jimrtex
All American
Posts: 2,577
Joined: Aug 2021
Reputation: 263
I Root For: Houston, Tulsa, Colorado
Location: |
RE: 96-team tournament.
(03-24-2022 12:41 AM)Milwaukee Wrote: (03-23-2022 11:00 PM)jimrtex Wrote: This is a possible format for a 96-team NCAA Tournament by which all teams are selected by objective criteria, not by a selection committee.
The 20 lowest-ranked conferences get two AQ. The top 12 conferences (ACC, B1G, B12, SEC, P12, B East, A100, AAC, WCC, MVC, MtW, and CUSA) share 56 AQ. The allocation of those 56 AQ is determined by tournaments played in December-January. Teams would play 3-games with a full championship and consolation bracket. The conference would be awarded one point for each win, with a bonus point for the tournament winner.
I simulated the results of these tournaments using KenPom ratings with randomized results. Teams were assigned randomly to these tournaments, making sure that there was only one team from a conference in any tournament.
If the 19 tournaments were spread out over five weeks, that would be four tournaments per week. If some tournaments were Tuesday-Thursday-Saturday and others Wednesday-Friday-Sunday, most games could be televised with two networks...
There seems to be a lot of support for the idea of an expanded, 96-team NCAA tournament, and trying to figure out ways to replace or minimize the role of the selection committee and to devise another method for selecting and scheduling teams.
Regarding the specifics in your post:
1) At most, there might be enough support for granting automatic bids to all the regular season and conference tourney champs in a 96-team NCAA, but there would probably be strong resistance to granting every conference a second auto-bid if their regular season champ were to win their conference tourney.
Each conference would determine who received their berths, or rather they would determine their criteria before the season. That is, entrants in the tournament would be determined on the court - not by some committee where bias is unavoidable. The number of berths for each conference would have to be fixed, otherwise you would have to award conditional berths to other conferences
For example if the American East only has one berth because Vermont was both the regular season and tournament champion which conference gets another berth? You end up with the same problem you have now where a lucky tournament winner takes a berth away from another conference.
2) The idea of replacing seedings has a great deal of merit, but the idea of holding mini-tournaments earlier in the season seems overly cumbersome and a potential scheduling nightmare, so other alternatives might be worth considering.
I have not replaced seedings for the tournament itself. The Conference Challenge tournaments would not be seeded (I suppose they could be, but in my simulations they were not). There could also be some spreading out of the talent. For example the Sweet 16 (15, excluding St.Peter's) could be guaranteed a tournament without another Sweet 16 team. So for example, you could not end up with Arizona, Kansas, and Gonzaga in the same tournament.
This tournament for the Top 12 conferences, there are 15 Sweet 16 teams (all but St. Peter's), 14 2nd Round losers (all but Murray State and New Mexico State) and 19 teams that lost in the first round or play-in).
So you have one vase with balls containing numbers 1 to 19, and a second vase with balls containing the names of the 15 Sweet Sixteen Teams from the Top 12. You draw one ball from each vase and assign the teams to a tournament group. After the Team vase is empty, you will still have 4 balls remaining in the Tournament vase. You bring out the vase containing balls for the 14 Second Round losers, and so on.
I'm assuming you have watched a World Cup draw.
This draw could be held immediately after the previous tournament. I'm assuming that basketball schedules for next season have not been finalized or even all the opponents determined. I don't know how the sites and dates for the tournaments would be determined. Perhaps they could be bid out.
The groups might be assigned based on proximity of the teams to the site. If you grouped the conferences West: P12, WCC, MtW; Midwest: B1G, B12, MVC; South: SEC, CUSA, AAC; and East: ACC, Big East, and A10. Then if a tournament group had three teams from a region, it could be assigned a site in that region.
Here is the rationale for Challenge Tournaments: How do you compare conferences? The only proper way is to have the conferences play games against each other. The 219 games will also make the NET ratings more reliable.
The draw could take place on Sunday of the Final Four. You could do like for the tournament selection show and have remotes to each school, with cheerleaders, pep band, mascots, and players.
3) The following is one alternative method for replacing selection committee seedings with a computerized procedure that would schedule tournament matchups based on the teams' geographical proximity to the regional and sub-regional sites, with algorithms minimizing the number of games that would be played between teams that have already played 1 or 2 games earlier in the season:
- All 96 teams would be assigned to one of eight sub-regional sites (12 teams per site) based on geographic proximity. For example, using the 2022 sites, there would be 12 teams from the northwestern U.S. competing in Portland.
.
---With all teams in each sub-region being relatively close to the site where the games would be played, more fans could attend the play-in games, boosting tournament attendance.
.
- The 32 regular season champions (4 per sub-region, based on proximity) would get a first-round bye, based on the rationale that teams that fail win their regular season conference championship shouldn't be favored to win the national championship.
.
---For example, the 4 regular season champs whose campuses are located nearest to the NW sub-region would play the winners of the 4 Portland play-in games.
.
- Rather than being seeded, the matchups for the 4 play-in games in each sub-region could be selected by computer programs that would minimize the number of rematches of teams that have played earlier in the season. Ties would be broken using a randomization algorithm. "Full round 1" games vs. the 4 sub-regional conference champs would be scheduled the same way.
.
---For example, if two teams from one conference (e.g., BYU & San Francisco) that had already played twice that season were among the Portland play-in teams, they would not play each other in a play-in game, and the reg. season champ of the WCC conference (Gonzaga) would not play either BYU or SF in round 1.
.
- The 4 teams that advance from each sub-region would compete for the four regional championships in a series of three games. The matchups would be determined by computer programs that would (a) minimize home court advantage by minimizing the number of games between teams nearest and farthest from the four regional sites, and (b) minimize the number of rematches of teams that have played earlier in the season, with ties being broken via randomization.
.
---For example, if San Francisco were the regional site, matchups would be scheduled pitting the most distant teams against each other and pitting the nearest teams against each other, while minimizing the number of matchups of teams that have played each other earlier in the season.
.
---To further illustrate: If teams such as Arizona and UCLA were to advance to the Regional quarterfinals, they would not be matched up in a way that would make it likely for them to play each other until the final round of the West (San Francisco) Regional, and the computer would also minimize the probability that they would play another team from their conference in the West Region semifinals or finals. This would make it unlikely that Arizona and UCLA would play each other before the West Region final.
In my scheme, there would be 4 #1 seeds, 4 #2 seeds, ... 4 #8 seeds, then 8 #9 seeds, 8 #10 seeds, ... 8 #16 seeds. This recognizes that it is harder to distinguish lower seeds, and that many of the teams will come from lower ranked conferences.
My first round groups will either be 8 4-teams groups: #1 vs #8, #4 vs. #5; or #2 vs. #7, #3 vs. #6. Just like now they will play two games the first week on Thursday-Saturday or Friday-Sunday.
The 8 8-teams groups will be: #9 v. #16, #12 v. #13, #11 v. 14, #10 v. 15. They will be play either Tuesday-Thursday-Saturday or Wednesday-Friday-Sunday. It would be possible to have an 8-team group playing at the same location as a 4-team group. You would have 4 games on Tuesday/Wednesday, 4 games on Thursday/Friday and 2 games on Saturday/Sunday.
4) Since there would be no need for a selection committee to seed the teams, the option would exist to completely eliminate the need for a selection committee, altogether.
---The decisions that have been made by selection committees could easily be done using computer programs that would factor in all the determinants (e.g., NET rankings, "Q1, Q2 records," kenpom.com rankings, etc.) that the selection committees have used through the years. Such computer algorithms could also adjust the number of at-large bids to be assigned per conference, to make it less likely that a conference such as the Big Ten would continue to get a disproportionate share of at-large bids.
The Challenge Tournaments replace the need to adjust the number of berths based on some sort of rating system.
(03-23-2022 11:45 PM)IWokeUpLikeThis Wrote: I think you’d see 32 AQ and 64 at-large.
32 AQ and 64 at-large seems likely, with the crux of the discussion being about whether favoritism should continue to be given to the top-ranked teams or to the 32 conference champions.
(03-23-2022 11:45 PM)IWokeUpLikeThis Wrote: The 96-team format was discussed around 2011 when the Turner deal happened but tv wasn’t willing to pay extra for it.
When one considers the massive viewership numbers that the tournament generates and the increasingly deep pockets of the networks, it seems unlikely that any reservations that the broadcasters may have had in 2011 or 2012 would continue into the mid-2020's.
.
|
|
