Results 1 to 11 of 11
  1. #1

    NCAAT Selection Process

    Latest issue in ESPN's "Change the Game" series is the committee's selection process.

    http://espn.go.com/blog/collegebaske...ting-selection

    Among Eamonn Brennan's arguments and suggestions:
    • "Despite its protests to the contrary, the NCAA organizes the committee’s information almost exclusively by RPI."
    • "RPI defenders are increasingly the exception."
    • "If the committee is going to use a statistical construct to organize teams, it should be given the best, most updated statistical construct available."
    • "Replace the RPI with a weighted average of all of the best and most accurate rankings systems in the sport. Include RPI in the formula if you really want to."
    • "The process needs to be more transparent."
    • "The only way we can really understand it is if we’re allowed in the room with the committee members while they’re doing their job."
    • "In the process, you could hammer home to everyone just how difficult and harried the selection process can be. The conflicts with travel, the different exemptions, the logistical tangles. That’s the real benefit: Letting the audience in on the unique challenge of building a 68-team tournament to be ready just a few hours after the season’s final games have been decided."

  2. #2
    Quote Originally Posted by gumbomoop View Post
    Latest issue in ESPN's "Change the Game" series is the committee's selection process.

    http://espn.go.com/blog/collegebaske...ting-selection

    Among Eamonn Brennan's arguments and suggestions:
    • "Despite its protests to the contrary, the NCAA organizes the committee’s information almost exclusively by RPI."
    • "RPI defenders are increasingly the exception."
    • "If the committee is going to use a statistical construct to organize teams, it should be given the best, most updated statistical construct available."
    • "Replace the RPI with a weighted average of all of the best and most accurate rankings systems in the sport. Include RPI in the formula if you really want to."
    • "The process needs to be more transparent."
    • "The only way we can really understand it is if we’re allowed in the room with the committee members while they’re doing their job."
    • "In the process, you could hammer home to everyone just how difficult and harried the selection process can be. The conflicts with travel, the different exemptions, the logistical tangles. That’s the real benefit: Letting the audience in on the unique challenge of building a 68-team tournament to be ready just a few hours after the season’s final games have been decided."
    Yeah, everyone really wants to see sausage being made.

  3. #3
    Quote Originally Posted by Indoor66 View Post
    Yeah, everyone really wants to see sausage being made.
    I hope this isn't a good analogy, but maybe you're right.

    In any case, it's clear that some do at least want to understand the process much better, as sometimes it has produced a few strange seedings, pairings, travel oddities, etc. The NCAA a few years ago began conducting media mock bracket sessions, to give a feel for the complications of the process.

    Also, just to clarify: Brennan's piece discusses 2 distinct, if related, issues, the first of which - over-reliance on RPI - has been a topic of some interest on EK.

  4. #4
    Quote Originally Posted by gumbomoop View Post
    Latest issue in ESPN's "Change the Game" series is the committee's selection process.

    http://espn.go.com/blog/collegebaske...ting-selection

    Among Eamonn Brennan's arguments and suggestions:
    • "Despite its protests to the contrary, the NCAA organizes the committee’s information almost exclusively by RPI."
    • "RPI defenders are increasingly the exception."
    • "If the committee is going to use a statistical construct to organize teams, it should be given the best, most updated statistical construct available."
    • "Replace the RPI with a weighted average of all of the best and most accurate rankings systems in the sport. Include RPI in the formula if you really want to."
    • "The process needs to be more transparent."
    • "The only way we can really understand it is if we’re allowed in the room with the committee members while they’re doing their job."
    • "In the process, you could hammer home to everyone just how difficult and harried the selection process can be. The conflicts with travel, the different exemptions, the logistical tangles. That’s the real benefit: Letting the audience in on the unique challenge of building a 68-team tournament to be ready just a few hours after the season’s final games have been decided."

    While I agree with Brennan that too much reliance on only 1 ranking system is flawed, I feel he undercuts his own argument when he admits "The committee typicaly gets it mostly right, especially in selecting the 37 at-large teams to fill out the last two 68-team tournaments. Valid complaints have been rare."

    So the cynic in me leads me to believe this is ESPN's way of creating an argument/controversy that just isn't really their in order to make themselve seem more cutting-edge(see ESPN'S QB Rating introduced last year to "fix" the "outdated" NFL's QB rating). I think ESPN is so paronoid about becoming irrelevant and old-school (see SI, Sporting News, etc.) that they rely much too heavily on "sports science" as opposed to just good old-fashioned professional reporting.

    Don't be shocked if ESPN introduces a new "college BBall rating" next year to provide "THE" most accurate judging of teams. During the season, Bilas, Digger, Katz, and the rest base all of their analysis off of ESPN's system, setting up the template. Then when the last 4 teams Lunardi has "in" (using ESPN's ranking) are completely different then the selection committee...BOOM!...Controversy initiated.

  5. #5
    Join Date
    Jan 2009
    Location
    Halifax, Nova Scotia
    Quote Originally Posted by BlueDevilBrowns View Post
    While I agree with Brennan that too much reliance on only 1 ranking system is flawed, I feel he undercuts his own argument when he admits "The committee typicaly gets it mostly right, especially in selecting the 37 at-large teams to fill out the last two 68-team tournaments. Valid complaints have been rare."

    So the cynic in me leads me to believe this is ESPN's way of creating an argument/controversy that just isn't really their in order to make themselve seem more cutting-edge(see ESPN'S QB Rating introduced last year to "fix" the "outdated" NFL's QB rating). I think ESPN is so paronoid about becoming irrelevant and old-school (see SI, Sporting News, etc.) that they rely much too heavily on "sports science" as opposed to just good old-fashioned professional reporting.

    Don't be shocked if ESPN introduces a new "college BBall rating" next year to provide "THE" most accurate judging of teams. During the season, Bilas, Digger, Katz, and the rest base all of their analysis off of ESPN's system, setting up the template. Then when the last 4 teams Lunardi has "in" (using ESPN's ranking) are completely different then the selection committee...BOOM!...Controversy initiated.
    Already done. I would have written that as the link like Duvall often does, but I haven't figured out how yet.
    http://espn.go.com/mens-college-bask...rankings-final

  6. #6
    Join Date
    Feb 2007
    Location
    Steamboat Springs, CO

    The Whole Process is Flawed: Redux

    Quote Originally Posted by gumbomoop View Post
    Latest issue in ESPN's "Change the Game" series is the committee's selection process.

    http://espn.go.com/blog/collegebaske...ting-selection

    Among Eamonn Brennan's arguments and suggestions:
    • "Despite its protests to the contrary, the NCAA organizes the committee’s information almost exclusively by RPI."
    • "RPI defenders are increasingly the exception."
    • "If the committee is going to use a statistical construct to organize teams, it should be given the best, most updated statistical construct available."
    • "Replace the RPI with a weighted average of all of the best and most accurate rankings systems in the sport. Include RPI in the formula if you really want to."
    • "The process needs to be more transparent."
    • "The only way we can really understand it is if we’re allowed in the room with the committee members while they’re doing their job."
    • "In the process, you could hammer home to everyone just how difficult and harried the selection process can be. The conflicts with travel, the different exemptions, the logistical tangles. That’s the real benefit: Letting the audience in on the unique challenge of building a 68-team tournament to be ready just a few hours after the season’s final games have been decided."
    I think there are improvements to be made in the statistical system used for NCAA selection. However, as I have said before, there is no statistical system on earth that can correct for HAVING NO GOOD DATA WHATSOEVER.

    Let me explain. Here are some assumptions that make this point:

    1. Many college teams change in capability throughout the course of the season. Usually they improve -- some dramatically.

    2. Accordingly, the record for teams in the six weeks between mid- November and January 1 are not representativie of the capabilities of the teams at tournament time.

    3. Unhappily, almost all interconference games that provide a basis of comparison occur in November and December (a few are in the first week of January). Therefore, it is almost impossible to rank teams correctly BETWEEN conferences.

    4. At the same time, between conference schedules and conference tournaments, there is a huge amount of data of how teams rank WITHIN a conference.

    Therefore, there is no statistic under the sun that can rank conferences when there is no meaningful interleague play, if, as I posited (is this a word?), teams change dramatically between the first one-third of the season and the end.

    [I did look at how teams were ranked at the end of December vs. the final per-tournament ranking. In 2011, once one removed the dominant ten teams, there was virtually no correlation among top 25 rankings among the rest of the major conference teams. In 2012 the degree of upheaval was less, but still significant.]

    Consequently, I would totally upend the applecart. If you turn it over to the right, you leave the schedules unchanged, but you totally change the way teams are selected. For example:

    1. Give each conference a given number of teams to make the conference.
    a. I would give the major conferences one-half of their total membership, invoking the rule that if you aren't in the top one-half, you don't deserve to be in the NCAA's.

    b. WRT to other conferences, I would let past performance affect the number of teams selected. Use a relegation system to give conferences more or fewer selections.)

    2. Let the goldarned conferences decide who represents them.

    3. I would probably use a formula for seeding teams, although this is a job that can be given to the TSC -- especially for the top 8-12 seeds.

    Alternatively, turning the applecart over to the left, I would change the way college basketball games are scheduled.

    1. Set aside two weeks in February for inter-conference matches (not just Big ten vs. ACC). 3-4 interconference games per team would provide 120 to 160 games results to help in ranking teams across conferences.

    2. Start the conference schedules two weeks earlier.

    3. Use statistical models (improved ones, please) to select and rank the teams.

    BTW, an interconference interlude in February would add to the appeal of college basketball.

    Just my two cents (puffed up to represent $ millions).

    sagegrouse

  7. #7
    Join Date
    Mar 2008
    Location
    New Orleans, Louisiana
    You can imagine a situation where the best possible metric for evaluating college basketball teams is adopted for purposes of tournament selection. There's a list of every team, from 1 to 345. "Selection" consists of highlighting the 31 automatic qualifiers and then taking the top 37 teams that aren't highlighted.

    The problem with this is that it gives the Selection Committee less to do. Sure, they still seed those 68 teams and place them in a bracket. But they've voluntarily eliminated the (political) power of inviting teams and severely reduced the privilege of what it means to be on the Selection Committee. That means they'd have to work harder to justify the lavish travel expenses and lodging that they incur to do something they could just as easily do by teleconference in a few hours.

    Never forget that the human element of the tournament selection process involves, well, humans.

  8. #8
    Join Date
    Feb 2007
    Location
    Inman, SC & Fort Myers, FL
    I can imagine an algorithm which would automatically select the best teams. Look how well the BCS has worked!

    I agree with Indoor66 -- I don't want to watch the sausage being made. Besides, if you had had the magic algoprithm, what would all the complainers find to complain about? Too automatic, too inhuman?

  9. #9
    Quote Originally Posted by NSDukeFan View Post
    Already done. I would have written that as the link like Duvall often does, but I haven't figured out how yet.
    http://espn.go.com/mens-college-bask...rankings-final
    Wow, I can't believe I didn't notice it last year. Well, again, it brings to my opinion that ESPN, like many other news organizations, create news instead of report news. While I appreciate the opinions of "experts" and seek their analysis and insights, I prefer they keep the opinions in the Editorial Section and out of the Front Page.

  10. #10
    Join Date
    Feb 2011
    Location
    High Point
    "1. Give each conference a given number of teams to make the conference.
    a. I would give the major conferences one-half of their total membership, invoking the rule that if you aren't in the top one-half, you don't deserve to be in the NCAA's.

    b. WRT to other conferences, I would let past performance affect the number of teams selected. Use a relegation system to give conferences more or fewer selections.)

    2. Let the goldarned conferences decide who represents them."

    Thank goodness you have nothing to do with the process.

  11. #11
    Join Date
    Aug 2009
    Location
    Maryland
    The RPI wasn't designed to be fair or predictive, and therefore it isn't. It rewards schedule strength to a very large fault. Both Sagarin Ratings and Pomeroy are much better. but, since any team with even a remote chance of winning gets in anyway -- it's only a matter of early round dancing. But then there is football.

Similar Threads

  1. Jay Bilas' suggestions to improve selection process
    By Nugget in forum Elizabeth King Forum
    Replies: 7
    Last Post: 03-23-2011, 02:21 AM
  2. Women's NCAAT selection show
    By Bluedawg in forum Elizabeth King Forum
    Replies: 4
    Last Post: 03-18-2008, 10:58 AM
  3. Admissions Process
    By Fish80 in forum Off Topic
    Replies: 0
    Last Post: 12-13-2007, 09:29 AM
  4. Surprising process
    By Classof06 in forum Elizabeth King Forum
    Replies: 14
    Last Post: 03-22-2007, 03:19 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
  •