no real basis to measure that.

    I do think (hope) that in moving to the NET rankings, the NCAA has calculated historical rankings and compared them to the RPI to ensure that the system they are moving to is going to be a better overall predictor of the NCAA field and seeding.

    The big plus with the NET for me is the way the team sheets have upper and lower quadrant records now.  I believe that will be key because it is a closer approximation of strength of record.
  • AdventurooAdventuroo Posts: 1,342
    Texpack said:
    tOSU and Shaka are both bringing back their suspended players for their play in games. Will the Pack follow suit or wait for the Final Four?
    Posted a comment about email exchanges between AD Yow & Dr. Woodson & myself.

    Woodson was apathetic, which is disappointing.  Yow says it is NOT an AD call...which Woodson said it WAS.  I REALLY think that the ADMINISTRATION folks that “advise” the Honor Code Board or whatever are dragging their feet.  I sense, with little other verification, that the GOTT issues have stirred the ghosts of the Jimmy V Witch Hunt Faculty squad and that there is a decidedly jealous or resentful attitude rising from the ashes.

    Lockett may be “OK’ed” tomorrow, but he will be of little value due to lack of conditioning...

    But my meager bet will be on LATER and not sooner “decision”
  • AdventurooAdventuroo Posts: 1,342
    Texpack said:
    tOSU and Shaka are both bringing back their suspended players for their play in games. Will the Pack follow suit or wait for the Final Four?
    Posted a comment about email exchanges between AD Yow & Dr. Woodson & myself.

    Woodson was apathetic, which is disappointing.  Yow says it is NOT an AD call...which Woodson said it WAS.  I REALLY think that the ADMINISTRATION folks that “advise” the Honor Code Board or whatever are dragging their feet.  I sense, with little other verification, that the GOTT issues have stirred the ghosts of the Jimmy V Witch Hunt Faculty squad and that there is a decidedly jealous or resentful attitude rising from the ashes.

    Lockett may be “OK’ed” tomorrow, but he will be of little value due to lack of conditioning...

    But my meager bet will be on LATER and not sooner “decision”
    Eric Lockett reinstated to NC State basketball team
    http://wr.al/1EaSy

    And NO...I cannot fix speeding or parking tickets

    ?️‍♂️??
  • choppack1choppack1 Posts: 575
    Really interesting piece here from Trout on ats....he points out only one team with a similar resume has been left out in a while, 2008 USC Trojans. Good stuff, herein and he does include necessary qualifiers.

    http://adamcwisports.blogspot.com/
  • TheAliasTrollTheAliasTroll Posts: 1,484PFN Referee
    Yeah the only sure thing is that it'll be a total mystery until Selection Sunday.
Sign In or Register to comment.