Results 1 to 6 of 6
  1. #1
    scottdude8's Avatar
    scottdude8 is online now Moderator, Contributor, Zoubek disciple, and resident Wolverine
    Join Date
    Jan 2010
    Location
    Storrs, CT

    NET News (February 10): How did the NET influence the mid-season Top 16 seeds?

    Hi everyone! My apologies for missing my weekly NET News post last week (I'm hoping maybe some of you missed it too? Haha), but I was off on vacation and enjoying thinking about nothing on the beach, including my over-analysis of the NET.

    Anyways, considering the big news from a NET/Bracketology perspective this weekend was the release of the mid-season Top 16 seeds (story here), I was curious as to how the mock-committee weighed two factors: a team's NET ranking and a team's Team Sheet/Resume (of which NET rankings obviously play a big role). I was also curious how things compared to the old RPI system to see where the NET might have affected the mock-seedings in a way that made sense or might have been a step backwards.

    Obviously, the seedings were released before the outcomes of yesterday's games, so the NET rankings I'm going to refer too (here) aren't the exact same as what the mock-committee used (although the only major result affecting the top of the bracket yesterday was our big win, so they should be quite similar). I'm going to focus on a couple of key points in detail rather than do a cursory analysis of everything, but I'm sure the discussion will naturally go where it needs to go. With that in mind, here were the major takeaways for me from the NET's influence on the mid-season seeds:

    1) When it came to the very top seeds, it appears NET ranking may have trumped strength of resume (in terms of last year's "quadrant" system). The relative ranking of Gonzaga, Kentucky, and Michigan are what I'm specifically referring to here... I think most people agree that, at this moment, the top 3 teams have started to differentiate themselves from the rest of the pack (although with a ton of big games coming up that could change quickly!). If we go back to my last thread on the "blind resumes", the consensus seemed to be that Michigan had the best claim to a No. 1 seed amongst these three teams (and arguably Michigan had the best win amongst them yesterday, beating NET No. 11 Wisconsin). However, Gonzaga claimed the fourth No. 1 seed from the mock-committee, and Kentucky beat out Michigan for the top No. 2 seed. This makes sense if you look purely at the NET rankings, where Gonzaga is now No. 1, Kentucky is No. 5, and Michigan is No. 6. However, if you go deeper and look at the team sheets, I think Michigan has the best overall resume amongst these three teams.

    Gonzaga is 4-2 in Q1 games, but importantly only one of those wins is a "top-half" Q1 win (obviously, it is against us). The rest of their Q1 wins are a tight home win against Washington and two away victories against inferior WCC opponents. Kentucky is 7-2 in Q1 games, but again three of those wins are "bottom-half" Q1 wins (I'm emphasizing this because it is now emphasized on the team sheets). Additionally, not only do the Wildcats have more losses than Gonzaga and Michigan, they arguably have the two worst losses amongst the group (at Alabama and a Q2 neutral site loss to Seton Hall). Michigan, meanwhile, is 6-2 in Q1, with both losses coming in the top-half of Q1 (at Wisconsin and at Iowa), and four top-half Q1 wins that are amongst the best in the country (home against UNC, Purdue and Wisconsin and at Villanova).

    All of that is to say the following: if you don't look at the NET ranking and look purely at resume, I don't think anyone would put Gonzaga above Michigan based upon their relative quality of victories. I also think that Michigan would be ahead of Kentucky considering they have similar quality victories, but Kentucky not only has an additional loss, but two significantly worse losses than Michigan. Add into the equation the fact the B1G is considered a superior conference to the WCC and SEC leads me to the conclusion that this ordering (Gonzaga, Kentucky, then Michigan) was influenced heavily by the NET rankings. Obviously the debate can be had whether the analytics/NET are more important than resume (I lean towards the later, obviously), but this data point seems to be evidence that the committee may be favoring the former to differentiate the top-tier.

    FWIW, in the RPI (here) Michigan is No. 4, Gonzaga No. 5, and Kentucky No. 7. The RPI gets WAY more wrong than it does right (more on that to follow) but the fact that IMHO it has the relative ordering of those three teams correct is fairly interesting.

    2) The NET seems to have improved upon some significant outliers in the RPI, and that was reflected in the mock-seeding. In the RPI, Kansas (somehow?!?!) is still No. 1 and Houston is No. 3. Meanwhile, Tennessee is way down at No. 9. Given our common sense and intuition on these teams, the NET seems to rank them in a more reasonable fashion (Tennessee at No. 4, Houston at No. 7, and Kansas at No. 18). The seeding seems to reflect the NET more than the RPI, with Tennessee earning a No. 1 seed and Houston and Kansas earning 3 seeds (as opposed to 1 or 2 seeds if you went solely by the RPI). The argument could be made that 3 seeds are a touch high for Houston and Kansas if you dig deeper into the resumes (Houston because of their strength of schedule, Kansas because of how many losses they've had), but in this area the NET seems to be at the very least an improvement from the RPI, and it shows in the seeding.

    3) Marquette is the one team that seemed to get a significant bump from their NET ranking, potentially based on resume. Marquette is still No. 21 in the NET rankings following yesterday's win over Villanova, but earned a 3 seed in the mock-seeding. Digging a little deeper, this is likely due to their Q1 strength... Marquette has 8 Q1 wins, including impressive non-conference victories over Wisconsin and Louisville, and three of their four losses are of the Q1 variety. So while NET ranking seemed to trump resume at the very top of the bracket, it appears resume and the "quadrant" system may be doing what it should (i.e. balancing out apparent outliers in the NET rankings) as you go down the line.

    Obviously, a lot of things here will end up resolving themselves (Kentucky and Tennessee still have to play at least twice, Michigan's resume has a lot more opportunities to improve than Gonzaga's, etc.), but I think that we've gained some valuable insight into how the committee views the new tool provided by the NET via these mid-season seedings, and seen a couple of clear instances where this tool is an improvement over the RPI. I'm very curious to hear the board's thoughts on the three specific points I highlighted above, as well as anything else they've found interesting in the latest NET rankings and mock-seedings!
    Scott Rich on the front page

    Trinity BS 2012; University of Michigan PhD 2018
    Duke Chronicle, Sports Online Editor: 2010-2012
    K-Ville Blue Tenting 2009-2012

    Unofficial Brian Zoubek Biographer
    If you have questions about Michigan Basketball/Football, I'm your man!

  2. #2
    Join Date
    Nov 2007
    Location
    Vermont
    Kansas still at #1 in the RPI is quite something...good reason it's fallen into disfavor.

  3. #3
    Thank you for the time and energy you expended putting this together.

    As long as Duke stays #1 overall I'm good with it.

  4. #4
    Join Date
    Feb 2007
    Location
    Hot'Lanta... home of the Falcons!
    Wait, the RPI is still a thing?

    Does it get included on the team sheets provided to the Selection Committee? I think the team sheets now only have NET, KenPom, ESPN BPI, ESPN SOR, and Sagarin. I think the RPI is not even in the room when the seeds/selections are made.
    Why are you wasting time here when you could be wasting it by listening to the latest episode of the DBR Podcast?

  5. #5
    Quote Originally Posted by JasonEvans View Post
    Wait, the RPI is still a thing?

    Does it get included on the team sheets provided to the Selection Committee? I think the team sheets now only have NET, KenPom, ESPN BPI, ESPN SOR, and Sagarin. I think the RPI is not even in the room when the seeds/selections are made.
    The RPI is not used, people just bring it up now and then to compare it to the new NET to seem if the move to the NET has been an improvement, and it clearly has.

  6. #6
    scottdude8's Avatar
    scottdude8 is online now Moderator, Contributor, Zoubek disciple, and resident Wolverine
    Join Date
    Jan 2010
    Location
    Storrs, CT
    Quote Originally Posted by JasonEvans View Post
    Wait, the RPI is still a thing?

    Does it get included on the team sheets provided to the Selection Committee? I think the team sheets now only have NET, KenPom, ESPN BPI, ESPN SOR, and Sagarin. I think the RPI is not even in the room when the seeds/selections are made.
    Correct, the RPI is no longer on the Team Sheets. I only included it in the discussion to highlight the differences between the NET and the RPI and show how, generally speaking, the NET is an improvement (albeit an imperfect one!) over the RPI. Since the RPI formula was easily replicable there are websites who are still keeping track of it just for kicks. Sorry if that wasn't clear!
    Scott Rich on the front page

    Trinity BS 2012; University of Michigan PhD 2018
    Duke Chronicle, Sports Online Editor: 2010-2012
    K-Ville Blue Tenting 2009-2012

    Unofficial Brian Zoubek Biographer
    If you have questions about Michigan Basketball/Football, I'm your man!

Similar Threads

  1. Replies: 100
    Last Post: 02-11-2019, 05:30 PM
  2. ACC late off-season news
    By Olympic Fan in forum Elizabeth King Forum
    Replies: 24
    Last Post: 08-30-2017, 11:53 AM
  3. Duke Women's Basketball off-season news
    By DU82 in forum Elizabeth King Forum
    Replies: 0
    Last Post: 04-20-2017, 08:31 PM
  4. Bad news for Duke baseball-Marconcini to miss season
    By jimsumner in forum Elizabeth King Forum
    Replies: 1
    Last Post: 02-15-2012, 08:00 PM
  5. UNC News...Ginyard and Graves out for season
    By roywhite in forum Elizabeth King Forum
    Replies: 56
    Last Post: 02-04-2009, 07:39 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •