The NCAA Tournament Selection Committee is on a Fool's Errand

[sagegrouse]

Def: Fool’s errand. A task or activity with no hope of success. Or, a fruitless or thankless undertaking.

The Tournament Selection Committee is on a fool’s errand. It is charged with the selecting the 36 at-large teams in the tournament and then seeding all 68 teams. It has no hope of success in either undertaking.

Here’s why:

1. The ranking of teams to pick NCAA at-large selections can be thought of as having two parts: (a) ranking teams within conferences and (b) ranking or calibrating conferences with each other.

2. Ranking or rating or comparing teams within conferences is comparatively easy: each team plays 16-18 scheduled conference games and a few more in the conference tournament. An analyst or committee can do all sorts of analysis with such rich data. Moreover, the games are almost all between early January and early March and are pertinent for a March tournament.

3. Measuring the conferences against each other is really difficult because the data are lousy.


a. Almost all nonconference matchups are played between mid-November and the first few days of January. Therefore, they are somewhat stale compared to the performance of the teams 2-3 months later. Results in, say, November may have little bearing on how good the teams are come tournament time (and the TSC purports to measure improvements).

b. The nonconference schedules are a ridiculous hodgepodge of games with little rhyme or reason. For example, there are 75 teams in what I call the “Top Six” conferences. I include the Big East rather than the AAC or the A-10. (If anyone wants to extend the analysis, be my guest!) There are only 141 games played among these 75 teams, or slightly fewer than 2.0. and it’s not like there are only a few nonconference games. Each team plays between 12 and 14 plus a few more for some teams in exempt tournaments. The Big Ten-ACC and the Big 12-SEC challenges create good match-ups of peer teams, but they are the exceptions.

c. To give a measure of the inadequacies of the interconference data, consider this: Only three times this season did an AP Top Ten team (current rankings) lose to a team from another Top Six conference: ND (to Providence); Wisconsin (to Duke) and Kansas (annihilated by Kentucky).


So, when the Tournament Selection Committee goes into the locked room in March, they are just guessing. Joe Lunardi is in the same boat.

As I said in the title to this thread, the TSC is on a fool’s errand. Reconsidering, it is also a public relations stunt designed, intentionally or not, to build interest in the NCAA basketball tournament. The somewhat different process used by the College Football Playoff Committee was clearly geared to build interest and ratings. So I don’t think the TSC is gonna change.

If anyone in the sports networks or the NCAA is interested, there is at least one way to improve the process by adding more relevant information. Create an “interconference week” in February that systematically schedules two interconference games for each team, matching teams between peer conferences. That would require two (or two more) conference games to be played in December, but what’s the harm in that? Or, the idea could be expanded to include two weeks of interconference games, which would provide more data and spur interest in college basketball.

Comments are freely solicited.

Sage

[hurleyfor3]

If it's such a bad process, why hasn't there been any sort of consensus to change it in the last 35 years? The last significant change in the process was to seed the teams in 1979. Everything else since then was either expanding the field or adding/changing relatively minor rules, mostly regarding who can go where or play whom when.

College football has had what, four different "championship eras" in the last 20 years? The basketball tournament doesn't need to be fixed because it ain't broke.

[jacone21]

Good stuff.

All I know is... the Joe Lunardi split screen during live action needs to go. I'm pretty tolerant of ESPN's nonsense, but that is just stupid.

What would we do without the annual tradition of bubble teams and last four outs and people who don't know the difference between an over and back and an over the back picking the winner of a game between Southern Methodist and Georgia Southern?

Just enjoy the ride.

[Wander]

Yeah, I don't think the process needs to be changed, except that having exactly 64 teams would be nice. If it was as much of a crapshoot as you're saying, it wouldn't be so easy to predict the vast majority of teams.

However, the interconference week would be cool. Note that they DID have such a week for non-BCS conferences called BracketBusters, but it was canceled. The George Mason and VCU teams that made the Final Four might not have even gotten into the tournament without it.

[Duvall]

The simple answer is that the real task of the Selection Committee is not to accurately compare and rank 68 conference champions and at-large teams - that would be impossible to do with any precision, as you say. Their real task is to create a compelling championship event. This only requires the Committee to be mostly right, or even somewhat right, with the good teams at the top, the okay teams in the middle and the Cinderellas sprinkled at the bottom. This they can manage, and it's enough to make the Tournament a success each year.

[Kedsy]

Here's my question, sage: did you just cut and paste your opening post from your similar post last year? I seem to remember you bringing up this same subject on more than one occasion.

As far as the substance, I think you're right that the selection committee has a hard job. I think your interconference week is a fun idea, though I'm not sure it would add enough data points to solve the problem you're attempting to solve. I also agree with those who say the committee doesn't have to get it perfect; they just have to do well enough to provide a compelling tournament, which they achieve more often than not.

[sagegrouse]

Here's my question, sage: did you just cut and paste your opening post from your similar post last year? I seem to remember you bringing up this same subject on more than one occasion.

As far as the substance, I think you're right that the selection committee has a hard job. I think your interconference week is a fun idea, though I'm not sure it would add enough data points to solve the problem you're attempting to solve. I also agree with those who say the committee doesn't have to get it perfect; they just have to do well enough to provide a compelling tournament, which they achieve more often than not.

The title and definition are the same; the rest of the post is totally rewritten.

[Kedsy]

The title and definition are the same; the rest of the post is totally rewritten.

OK, thanks. I couldn't remember the exact language of your previous posts. It's certainly a subject worth revisiting.

[DukeandMdFan]

[ people who don't know the difference between an over and back and an over the back [/QUOTE]

{Somewhat OT, but I was just at a basketball officials website. I'm a high school basketball official and attend officials camps with college officials as instructors. One thing that "amuses" officials is when coaches, players, and fans yell "over be the back" when there is nothing in the rule book prohibiting "over the back". A taller/better leaper may reach over a smaller player - and even make incidental contact - for a rebound without it being a foul. It is only a foul if the player being boxed out "displaces" the player in front of him/her. In practice, Cook can box out Okafor, but Okafor could still get the rebound without commiting a foul.}

I agree that Lunardi gets too much air time. I would rather that effort be put into scheduling games in February to gauge which conferences are the best and how the teams from the top of the mid-major conferences compare with the teams from the middle of the power conferences.

Thinking out of the box, they could have challenges between conferences and the number of bids to the tourney be based on interconference records. For example, the NCAA determines that the ACC, B1G, and SEC get a combined 15-18 bids to the NCAAA tourney. Each conference gets 1 automatic bid to their champion. After "Rivalry Week" in January, play 4 games between ACC & B1G; 4 games between ACC & SEC; 4 games between ACC and SEC. Each conference plays 2 games against teams from lesser conferences. For every game a conference wins, they get a bid to the tourney. I haven't thought this through, but all the effort spent by Lunardi, etc. to rank the teams could be used to develop a method to settle it on the court.

[BigWayne]

The simple answer is that the real task of the Selection Committee is not to accurately compare and rank 68 conference champions and at-large teams - that would be impossible to do with any precision, as you say. Their real task is to create a compelling championship event. This only requires the Committee to be mostly right, or even somewhat right, with the good teams at the top, the okay teams in the middle and the Cinderellas sprinkled at the bottom. This they can manage, and it's enough to make the Tournament a success each year.

This is right on. People like sage have a difference of opinion of what the goal should be compared to what the NCAA has set out to do. They will therefore never be happy. The NCAA has set out the parameters of how they are going to do the selection and seeding and teams then need to schedule and perform to meet those parameters. That does not necessarily pick the best teams for the season or especially for the end of the season.

As for Lunardi, putting him on the TV incessantly is annoying, but his statements and opinions are also not about the best teams. He is just trying to predict how the teams performance will be measured by the committee's stated goals and past performance. He is more focused on the committee actions than which teams are really the best.

In the end, for teams that want highs seeds or to be on the right side of the bubble, it usually comes down to winning more games this time of year, which is pretty easy to understand.

[BigWayne]

{Somewhat OT, but I was just at a basketball officials website. I'm a high school basketball official and attend officials camps with college officials as instructors. One thing that "amuses" officials is when coaches, players, and fans yell "over be the back" when there is nothing in the rule book prohibiting "over the back". A taller/better leaper may reach over a smaller player - and even make incidental contact - for a rebound without it being a foul. It is only a foul if the player being boxed out "displaces" the player in front of him/her. In practice, Cook can box out Okafor, but Okafor could still get the rebound without commiting a foul.}


Reminds me of Billy Packer for some reason. Yes, you can jump completely over an opponents back and, if you don't touch him, there is no foul. Likewise, "reaching in" is not a foul if you reach in and don't hit anything. Bad announcers and bad referees that guess a lot are the ones that believe in "over the back" and "reaching in" fouls.

[-jk]

I like the idea of an out-of-conference week. Not only would make for better comparisons for the committee, it would also help the dork polls and give the TV audience some interesting games.

K has said he likes to have a tough out-of-conference game in February to get outside eyes on his team so he can find weaknesses before March.

(And take the NCAAs back to 64.)

-jk

[MarkD83]

Def: Fool’s errand. A task or activity with no hope of success. Or, a fruitless or thankless undertaking.

If anyone in the sports networks or the NCAA is interested, there is at least one way to improve the process by adding more relevant information. Create an “interconference week” in February that systematically schedules two interconference games for each team, matching teams between peer conferences. That would require two (or two more) conference games to be played in December, but what’s the harm in that? Or, the idea could be expanded to include two weeks of interconference games, which would provide more data and spur interest in college basketball.

Comments are freely solicited.

Sage

What I am about to add was suggested by jk and Dukeandmdfan but I will embellish. The committee already has a time where they sit down with the media and do a "mock" seeding. Make this a first official seeding like NCAA football just did. Then pick the 8-16 teams that are just in or just out. Take the week in Feb and have these teams play a mini 2 game tournament. 1 plays 16; 2 plays 15 etc with a winners and losers bracket. They only need 2 games each. During the same week you can take the top seeds in the mock bracket and have them play 2 games.

At that point you have more data like sage suggested to use to rank the teams. In addition the bubble teams now know what they have to do to get in down the stretch....win their conference tournament...don't have bad losses...

[Tom B.]

{Somewhat OT, but I was just at a basketball officials website. I'm a high school basketball official and attend officials camps with college officials as instructors. One thing that "amuses" officials is when coaches, players, and fans yell "over be the back" when there is nothing in the rule book prohibiting "over the back". A taller/better leaper may reach over a smaller player - and even make incidental contact - for a rebound without it being a foul. It is only a foul if the player being boxed out "displaces" the player in front of him/her. In practice, Cook can box out Okafor, but Okafor could still get the rebound without commiting a foul.}





Tim Duncan was a master at this. He'd jump straight up and reach his long arms out over the top of the player who was boxing him out, and grab the rebound. He often did this on Wake's offensive end of the floor, resulting in a lot of stickback points for him. Opposing coaches would scream, "Over the back!" but it wasn't a foul. It was just good, sound play by a tall guy with long arms.

[Kedsy]

At that point you have more data like sage suggested to use to rank the teams.

If you're going to do all that (and I like the suggestion), why not give the winner an automatic bid? We're talking about bubble teams that have a decent chance to get in anyway, so it's not like giving another automatic bid to a low-major or anything. It would give us more information and make for riveting TV. Win-win!

[Mtn.Devil.91.92.01.10.15]

I disagree with the presumption that if you can't do this perfectly, then it is a fool's errand. So, would you suggest that the only way that one would recognize a correctly seeded and selected tournament would be if higher seeds won every game by smaller and smaller margins each round and all #1 seeds advanced to the final four to play three very close games?

Does that sound like any fun?

Besides, even that wouldn't necessarily prove that the committee "got it right." Just because a 15 upsets a 2 doesn't mean there was a mistake on the part of the NCAA committee. Injuries, hot streaks, cold streaks, bad calls, fouls, etc can swing things wildly without necessarily reflecting the "goodness" of the team.

I would say that the most important tasks of the committee (from my totally biased POV) are selecting the best teams at large, and keeping an eye towards "fairness." By that I mean, keeping in mind travel distances, overloading one bracket, etc. Beyond that, I have very little quarrel. Perhaps that's because as a Duke fan, I feel it's on our team to win six games and make all the other considerations moot.

Anyways, to the OP's point, I agree that consolidating the information available to them and attempting to discern who the most qualified teams are is very difficult. But, I doubt very seriously that many teams that were left out all together would have had a realistic chance to win.

What I love about college basketball is that even now - with 80% of the regular season in the books - every team still controls their own destiny in regards to winning it all. Well, every team except Syracuse.

[Olympic Fan]

Frankly, I don't have a problem with the process. What would you replace it with -- a strict mathematical formula? Just check out Pomeroy or RPI or Sagarin t see how absurd those rankings would be.

My only problem is the makeup of the committee. Its almost entirely made up of administrators -- not basketball people. Last year, I believe one selection member had a basketball background. In recent years, the committee has been dominated by mid-major and low major representatives. They've done a good job of taking care of their teams at the expense of the power conferences. The good news for Duke is that Kevin White will join the committee next year and despite what they say about "stepping out of the room", committee members' teams and conferences always get taken care of.

I'd also like to see the field reduced to 64 again.

Still, we have a large enough field to overcome the mistakes of he committee. There is no chance in my mind that a team capable of winning the championship misses the tournament (other than probation). A contender might face a difficult -- and unfair -- seed (as Duke did in 2013), but it might also get a favorable path to the title (as Duke did in 2010). Either way, as Jim Valvano once said, "If you get in the field, you can play your way out of a bad seed."

[bedeviled]

In recent years, the committee has been dominated by mid-major and low major representatives. They've done a good job of taking care of their teams at the expense of the power conferencesHere's an article (http://keepingscore.blogs.time.com/2013/03/15/predicting-the-ncaa-mens-basketball-field-and-discovering-the-selection-committees-biases/) about some bracketologists making the "Dance Card" bracket predictions. I don't think it's worth reading, but I wanted to point out that they had to incorporate factors for (committee members) and (major vs mid-major) into their model to more accurately predict invitations and seeding. According to the article,

Looking back on the data from 1999-2008, they discovered some shocking stats. When comparing more objective metrics with what the committee decided each year, being in the Pac-12 (or Pac-10 at the time) was equivalent to being 31 spots higher in the RPI. Being in the Big 12 equaled to being 17 spots higher. For the Missouri Valley, being in that conference was like they were 32 spots lower.

[throatybeard]

I like the idea of an out-of-conference week. Not only would make for better comparisons for the committee, it would also help the dork polls and give the TV audience some interesting games.

K has said he likes to have a tough out-of-conference game in February to get outside eyes on his team so he can find weaknesses before March.

-jk

Yeah, we used to do this more reliably. Having additional conference games certainly doesn't facilitate it.

[sagegrouse]

I disagree with the presumption that if you can't do this perfectly, then it is a fool's errand. So, would you suggest that the only way that one would recognize a correctly seeded and selected tournament would be if higher seeds won every game by smaller and smaller margins each round and all #1 seeds advanced to the final four to play three very close games?

.


Frankly, I don't have a problem with the process. What would you replace it with -- a strict mathematical formula? Just check out Pomeroy or RPI or Sagarin t see how absurd those rankings would be.



What would be better? Did the little kid who yelled, "The Emperor has no clothes!" offer a solution?

I'd probably work to change the interconference schedules to create stronger and fairer tests on a national stage. I mean, the "Cupcake City" aspects of today's major programs is really ridiculous, except for the economic benefits to the lesser programs. Also, I like the idea of inter-conference play in January or February. Incidentally, the sports networks would probably love the ability to get more meaningful games.

Barring that change, I would probably give the job back to the conferences. For the major conferences (still to be adequately defined), I'd give the top one-half automatic bids and let the conferences decide who those teams are. (The prospect of the bloodletting is strangely thrilling.) If you're not in the top half of your conference, you don't go to the NCAA's. (Yeah, yeah! The ACC as the "odd man out" would alternate between seven and eight.) There would be some complexities, but I think there's a workable system out there.

Or maybe it's good enough just to admit that there is virtually no sound way to compare the relative strength of the conferences, so "We are just guessing." Which was my point to begin with.

Meanwhile, there is a special circle in hell for Joe Lunardi, who knows this is ludicrous but needs to make a living some way.

[ChillinDuke]

What would be better? Did the little kid who yelled, "The Emperor has no clothes!" offer a solution?

I'd probably work to change the interconference schedules to create stronger and fairer tests on a national stage. I mean, the "Cupcake City" aspects of today's major programs is really ridiculous, except for the economic benefits to the lesser programs. Also, I like the idea of inter-conference play in January or February. Incidentally, the sports networks would probably love the ability to get more meaningful games.

Barring that change, I would probably give the job back to the conferences. For the major conferences (still to be adequately defined), I'd give the top one-half automatic bids and let the conferences decide who those teams are. (The prospect of the bloodletting is strangely thrilling.) If you're not in the top half of your conference, you don't go to the NCAA's. (Yeah, yeah! The ACC as the "odd man out" would alternate between seven and eight.) There would be some complexities, but I think there's a workable system out there.

Or maybe it's good enough just to admit that there is virtually no sound way to compare the relative strength of the conferences, so "We are just guessing." Which was my point to begin with.

Meanwhile, there is a special circle in hell for Joe Lunardi, who knows this is ludicrous but needs to make a living some way.

But Sage, I lose you right off the bat when you say that they are tasked with selecting the 36 at-large teams and have no hope.

Surely you would admit that at least 20 of those teams (and probably 25-30) are relatively easy to identify and select, regardless of conference strength.

So, in my eyes, it really comes down to selecting the last 16 (and probably 6-11) teams for at-large spots. While 16 teams is starting to approach a larger undertaking, a more realistic selection of 6-11 teams that have a reasonable amount of debate as to inclusion doesn't seem to equate to "no hope". At least that's how I see it. Especially when these teams are generally in the 7-12 seed range, meaning their importance as to inclusion is not extremely high (from the perspective of the pool of teams as a whole).

I guess I just don't see many problems with the overall selection of teams by the TSC.

Now seeding, that's a different story. And I could probably agree with you that there's "no hope" there depending on your definition. But seeding seems less important to me than selection as I've explained in other threads that my view is you play to win. So while accurate seeding is great and all, you still have to win. You don't get points for being seeded higher. And I don't really subscribe to chalking a loss up to "getting a bad seed".

So I come back to selection. Which I feel the TSC generally does a fine job at. Perfect? By no means. But adequate to more than adequate.

- Chillin (tab: $16.88)

[Mtn.Devil.91.92.01.10.15]

What would be better? Did the little kid who yelled, "The Emperor has no clothes!" offer a solution?

I'd probably work to change the interconference schedules to create stronger and fairer tests on a national stage. I mean, the "Cupcake City" aspects of today's major programs is really ridiculous, except for the economic benefits to the lesser programs. Also, I like the idea of inter-conference play in January or February. Incidentally, the sports networks would probably love the ability to get more meaningful games.

Barring that change, I would probably give the job back to the conferences. For the major conferences (still to be adequately defined), I'd give the top one-half automatic bids and let the conferences decide who those teams are. (The prospect of the bloodletting is strangely thrilling.) If you're not in the top half of your conference, you don't go to the NCAA's. (Yeah, yeah! The ACC as the "odd man out" would alternate between seven and eight.) There would be some complexities, but I think there's a workable system out there.

Or maybe it's good enough just to admit that there is virtually no sound way to compare the relative strength of the conferences, so "We are just guessing." Which was my point to begin with.

Meanwhile, there is a special circle in hell for Joe Lunardi, who knows this is ludicrous but needs to make a living some way.

I still say that the presumption of your argument - that the existing system isn't working - is at the very least up for debate.

Are you suggesting that scheduling needs to change to give better information for evaluation? Or that they process needs to be changed to bring in better scheduling?

And yes, Lunardi is the worst kind... the idiot that baits us in and gets us to click on his links every week and debate them here as though they have meaning.

[sagegrouse]

But Sage, I lose you right off the bat when you say that they are tasked with selecting the 36 at-large teams and have no hope.

Surely you would admit that at least 20 of those teams (and probably 25-30) are relatively easy to identify and select, regardless of conference strength.

So, in my eyes, it really comes down to selecting the last 16 (and probably 6-11) teams for at-large spots. While 16 teams is starting to approach a larger undertaking, a more realistic selection of 6-11 teams that have a reasonable amount of debate as to inclusion doesn't seem to equate to "no hope". At least that's how I see it. Especially when these teams are generally in the 7-12 seed range, meaning their importance as to inclusion is not extremely high (from the perspective of the pool of teams as a whole).

I guess I just don't see many problems with the overall selection of teams by the TSC.

Now seeding, that's a different story. And I could probably agree with you that there's "no hope" there depending on your definition. But seeding seems less important to me than selection as I've explained in other threads that my view is you play to win. So while accurate seeding is great and all, you still have to win. You don't get points for being seeded higher. And I don't really subscribe to chalking a loss up to "getting a bad seed".

So I come back to selection. Which I feel the TSC generally does a fine job at. Perfect? By no means. But adequate to more than adequate.

- Chillin (tab: $16.88)


My intellectual journey began, much as Archimedes in the bathtub, with a simple observation. In 2006 the Missouri Valley Conference got four bids to the NCAA's. Why? There was only one ranked team (#25 No. Iowa) and one other that received AP votes (Wichita State). Aside from the fact that a persuasive conference official was on the TSC, what was there to boost so many teams from the MVC over teams from apparently stronger conferences? I looked at the records of all the teams and the only thing I found was one game: On December 19, Northern Iowa beat LSU by four points in Baton Rouge; there were no other significant positive results for the conference. One game results in three at-large bids? How could that be?

"Eureka!" I said, as I suddenly realized that there are few useful interconference results to align conferences and those few meaningful games occurred almost exclusively in November and December, when many teams are just in the formative stage. The Tournament Selection Committee has to rely on heuristics (Well, the Big Ten always has a lot of good teams), histrionics (State U. belongs in the Tournament!), optics (So-and-so is a terrific player), and political haggling among competing conferences (Most of whom are in the room).

In any event, whatever we do or do not do about the NCAA selection process, please recognize the TSC doesn't have the data to properly evaluate the teams it selects. It is almost like a group of ancient religious scholars pondering murky and conflicting sacred writings and then coming out of their temple with a series of pronouncements.

Anyway, that's my excuse and I am sticking to it.

Kindly, Sage
'Oh, and as it turns out, that was the LSU team that upset #1 Duke and went to the Final Four. But I didn't know that would happen when I did my MVC analysis'

[Mtn.Devil.91.92.01.10.15]

My intellectual journey began, much as Archimedes in the bathtub, with a simple observation. In 2006 the Missouri Valley Conference got four bids to the NCAA's. Why? There was only one ranked team (#25 No. Iowa) and one other that received AP votes (Wichita State). Aside from the fact that a persuasive conference official was on the TSC, what was there to boost so many teams from the MVC over teams from apparently stronger conferences? I looked at the records of all the teams and the only thing I found was one game: On December 19, Northern Iowa beat LSU by four points in Baton Rouge; there were no other significant positive results for the conference. One game results in three at-large bids? How could that be?

"Eureka!" I said, as I suddenly realized that there are few useful interconference results to align conferences and those few meaningful games occurred almost exclusively in November and December, when many teams are just in the formative stage. The Tournament Selection Committee has to rely on heuristics (Well, the Big Ten always has a lot of good teams), histrionics (State U. belongs in the Tournament!), optics (So-and-so is a terrific player), and political haggling among competing conferences (Most of whom are in the room).

In any event, whatever we do or do not do about the NCAA selection process, please recognize the TSC doesn't have the data to properly evaluate the teams it selects. It is almost like a group of ancient religious scholars pondering murky and conflicting sacred writings and then coming out of their temple with a series of pronouncements.

Anyway, that's my excuse and I am sticking to it.

Kindly, Sage
'Oh, and as it turns out, that was the LSU team that upset #1 Duke and went to the Final Four. But I didn't know that would happen when I did my MVC analysis'

Well, that makes a lot of sense.

[Wander]

My intellectual journey began, much as Archimedes in the bathtub, with a simple observation. In 2006 the Missouri Valley Conference got four bids to the NCAA's. Why? There was only one ranked team (#25 No. Iowa) and one other that received AP votes (Wichita State). Aside from the fact that a persuasive conference official was on the TSC, what was there to boost so many teams from the MVC over teams from apparently stronger conferences? I looked at the records of all the teams and the only thing I found was one game: On December 19, Northern Iowa beat LSU by four points in Baton Rouge; there were no other significant positive results for the conference. One game results in three at-large bids? How could that be?

That's it? You realize the year that you're using is the perfect example of why the smaller conferences deserve at-large bids, right? Billy Packer threw a legendary hissy fit over the MVC getting as many at-large teams as the ACC and George Mason getting in, and then looked like a fool when the MVC got as many Sweet 16 teams as the ACC and George Mason made the Final Four.

[Mtn.Devil.91.92.01.10.15]

Well, that makes a lot of sense.

To clarify, I mean your confusion makes sense - not the actions of the committee.

[sagegrouse]

That's it? You realize the year that you're using is the perfect example of why the smaller conferences deserve at-large bids, right? Billy Packer threw a legendary hissy fit over the MVC getting as many at-large teams as the ACC and George Mason getting in, and then looked like a fool when the MVC got as many Sweet 16 teams as the ACC and George Mason made the Final Four.

Isn't my point that there is little or no useful data in deciding who's in and who's out, not whether the MVC deserved the bids. I had no axe to grind against the MVC; I just wondered about the rationale. In this case, there was exactly one game that seemed relevant.

[Indoor66]

[ people who don't know the difference between an over and back and an over the back

{Somewhat OT, but I was just at a basketball officials website. I'm a high school basketball official and attend officials camps with college officials as instructors. One thing that "amuses" officials is when coaches, players, and fans yell "over be the back" when there is nothing in the rule book prohibiting "over the back". A taller/better leaper may reach over a smaller player - and even make incidental contact - for a rebound without it being a foul. It is only a foul if the player being boxed out "displaces" the player in front of him/her. In practice, Cook can box out Okafor, but Okafor could still get the rebound without commiting a foul.}[/QUOTE]

You can thank El Deano from the Dump on the Hump for the "Over the back" insanity. He used to scream that all the time. It would drive many of us crazy because the refs started reacting to his constant harping and calling fouls when there was no foul. Thanks dean!

[Wander]

Isn't my point that there is little or no useful data in deciding who's in and who's out, not whether the MVC deserved the bids. I had no axe to grind against the MVC; I just wondered about the rationale. In this case, there was exactly one game that seemed relevant.

Fair enough, but it's IMO foolish to think that only one game counted. Northern Iowa over LSU wasn't the only MVC win against a tournament team - it wasn't even the only MVC win against a Final Four team. Both Kenpom and RPI had all 3 of the at-large MVC teams from that year in the top 35 (post-tournament, unfortunately I don't see pre-tournament numbers, but still). A fifth team, Missouri State, is 39 in kenpom, and at the time was the highest RPI to get left out of the tournament. These metrics, agree or disagree with them, of course take into account all games.

[sagegrouse]

Fair enough, but it's IMO foolish to think that only one game counted. Northern Iowa over LSU wasn't the only MVC win against a tournament team - it wasn't even the only MVC win against a Final Four team. Both Kenpom and RPI had all 3 of the at-large MVC teams from that year in the top 35 (post-tournament, unfortunately I don't see pre-tournament numbers, but still). A fifth team, Missouri State, is 39 in kenpom, and at the time was the highest RPI to get left out of the tournament. These metrics, agree or disagree with them, of course take into account all games.

I welcome your results. I didn't build a spreadsheet but did paper notes. Still, if there are very few games between MVC teams and other conferences, both KenPom and RPI are pretty unstable measures, even though they "take into account all games."

[Wander]

I welcome your results. I didn't build a spreadsheet but did paper notes. Still, if there are very few games between MVC teams and other conferences, both KenPom and RPI are pretty unstable measures, even though they "take into account all games."

I just don't think they're as unstable as you think they are. Again, if it was as much of a guessing game as you think it is, then why is it so easy for everyone to correctly pick the vast majority of the teams, sometimes literally all the teams, that the selection committee picks every single year?

The bigger point about that year with the MVC teams is that it shows why the idea of just locking in the top half of teams from certain conferences is absurd. Some years, like in 2006, there is going to be a non-power conference like the MVC that is so good that any team a couple games over .500 in the league deserves a look. Some years, like the SEC this year, there is going to be a power conference that is weak enough that it doesn't deserve that (Alabama would currently make the tournament under your rule). Actually, nearly every year these two things are going to occur. Conference strength is in flux, and one of the nice things about college basketball is that there are no formal advantages given to certain conferences in the selection format, in part for this reason. It's the last thing that needs to change.

I agree with your point that it would be nice if we had some non-conference games later in the season.

(In case you're wondering why I have an oddly specific memory about the 2006 tournament, it was the first year I got really into college basketball beyond Duke, and won all my bracket pools by going all-in on the MVC and Northwestern State).

[sagegrouse]

Fair enough, but it's IMO foolish to think that only one game counted. Northern Iowa over LSU wasn't the only MVC win against a tournament team - it wasn't even the only MVC win against a Final Four team. Both Kenpom and RPI had all 3 of the at-large MVC teams from that year in the top 35 (post-tournament, unfortunately I don't see pre-tournament numbers, but still). A fifth team, Missouri State, is 39 in kenpom, and at the time was the highest RPI to get left out of the tournament. These metrics, agree or disagree with them, of course take into account all games.

If you recall, my "parable of the MVC" came up as the source of my thinking about the data set (and its glaring weaknesses) used by the NCAA TSC. In that sense, the 2006 season is "beside the point."

[Neals384]

So, when the Tournament Selection Committee goes into the locked room in March, they are just guessing. Joe Lunardi is in the same boat.

Sage

Darn. Just when I was thinking your entire post would not mention HIM...

[Neals384]

Besides, even that wouldn't necessarily prove that the committee "got it right." Just because a 15 upsets a 2 doesn't mean there was a mistake on the part of the NCAA committee.

Now why would you bring that up?

[Henderson]

Ever look at a piece of modern art hanging in a fancy gallery and think, "My 6 year old could do that."?

Joe Lunardi's genius is in painting like a 6 year old on a paint-by numbers easel and getting some weird art dealer to think it was "important."

[Turk]

I'm puzzled by all the Lunardi bashing. The premise is simple: he tries to apply the rules of the selection committee to create a bracket as if the tournament would start tomorrow. This is useful because like everything else the NCAA does, the bracketing rules are more complicated than they need to be. He was the first one to do it, and now we have Palm and a bunch of others. What's wrong with that? No one is claiming that predicting brackets is "art" or "clickbait" or anything like that. Feel free to take it or leave it; seems like there are more important things to get worked up about.

When Lunardi makes a case for or against a team, he makes fact-based arguments. I like the rule that Lunardi and many others have proposed: if you're not .500 or better in your conference, you don't get a bid.

I do agree that having him split-screen on the telly with live action isn't working, but I hardly think it was his idea. The same thing goes for when they split-screen to interview a player's mom or some retired coach or celebrity. If you split my screen, I better be seeing another ballgame.

[Henderson]

I'm puzzled by all the Lunardi bashing. The premise is simple: he tries to apply the rules of the selection committee to create a bracket as if the tournament would start tomorrow.

Millions of fans do the same thing, and they all have the information Lunardi has. He says he's just trying to imagine what the selection committee might do. But that's what fans do. He's not special. His facts are available to us all. So what is his useful role?

[Kedsy]

Millions of fans do the same thing, and they all have the information Lunardi has. He says he's just trying to imagine what the selection committee might do. But that's what fans do. He's not special. His facts are available to us all. So what is his useful role?

Same as many other sports commenters and reporters. What's their useful role?

[Neals384]

I'm puzzled by all the Lunardi bashing.

I hope you enjoyed the endless split-screen views of Mr. Lunardi bloviating during the first half of today's game.

[uh_no]

I'm puzzled by all the Lunardi bashing. The premise is simple: he tries to apply the rules of the selection committee to create a bracket as if the tournament would start tomorrow. This is useful because like everything else the NCAA does, the bracketing rules are more complicated than they need to be. He was the first one to do it, and now we have Palm and a bunch of others. What's wrong with that? No one is claiming that predicting brackets is "art" or "clickbait" or anything like that. Feel free to take it or leave it; seems like there are more important things to get worked up about.

When Lunardi makes a case for or against a team, he makes fact-based arguments. I like the rule that Lunardi and many others have proposed: if you're not .500 or better in your conference, you don't get a bid.

I do agree that having him split-screen on the telly with live action isn't working, but I hardly think it was his idea. The same thing goes for when they split-screen to interview a player's mom or some retired coach or celebrity. If you split my screen, I better be seeing another ballgame.

To some degree I agree, and some I don't. Here's my take:

1) most fans are really idiots, and I learned this by reading the Chats those guys used to (and still?) do on ESPN....people would seriously ask stuff like (perhaps this year) "do you think this is the year VT will get over the hump and get an at large?" and of course reasonable people say "NO, of course not"....but read the first 5 words of this bullet point again
2) the rules of the bracket are really quite complicated....who can play where, what days can teams play on, how far can you deviate from s-curve seed in order to make the bracket work, and who gets in. JB brackets based on all the rules, and people like brackets...so there you go
3) as you say, he was the first to really do it throughout the season, and it stuck
4) in terms of picking teams to get in, he's really any better than most pundits. There're 36 (?) at large teams, and he might get 33 right....the problem is 30 of them are effectively locks, so he's really only batting 50%....and those are often big misses that everyone misses....so a random schmuck on the street could probably hit 30 at large teams out of 36 with no sweat
5) obviously after you get off the 1 seed line, slotting teams into the actual bracket is a crap shoot....even if you have the s-curve dead on, there might be 10,000 different ways you can adjust it to fit the bracketing rules.

So he is what he is. every once in a while i peek in to get an idea where he things duke or uconn will be....but GDI stay off my split screen!

[juise]

Mods - please feel free to move this if this isn't a good spot. I didn't think this merited it's own thread.

In this week's Mark Titus mailbag (http://grantland.com/the-triangle/ncaa-college-basketball-rankings-top-12-wichita-state-shockers-big-12-maryland-terrapins-notre-dame-fighting-irish-gonzaga-bulldogs-kansas-jayhawks-wisconsin-badgers-arizona-wildcats-duke-blue-devils/), a reader suggested an alternate bracket method. Essentially, the committee would just release a ranked list of teams 1-68 (with 1-4 slotted as #1's, 5-8 as #2's, etc.) and the team's would get to pick their spot in the bracket. It would be televised like a draft with each team having only a few minutes "on the clock." This would give the teams a chance to weigh location versus opponent and any other consideration. I assume that the first 16 teams would also be able to select their round of 64/32 site.

I love the idea and I would love the theater of the selection process. I am not sure how the play-in game teams would be involved in the process, but I would guess that they would just be assigned the last of their seed (12, for instance) and they would get stuck in the slot nobody else picked. Or better yet, we just go back to 64 teams (I know, not happening). I imagine that there wouldn't be enough teams purposely choosing locations far away from campus to hurt ticket sales. I'm not sure I see a downside.

What do you all think?

[uh_no]

Mods - please feel free to move this if this isn't a good spot. I didn't think this merited it's own thread.

In this week's Mark Titus mailbag (http://grantland.com/the-triangle/ncaa-college-basketball-rankings-top-12-wichita-state-shockers-big-12-maryland-terrapins-notre-dame-fighting-irish-gonzaga-bulldogs-kansas-jayhawks-wisconsin-badgers-arizona-wildcats-duke-blue-devils/), a reader suggested an alternate bracket method. Essentially, the committee would just release a ranked list of teams 1-68 (with 1-4 slotted as #1's, 5-8 as #2's, etc.) and the team's would get to pick their spot in the bracket. It would be televised like a draft with each team having only a few minutes "on the clock." This would give the teams a chance to weigh location versus opponent and any other consideration. I assume that the first 16 teams would also be able to select their round of 64/32 site.

I love the idea and I would love the theater of the selection process. I am not sure how the play-in game teams would be involved in the process, but I would guess that they would just be assigned the last of their seed (12, for instance) and they would get stuck in the slot nobody else picked. Or better yet, we just go back to 64 teams (I know, not happening). I imagine that there wouldn't be enough teams purposely choosing locations far away from campus to hurt ticket sales. I'm not sure I see a downside.

What do you all think?

haha that would be fun!

the only issue is that lower seeds have the advantage of selecting matchups more advantageous to them....so you would likely see more upsets

not that that's a BAD thing.

[gumbomoop]

Sort of.

Watching Stanford @ Arizona St, second half. Stanford player at FT line. AZSt crazies set up a kind of curtain, and just as Stanford guy, um, puts it up, the curtain is opened to reveal ..... 2 fanatics, faces masked with Devil-heads, going at it, simulating smooching and other things.

Replayed several times. Well, it was 12:45 a.m. EDT, so I guess, no problem. Indeed, after commercial, we get a behind the scenes look at the fanatics planning their next performance. What hath the Speedo Guy wrought?

Thread relevance? Stanford presumably has to win this one and maybe @ Arizona this weekend to get back into the picture. Hope Johnny's guys can stay focused and make those FTs. If not, they're screwed.

[JasonEvans]

Watching Stanford @ Arizona St, second half. Stanford player at FT line. AZSt crazies set up a kind of curtain, and just as Stanford guy, um, puts it up, the curtain is opened to reveal ..... 2 fanatics, faces masked with Devil-heads, going at it, simulating smooching and other things.

Replayed several times. Well, it was 12:45 a.m. EDT, so I guess, no problem. Indeed, after commercial, we get a behind the scenes look at the fanatics planning their next performance. What hath the Speedo Guy wrought?

Thread relevance? Stanford presumably has to win this one and maybe @ Arizona this weekend to get back into the picture. Hope Johnny's guys can stay focused and make those FTs. If not, they're screwed.

Ahh, your first introduction to the Curtain of Distraction. Much has been written about the Curtain. The NY Times (http://www.nytimes.com/2015/02/14/upshot/how-arizona-state-reinvented-free-throw-distraction.html?_r=0) even covered it.

Stanford did lose to AzSt. I think Johnny's boys are going to get a #2 or #3 seed... in the NIT. Beating Arizona at Arizona is almost a must win for them now and it appears Stanford is nt up to that task, especially with the injuries the team has endured (another starting F went down last night).

-Jason "as an aside, I love love love love love Sage's notion of an intra-conference week of hoops... maybe the first week of February. Brilliant!" Evans

[ChillinDuke]

Same as many other sports commenters and reporters. What's their useful role?

Well, exactly. Next to nothing.

- Chillin

[gumbomoop]

Ahh, your first introduction to the Curtain of Distraction. Much has been written about the Curtain. The NY Times (http://www.nytimes.com/2015/02/14/upshot/how-arizona-state-reinvented-free-throw-distraction.html?_r=0) even covered it.

I blame myself, never having heard of the Curtain. Don't recall watching a game at AZSt in years, or ever. Nor seen simulated sex behind a curtain, very often. As for missing out on articles about the Curtain, I can only say I have to pay more attention to the literature. Which I will do, as soon as I finish A la recherche du temps perdu. I'm halfway through v.2, the Monty Python edition.

[Nugget]

In this week's Mark Titus mailbag (http://grantland.com/the-triangle/ncaa-college-basketball-rankings-top-12-wichita-state-shockers-big-12-maryland-terrapins-notre-dame-fighting-irish-gonzaga-bulldogs-kansas-jayhawks-wisconsin-badgers-arizona-wildcats-duke-blue-devils/), a reader suggested an alternate bracket method. Essentially, the committee would just release a ranked list of teams 1-68 (with 1-4 slotted as #1's, 5-8 as #2's, etc.) and the team's would get to pick their spot in the bracket. It would be televised like a draft with each team having only a few minutes "on the clock." This would give the teams a chance to weigh location versus opponent and any other consideration. I assume that the first 16 teams would also be able to select their round of 64/32 site.

I love the idea and I would love the theater of the selection process. I am not sure how the play-in game teams would be involved in the process, but I would guess that they would just be assigned the last of their seed (12, for instance) and they would get stuck in the slot nobody else picked. Or better yet, we just go back to 64 teams (I know, not happening). I imagine that there wouldn't be enough teams purposely choosing locations far away from campus to hurt ticket sales. I'm not sure I see a downside.

I think this would be a phenomenal idea. It would be great theater -- a 2-3 hour "draft" of building the bracket one team at a time through choices of each school would be far better television than the annoucement of the results by the black-box Committee, the Chairman answering every question about their iffy decisions with non-disclosure "body of work" garbage and then 2 hours of talking heads complaining about Duke having the easiest region.

It would lead to even better scrutiny/punditry/second-guessing of each team's individual choices; and there would be tremendous drama in those choices -- as noted, "who would shy away from Kentucky," "how far down will teams go to avoid Kentucky," "would Maryland choose to be in Georgetown's bracket or Duke's," etc.

And, it would be fairer than the current process where the Committee adopts blanket rules for placement/bracketing (e.g., "seed line over geography," "geography over matchups," and "we don't project ahead") that wouldn't necessarily play out the same way if any given school was making a context-specific decision. Just some obvious examples: Under the current system, Villanova might get #1 in the West (with #2 Arizona) rather than be #2 in the East (with #1 Virginia). But, other than wearing home jerseys, Villanova might well prefer to be the #2 seed playing U.Va. in Syracuse than the #1 playing Arizona in Los Angeles. Obviously, the Committee can't (or, at least, won't) account for that. But, if Villanova got to choose it could. Or, the one pointed out by Titus' emailer -- under the current system a #2 Wisconsin will be sent to Cleveland since it's closest. But, maybe they'd rather go further away or even be a #3 in a different region to avoid Kentucky. Or, in years' past, we had instances of sub-regional locations in Greensboro and Atlanta or Charlotte and DC, with Carolina set for one of the NC locations -- would Duke prefer to be slightly further away from home but not playing in an arena dominated by Carolina fans? Again, the Committee couldn't possibly account for those concerns. But, by letting teams choose, they'd be able to balance them individually in the way most fair to most teams.

So, while this will probably never happen, I think it would be fantastic and agree I don't see any downside at all, except maybe Lunardi and Jerry Palm would have to take pay cuts.

[JasonEvans]

It would lead to even better scrutiny/punditry/second-guessing of each team's individual choices; and there would be tremendous drama in those choices -- as noted, "who would shy away from Kentucky," "how far down will teams go to avoid Kentucky," "would Maryland choose to be in Georgetown's bracket or Duke's," etc.

I agree that on the surface the draft your spot idea seems like a great idea, but this season it would be an utter disaster. It would unnaturally skew the bracket to make Kentucky's path to the title even easier.

How far would teams go to avoid Kentucky? I would imagine the 8th best team would chose to be a #3 in another region versus being a #2 in Kentucky's bracket. This would probably continue for a while until teams were faced with either being a #4 or #5 seed in a different region or being the #2 in Kentucky's region. so, Kentucky's #2 is going to be something like the 15th - 20th best team team in the land versus being one of the top 8. It would hold true when it came time to fill in the #8-9 game as well, with teams likely choosing to be #10, #11, even #12 seeds (giving them a better shot at a winnable 2nd round game) versus being Kentucky's second round opponent. I could see Kentucky getting a ridiculously easy bracket under this scenario.

-Jason "am I wrong? I certainly think we would see a skew in the 8-9 seeds and probably the #2 seeds" Evans

[Nugget]

I agree that on the surface the draft your spot idea seems like a great idea, but this season it would be an utter disaster. It would unnaturally skew the bracket to make Kentucky's path to the title even easier.

How far would teams go to avoid Kentucky? I would imagine the 8th best team would chose to be a #3 in another region versus being a #2 in Kentucky's bracket. This would probably continue for a while until teams were faced with either being a #4 or #5 seed in a different region or being the #2 in Kentucky's region. so, Kentucky's #2 is going to be something like the 15th - 20th best team team in the land versus being one of the top 8. It would hold true when it came time to fill in the #8-9 game as well, with teams likely choosing to be #10, #11, even #12 seeds (giving them a better shot at a winnable 2nd round game) versus being Kentucky's second round opponent. I could see Kentucky getting a ridiculously easy bracket under this scenario.

-Jason "am I wrong? I certainly think we would see a skew in the 8-9 seeds and probably the #2 seeds" Evans

Nope. That is a huge flaw I'd overlooked. Though it would only be most prevalent in a year like this with a team everyone is looking to avoid.

[Hingeknocker]

Nope. That is a huge flaw I'd overlooked. Though it would only be most prevalent in a year like this with a team everyone is looking to avoid.

What if the committee assigned the seeds as well? I.e. UK-UVA-Duke-Nova are all 1-seeds, but they get to pick their regions in order of the 1-68 rankings? Same thing down the line for the pool of 2-seeds, 3-seeds, etc.

This would allow for the freedom, flexibility and fun of the Titus Plan, but also provide some protection to teams trying to avoid a powerful 1-seed. The same thing would happen lower in the bracket as well, if you didn't do this. Teams that would normally be an 8 or 9 seed would be have the incentive to drop down to the 12 or 13 lines, sacrificing a more difficult 1st round game in exchange for avoiding a 1-seed in the 2nd round.

If this were actually implemented as Titus proposed, it would actually be pretty comical (and take WAY more than 3 hours) to watch the coaches to play out all the permutations and strategies. You see what college coaches already do with a handful of timeouts in the last minute of a game!

Love the idea, overall.

[juise]

What if the committee assigned the seeds as well? I.e. UK-UVA-Duke-Nova are all 1-seeds, but they get to pick their regions in order of the 1-68 rankings? Same thing down the line for the pool of 2-seeds, 3-seeds, etc.

This is actually how I described it above, but I didn't remember correctly (I read the idea a day or so before posting). I'm not sure which way I like better, but I think either would be way better than what we have now. I think I prefer the slotted seed method, but I think you could also argue that a team as intimidating as UK should benefit of teams are that afraid to face them in the regional final.

I think if you just gave each team 2 minutes, the selection show wouldn't be too bad. I think the analysts would have some great reactions to the selections, too. I'd love to see coach interviews while other teams are on the clock.

[Nugget]

What if the committee assigned the seeds as well? I.e. UK-UVA-Duke-Nova are all 1-seeds, but they get to pick their regions in order of the 1-68 rankings? Same thing down the line for the pool of 2-seeds, 3-seeds, etc.

This would allow for the freedom, flexibility and fun of the Titus Plan, but also provide some protection to teams trying to avoid a powerful 1-seed. The same thing would happen lower in the bracket as well, if you didn't do this. Teams that would normally be an 8 or 9 seed would be have the incentive to drop down to the 12 or 13 lines, sacrificing a more difficult 1st round game in exchange for avoiding a 1-seed in the 2nd round.

If this were actually implemented as Titus proposed, it would actually be pretty comical (and take WAY more than 3 hours) to watch the coaches to play out all the permutations and strategies. You see what college coaches already do with a handful of timeouts in the last minute of a game!

Love the idea, overall.

That would work too, though there would still be some inherent skewing at each level away from a runaway #1 team like Kentucky (though I bet much less so in the seed ranges that wouldn't be facing them until the Elite 8).

I took a rough crack at how it might work out using Lunardi's current ranking of 1-68. I definitely had a few teams make moves, even down a seed line, to stay away from Kentucky. Other than it being a bit easy for Kentucky -- but, the truth is they are so dominant, pretty much any region anyone sets up will look "easy" for them -- I think it comes out not too bad (and I just kind of get the sense that clubs like Xavier and Cincy might not be scared to take a shot at UK from the 8-9 spots, since Kentucky won't schedule them in the regular season.

East
1 Virginia-No. Fla/Charls. So. (Charlotte)
8 St John’s-Ole Miss

4 Wichita St.- La Tech (Jacksonville)
5 Louisville-Murray St.

2 Villanova- Albany (Pittsburgh)
7 Mich St.-LSU

3 Utah-N. Mex St. (Port)
6 West Virginia-Colo St.

South
1 Duke-St Francis (Charlotte)
8 Ohio St.-NC St.

4 No. Iowa-Harvard (Jacksonville)
5 Arkansas-Wofford

2 Gonzaga-So. Dakota St. (Sea)
7 SDSU-Oregon

3 Kansas- Cent. Mich (Omaha)
6 Butler-BYU/Davidson

Midwest
1 Kentucky-Sac St./Texas So. (Louisville)
8 Xavier-Cinn

4 Baylor-UC Davis (Seattle)
5 Providence-Boise St./Temple

2 Maryland-Will. & Mary (Pittsburgh)
7 VCU-Okla St.

3 Iowa St.-Georgia St. (Omaha)
6 SMU-Purdue

West
1 Wisc-Bucknell (Louisville)
8 Dayton-Georgia

4 Oklahoma-Iona (Columbus)
5 N. Carolina-Stephen F. Austin

2 Arizona-NCCU (Portland)
7 Iowa-Texas A&M

3 Notre Dame- Valpo(Columbus)
6 Georgetown-Indiana

[sagegrouse]

An auction would add some interest to the event, but -- unless by (some miracle) we get an inter-conference schedule that enables us to compare teams across conferences -- I would toss the basketball to the conferences. Here's how:

Let the top 5-6 conferences automatically get 50 percent of their teams into the NCAA's. They know who's good because there have been -- for a 12-team conference -- around 120 games among the conference teams. The conferences themselves decide who's in and who's out and by whatever system they want to use. If you aren't in the top one-half of your conference, you don't make the playoffs. Sure, there'll be blood-letting, but who cares? The odd-man ACC gets seven bids one year and eight the next; the rest of the major conferences, as of now, have an even number of teams, although usually not the number in the name of the conference.

Then I would give a dwindling number of automatic bids for the next few conferences, who would also select their own teams.

Then the remainder will have one automatic bid.

To seed the teams, I would probably use some combination formula like college football does, mixing polls and computer rankings.

Now, how do you determine the top five or six conferences? Easy. I would use the number of wins in the previous five NCAA tournaments. The allows other conferences the chance to move up in the conference rankings, but it won't be easy.

This process is a system like the U.S. Constitution, with a lot of pieces and with responsibility spread out over many different institutions. The current system, the Tournament Selection Committee, is some monarchical, authoritarian institution that is positively un-American. The TSC is like the old Privy Council of England and of Scotland or the Roman Curia, which may now be reined in by Pope Francis. We didn't need the TSC, and we don't need it.

I'll check the wins by conference and come back with a practical application.

[Wander]

Guys, I love you, but these are all terrible ideas.

(except for sagegrouse's proposal to have a designated non-conference week in February - that would be super cool).

I agree the auction thing would make for some fun TV but ultimately doesn't even make sense. The big advantage in picking your spot happens only if most of the bracket is filled out already - so you get to decide who to play - meaning Kentucky would pick close to last, so other teams don't get the chance to "avoid Kentucky." Kentucky would get to decide who they want to avoid. We don't need that, plus it would throw out some of the best practices of the tournament, like not letting teams play conference teams in the first or second round (when permitted).

Sagegrouse, your non-conference week is a fantastic idea, but I'm sorry: the stuff about all these automatic bids for power conferences is just beyond awful. What you're doing is importing the absolute worst part of college football - the idea that what your conference has done historically somehow matters in what your team does this season.

Right now, for example, your system would end up putting Arizona State and Florida into the field. What if some random Missouri Valley team wins that conference tournament? It happens every year to a couple conferences. You're going to put in a Florida team that has a LOSING RECORD over either Wichita State or Northern Iowa, both top 15 teams?

I'll ask again: if the selection of teams is such a crapshoot, why is it so easy for everyone to correctly predict the vast majority of the teams - sometimes ALL the teams - that make the field every single year?

[sagegrouse]

Guys, I love you, but these are all terrible ideas.

(except for sagegrouse's proposal to have a designated non-conference week in February - that would be super cool).

...................................

Sagegrouse, your non-conference week is a fantastic idea, but I'm sorry: the stuff about all these automatic bids for power conferences is just beyond awful. What you're doing is importing the absolute worst part of college football - the idea that what your conference has done historically somehow matters in what your team does this season.

Right now, for example, your system would end up putting Arizona State and Florida into the field. What if some random Missouri Valley team wins that conference tournament? It happens every year to a couple conferences. You're going to put in a Florida team that has a LOSING RECORD over either Wichita State or Northern Iowa, both top 15 teams?

I'll ask again: if the selection of teams is such a crapshoot, why is it so easy for everyone to correctly predict the vast majority of the teams - sometimes ALL the teams - that make the field every single year?

Wander: The fact that everyone can pick the NCAA field is an example of the "reliability" of the selection process. But "reliability" and "validity" are not the same thing. The NCAA process has a hard team being "valid" with so little data to compare results across conferences, and none of it occurring within 75 days of the NCAA tournament (Duke-St. John's excepted).

I am busy assembling NCAA tournament wins by conference -- which is only one of the possible measures -- although it will take some work because of all the changes in conference membership.

Thanks for your kind words on "inter-conference week." You do realize that my extreme NCAA Tournament proposals are partially aimed at getting the NCAA and the sports channels to take action to enrich the poverty-stricken out-of-conference schedules of teams in the major conferences. They are a disgrace to the concept of competition in athletics.

[sagegrouse]

Here's John Feinstein's take (http://www.washingtonpost.com/news/sports/wp/2015/03/16/ncaa-basketball-committee-needs-to-be-fixed-heres-how/) on the ineffectiveness of the Tournament Selection Committee. His make -- if the TSC consisted of basketball coaches and experts, this would be a trivial problem to solve. Only one member of the TSC was ever a college basketball coach -- Joe Roby -- the Northeastern AD, who coached at Harvard.


As for the [selection] rules, there should be none. Forget all the silly talk about RPI and SOS and away wins and wins against the top 50 or top 100 or top million. Forget the numbers. They’re nothing but a crutch and an existence justification for the bracketologists, who should be banned from the earth by an act of Congress or a Presidential Executive Order.

No rules, just basketball. It’s wonderful that committee members cite the “eye-test” as a reason for picking a team. The problem is their eyes don’t know anything about basketball. Put the committee in the hands of those who know the game and give them one marching order: Pick the best field. Watch basketball — double-check numbers and away wins for backup if you want – and just pick the best teams.

And his candidates for the TSC:


Who should be on the committee? To name a few: Gary Williams; John Thompson Jr.; Denny Crum; Sonny Smith; Wimp Sanderson; Frank Sullivan; Nolan Richardson; Bobby Cremins; Terry Holland (who was the committee chairman in 1997 when he was an AD) and Mike Montgomery. All those men have extensive coaching experience and have had a lot of success.
Plus a few others from the press.

[Reilly]

We don't need some JF-designated panel of experts to do this right. Just let the kenpom rankings invite the field.

32 teams get automatic bids.

36 get invited. If kenpom did the inviting this year (the 36 highest rated non-automatic-qualifiers), if I'm reading his chart right, we'd have almost the exact same field. We'd lose Purdue and Indiana (quite a hit for the state of Indiana; wheres's the NCAA based?) and instead have Florida and Vandy. Yeah, yeah, Florida is 16-17. If you want a "no team with a losing record gets an at-large invite rule" then we'd have Vandy and Tex A&M rather than Purdue and Indiana ... and Stanford would've been the first team left out.

Do we really need *weeks* of Joe Lunardi nonsense, and a bunch of guys camped out in a hotel conference room, to decide between Purdue/Indiana on the one hand, and Florida/Vandy/TexA&M on the other? Cuz that's what it comes down to. Everybody knows the great, great majority of teams who should get invited. Just let KenPom do it. Then have a committee to seed it, trying to keep folks close to home, or maybe even *trying* to create cool story lines (instead of pretending they pay no attention to such considerations). It's not rocket science, and we don't need Feinstein-designated-demigods to populate a worthy field.

[Henderson]

We don't need some JF-designated panel of experts to do this right. Just let the kenpom rankings invite the field.

32 teams get automatic bids.

36 get invited. If kenpom did the inviting this year (the 36 highest rated non-automatic-qualifiers), if I'm reading his chart right, we'd have almost the exact same field. We'd lose Purdue and Indiana (quite a hit for the state of Indiana; wheres's the NCAA based?) and instead have Florida and Vandy. Yeah, yeah, Florida is 16-17. If you want a "no team with a losing record gets an at-large invite rule" then we'd have Vandy and Tex A&M rather than Purdue and Indiana ... and Stanford would've been the first team left out.

Do we really need *weeks* of Joe Lunardi nonsense, and a bunch of guys camped out in a hotel conference room, to decide between Purdue/Indiana on the one hand, and Florida/Vandy/TexA&M on the other? Cuz that's what it comes down to. Everybody knows the great, great majority of teams who should get invited. Just let KenPom do it. Then have a committee to seed it, trying to keep folks close to home, or maybe even *trying* to create cool story lines (instead of pretending they pay no attention to such considerations). It's not rocket science, and we don't need Feinstein-designated-demigods to populate a worthy field.

What committee determines the KenPom model that sets the field? Or does Mr. Pomeroy get to decide that?

[sagegrouse]

We don't need some JF-designated panel of experts to do this right. Just let the kenpom rankings invite the field.

32 teams get automatic bids.

36 get invited. If kenpom did the inviting this year (the 36 highest rated non-automatic-qualifiers), if I'm reading his chart right, we'd have almost the exact same field. We'd lose Purdue and Indiana (quite a hit for the state of Indiana; wheres's the NCAA based?) and instead have Florida and Vandy. Yeah, yeah, Florida is 16-17. If you want a "no team with a losing record gets an at-large invite rule" then we'd have Vandy and Tex A&M rather than Purdue and Indiana ... and Stanford would've been the first team left out.

Do we really need *weeks* of Joe Lunardi nonsense, and a bunch of guys camped out in a hotel conference room, to decide between Purdue/Indiana on the one hand, and Florida/Vandy/TexA&M on the other? Cuz that's what it comes down to. Everybody knows the great, great majority of teams who should get invited. Just let KenPom do it. Then have a committee to seed it, trying to keep folks close to home, or maybe even *trying* to create cool story lines (instead of pretending they pay no attention to such considerations). It's not rocket science, and we don't need Feinstein-designated-demigods to populate a worthy field.

First, because KenPom takes the entire season of basketball and reduces it to four scalar values for any game (plus a home field margin, if appropriate). Even if these are the optimally determined four scalars, basketball involves many, many more values and variables -- some explicitly measured and some tacit.

I am wiling to concede that he has good and useful models based on some reasonably constructed relationships. BUT THEY ARE JUST FOUR SCALAR VALUES. I am not willing to concede that basketball is best described by these values.

Second, he has no more basis for comparing results between conferences than the benighted TSC that both Feinstein and I despise. There are essentially no inter-conference games after the first couple of days of January; therefore, there is no true quantitative basis for comparing results across conferences. Let me give you a simple example -- although not an entirely simplistic one. There are two groups of basketball teams -- say, Atlantic and Pacific that play almost exclusively within their own region. There is one game per year between an Atlantic team and a Pacific team, but no others. KenPom and Sagarin, therefore, would have a numerical basis for constructing a rating system for all teams. But it is essentially meaningless, since there was only one game to set the alignment between Atlantic and Pacific teams. Thus, the 2.5 month absence of inter-conference games weakens any quantitative measure of the teams BETWEEN conferences, especially in regards to their strength at the end of the season. It doesn't matter which equations or estimating techniques are used by Pomeroy, Sagarin or 538 -- they got very little meaningful data. Capiche?

[Reilly]

First, because KenPom takes the entire season of basketball and reduces it to four scalar values for any game (plus a home field margin, if appropriate). Even if these are the optimally determined four scalars, basketball involves many, many more values and variables -- some explicitly measured and some tacit.

I am wiling to concede that he has good and useful models based on some reasonably constructed relationships. BUT THEY ARE JUST FOUR SCALAR VALUES. I am not willing to concede that basketball is best described by these values.

Second, he has no more basis for comparing results between conferences than the benighted TSC that both Feinstein and I despise. There are essentially no inter-conference games after the first couple of days of January; therefore, there is no true quantitative basis for comparing results across conferences. Let me give you a simple example -- although not an entirely simplistic one. There are two groups of basketball teams -- say, Atlantic and Pacific that play almost exclusively within their own region. There is one game per year between an Atlantic team and a Pacific team, but no others. KenPom and Sagarin, therefore, would have a numerical basis for constructing a rating system for all teams. But it is essentially meaningless, since there was only one game to set the alignment between Atlantic and Pacific teams. Thus, the 2.5 month absence of inter-conference games weakens any quantitative measure of the teams BETWEEN conferences, especially in regards to their strength at the end of the season. It doesn't matter which equations or estimating techniques are used by Pomeroy, Sagarin or 538 -- they got very little meaningful data. Capiche?

Sage, who are the worthies -- this year, or any year -- who have been left out of the dance? And how many injustices have been wrought?

We have 32 conferences and each determines its champion. That's cool -- be the best on your block, in your neighborhood, among those who you have thrown your lot in with.

Then, we want to invite some other worthies. OK. Do you really think Florida, Vandy, TexA&M, and Stanford (the first four KenPom teams left out) are better than Virginia, Duke, Utah and Oklahoma (the first four KenPom teams invited at-large)? We didn't get this decision right because we don't have enough linked data? This year's Florida and Vandy are better than Virginia and Duke? Of course you don't believe that. OK, move on to the next four at-large invited: Kansas, Wichita State, Baylor, and UNC. Keep on going down the line ... When does it start to get murky or questionable?

My point: we have enough data to get the worthies into the field. You want to play in the NCAA? Win your conference. You want an at-large invite? Be clearly good enough.

We are arguing at the far margins here: Florida and Vandy v. Texas A&M and Stanford v. Purdue and Indiana. Who really cares all that much? Do we really think -- in the real world -- that some super-deserving team has been slighted by our use of inappropriately linked data?

Henderson, it's the NCAA's tourney, so the NCAA would determine how to select its invitees -- just as it does now. I'm suggesting that we have models already that get us 95+% of the way to a very worthy tourney, without the human drama. Yes, the human drama comes in picking in a formula.

Make an invite objective (win your conference; or finish ___ in some objective poll). High schools use computer formulas to populate playoff fields. Nobody really deserving will be omitted, in my opinion.

[CDu]

I kind of agree with reilly's point. The only thing being debated here is the quality of the 34th/35th/36th/37th/38th/39th best teams that didn't win their conference. If you are considered by most metrics to be on the border of the top-50 teams in the country (and that's who we're talking about here), do you really have a gripe if you aren't invited to the tournament?

Of course, I'm one of those who thinks the tourney is TOO inclusive already. We don't really need 20% of Division-1 schools participating. It's fun, but it is just added noise. Not that we're going to have less than 68 teams in the tournament anytime soon (or perhaps ever again in my lifetime). But given that we have 68 teams, I don't see any problem with having the last few at-large teams be defined by some objective criteria rather than a Selection Committee. Even if the criteria have inherent limitations. As Reilly said, if you want to make sure you don't get left out, don't be the 50th best team in the country. And if you are the 50th best team in the country, don't lose your conference championship.

[FerryFor50]

I kind of agree with reilly's point. The only thing being debated here is the quality of the 34th/35th/36th/37th/38th/39th best teams that didn't win their conference. If you are considered by most metrics to be on the border of the top-50 teams in the country (and that's who we're talking about here), do you really have a gripe if you aren't invited to the tournament?

Of course, I'm one of those who thinks the tourney is TOO inclusive already. We don't really need 20% of Division-1 schools participating. It's fun, but it is just added noise. Not that we're going to have less than 68 teams in the tournament anytime soon (or perhaps ever again in my lifetime). But given that we have 68 teams, I don't see any problem with having the last few at-large teams be defined by some objective criteria rather than a Selection Committee. Even if the criteria have inherent limitations. As Reilly said, if you want to make sure you don't get left out, don't be the 50th best team in the country. And if you are the 50th best team in the country, don't lose your conference championship.

That, and, if you want a chance at playing (and beating) the top teams in the country in the NCAA tournament, perhaps try doing that in the regular season first. Don't beat up on the little guys all year and demand a shot at the big guys that you really haven't earned.

[Reilly]

... The only thing being debated here is the quality of the 34th/35th/36th/37th/38th/39th best teams that didn't win their conference ...

That's sort of how I see it b/c I mostly trust these Kenpom numbers as they accord with my own subjective assessment watching folks play over the years.

Sage can certainly speak for himself. I understand him to be saying "how do we know we are only arguing about the 34th best at-large team, because we have unreliable data?"

That is, Kenpom shows ...

#9 Oklahoma is 22-10
#49 Stanford is 20-13
#71 GWU is 22-12
#102 UTEP is 22-11
#139 W.KY is 20-12
#164 Montana is 20-13
#185 Norfolk State is 20-14
#212 Gardner-Webb is 20-14

These teams don't play each other, and their opponents don't all play each other, so we have unreliable data, and maybe GWU is #9 and Oklahoma is #164 ... or something ...

I follow Duke closely, and where the SRS sports-reference.com numbers and KenPom numbers slot Duke every year squares with my impressions of the various Duke teams over the years and how good/strong I think they are.

I don't understand the numbers and formulas, and I cannot speak to the reliability of the data, but the rankings (Kentucky at #1 and 2-27 Grambling State at #351) end up making sense to me ... so based on my own eye test over the years, I trust the numbers (though I don't understand them).