What would be the impact of an expanded NCAA Tournament?
Rumors are swirling once again about expanding March Madness. While it remains one of the worst ideas ever proposed, what would happen if the NCAA were to expand?
After a quiet summer and fall, dark clouds are once again forming near the horizon of the college basketball world. Certain perceived power brokers are once again discussing the possibility to expand the NCAA Men's Basketball Tournament.
I have made my personal opinions on this matter crystal clear: it is an objectively terrible idea. Any person in a decision-making role who advocates for such expansion should be publicly and repeated ridiculed until forced out of the sport.
I have written previously about the many reasons why expansion is such a bad idea. Briefly, there is no evidence to support the idea that the current format does not allow in enough worthy teams. Instead, expansion will almost exclusively benefit unworthy, mediocre high-major teams at the expense of mid- and low-major teams.
Expansion also will disrupt the current schedule and popular activities such as office pools. This very likely will impact fans enjoyment of the tournament. To this end, in a world where people can barely agree on the color of the sky, expansion is a concept that is perhaps nearly universally despised by many college basketball fans.
It is not apocryphal to suggest that expansion could eventually decrease viewership, costing television networks billions of dollars. Did that get your attention CBS/Paramount Executives?
But the complete lack of merit for expansion it not the purpose of today's contribution. Instead, I would like to analyze the likely result of the proposed expansion, if it were to happen.
The 76-Team Proposal
In the spring of 2025, the expansion discussion centered around a possible expansion of the current 68-team field to 72 teams. The four additional teams would be drawn from the pool of "at-large" teams that did not receive an automatic to the NCAA Tournament as a result of winning a conference tournament.
The addition of four teams on wrong side of the current NCAA bubble would result in the addition of four additional mid-week "play-in" games prior to the traditional start of first round of the NCAA Tournament on Thursday that for mathematical and logistical simplicity almost must have exactly 64 teams.
Back in the spring, no decision was made regarding expansion. Surprisingly, when rumors of expansion resurfaced this fall, the proposed expansion had ballooned to 76 teams.
In this new proposal, 24 teams of the 76-member field would need to start play mid-week in order to pare the field down to 64 teams by Thursday. This means that a total of 12 games would need to be played on the Tuesday and Wednesday between Selection Sunday and the start of the first round on Thursday.
In principle, these 12 games could be played at three different sites on Tuesday and Wednesday in the evening. Three games could be played in parallel on three different networks or streaming platforms. However, the proposals that I have seen suggest that only two sites will be used, which necessarily means that likely four of the 12 games would be played in the afternoon on that Tuesday and Wednesday.
Serious basketball fans already spent time following afternoon conference tournament games the week prior. Many serious fans (myself included) take time off of work to watch first round NCAA Tournament games on Thursday and Friday. These same fans will almost certainly not make time to watch the new play-in round Tuesday/Wednesday games. Viewership could be terrible.
Furthermore, the existence of 12 total play-in games means that 12 of the 64 slots in a standard office pool bracket (almost 20%) would contain an "either/or" pair of teams. The presence of just four of these slots in the current format already causes significant confusion and frustration with casual fans. The 76-team proposal would triple this negative impact.
Which New Teams Would Get In?
As mentioned above, an expansion to 76 teams would allow eight additional teams to play in the tournament, which otherwise would have not have made the field. What would the profile of these additional teams likely look like?
NCAA Tournament Selection and seeding is a highly controversial subject, so it is impossible to know for sure. But if we use the 2025 NCAA Tournament candidate field as a reference, we can get an idea of which teams would have been added.
Based on the available metrics, such as the NET, Kenpom efficiency, and my personal seeding algorithm, here is a list of the most likely additional teams, if the 2025 NCAA Tournament would have expanded by eight slots:
Indiana: 19-13, No. 9 in Big Ten (10-11), lost to No. 8 Oregon in Big Ten Tournament
West Virginia: 19-13, No. 8 in Big 12 (11-11), lost to No. 16 Colorado in Big 12 Tournament
Ohio State: 17-15, No 10 in Big Ten (9-12), lost to No. 15 Iowa in Big Ten Tournament
Boise State: 24-10, No. 5 in the Mountain West (16-7), lost to No. 1 New Mexico in Mountain West Tournament
SMU: 23-10, No. 6 in the ACC (14-8), lost to No. 3 Clemson in ACC Tournament
San Francisco: 24-9, No. 3 in the West Coast Conference (14-6), lost to No. 1 Gonzaga in the WCC Tournament
Wake Forest: 21-11. No. 4 in the ACC (13-8), lost to No. 5 North Carolina in ACC Tournament
UC Irvine: 28-6, No. 2 in the Big West (18-4), lost to No. 1 UC San Diego in the Big West Tournament
Nebraska: 17-14, No. 16 in the Big Ten (7-13), did not qualify for the Big Ten Tournament
Northwestern: 17-16, No. 13 in Big Ten (8-14), lost to No. 5 Wisconsin in Big Ten Tournament
Cincinnati: 18-15, No. 13 in the Big 12, (8-14), lost to No. 5 Iowa State in Big 12 Tournament
Villanova: 19-14, No. 6 in the Big East (12-10), lost to No. 3 UConn in Big East Tournament
Eight of the 12 teams above would have made the NCAA Tournament.
Six of the 12 teams finished at or below .500 in conference play. Seven of the 12 teams finished with at least 13 losses. The two ACC teams (SMU and Wake Forest) combined for just two "Quad One" wins and both had strengths of schedules outside of the top 50.
The list does contain a total of three mid-major teams: Boise State, San Francisco and UC Irvine. The addition of these potential Cinderella teams into the bracket is one of the few potentially compelling arguments for expansion.
But, let's be serious. The stated purpose of expansion from conference leaders such as SEC Commissioner Greg Sankey and Big Ten Commissioner Tony Petitti is to get more of their teams into the tournament. Can we honestly believe that more that one of the mid-majors above would have been included in this theoretically expanded field? I think not.
Seeding Implications
The general structure of a hypothetical 76-team NCAA Tournament was described above, but there are some additional subtleties in how this proposed tournament would be constructed. It is certain that 24 teams would need to play in 12 games prior to the first round of the tournament.
But which teams from the field of 76 would play in the expanded play-in round and how would the winning teams be incorporated into the main bracket?
There are several ways in which this could be accomplished, but two possible strategies have been outlined in the discussion last fall.
The most likely strategy was shared by Ross Dellenger on the On3 network last October. Half of the 12 play-in games would involve the 12 lowest-ranked teams in the tournament. This pool of teams are exclusively made up from conference champions of mid- and low-major conferences. In the current format, only four teams from this pool are forced to participate in the play-in round. The new proposal triples that number.
The other half of the 12 play-in games would be filled by the 12 lowest-ranked at-large teams. Basically all eight of the newly added teams would participate in the play-in round, along with the four teams that currently participate in the play-in games. Note that prior to 2010, only one of those 12 teams (the highest ranked) would have qualified for the NCAA Tournament at all.
But Big Ten Commissioner Tony Petitti unveiled a new idea last fall at Big Ten Basketball Media Day. He suggested an alternative model based on what he referred to as "straight seeding." His proposal is that the 24 lowest-ranked teams in the NCAA Tournament would all be moved into the play-in round. This group would include almost all of the automatic qualifiers from the mid- and low-major conferences.
In this scenario, the majority of the newly-added bubble teams that previously would not have qualified for the NCAA Tournament would instead be placed directly in the main portion of the tournament bracket in the first round games played on either Thursday or Friday.
Both 76-team proposals would result in a major shift in the way the NCAA Tournament is constructed.
The biggest impact would be a general compression and shift of the teams seeded No. 13 and below. In the mostly plausible scenario, all of the current No. 16 seeds and half of the current No. 15 seeds would shift into No. 16 play-in games. All of the current No. 14 seeds would morph into teams playing for the two remaining No. 15 seeds. All of the current No. 13 seeds would become No. 14 seeds.
Last March, No. 13 McNeese State upset No. 4 Clemson in the first round. In the expanded Tournament, McNeese State would have been a No. 14 seed.
No. 15 seed Bryant pushed No. 2 Michigan State in the first round. Bryant would slide to a No. 16 seed as well.
In Petitti's more aggressive reseeding proposal, this shift and compression is even more extreme. Up to 12 additional mid-major teams would effectively be shifted down the bracket. All four No. 14 seeds, No. 15 seeds and No. 16 seeds in the main bracket would originate in a play-in game.
A dangerous team such as McNeese State would have wound up in a No. 14 seed play-in game rather than as a No. 12 seed. A team such as Grand Canyon would have played in a No. 15 seed play-in game rather than as a No. 13 seed.
The full pool of No. 12 seeds and No. 13 seeds would almost entirely be composed of teams that previously would not have qualified to make the Tournament at all.
Quantifying the Madness
The analysis above provides a qualitative look at the impact of possible NCAA Tournament analysis. These changes will also have a significant impact on the very nature of the tournament in ways that cannot be measured qualitatively.
In order to clarify the more tangible impacts of expansion, I developed a more precise method to simulate the results of the NCAA Men's Basketball Tournament in each of the scenarios described above.
I used historical Kenpom Efficiency margin data and the lists of fully ranked seed lines from the NCAA Selection Committee back to 2011 and developed a correlation between the two data sets. That correlation is shown below in Figure 1.
Figure 1: Correlation between average Kenpom Efficiency Margins and the published seed line position (No. 1 to No. 68) for the NCAA Tournaments from 2011 to 2025
With this correlation in hand, it is trivial to set up a set of NCAA Tournament brackets where the participants represent perfectly average teams seeded in a perfect way.
The seed list is the perfected ordered with the best team in the NCAA Tournament as the No. 1 overall seed all the way down to the worst team in the tournament at the bottom of the list, which contains 64, 65, 68, 72 or 76 teams.
Seeding also follows a perfect "s-curve" where at every round of the tournament, the best team faces the worst remaining team, the second-best team faces the second-worst team, and so on, assuming the favored team wins every game.
Using this mathematical model, Kenpom data can be used to calculate the odds for the favored team to win any possible NCAA Tournament matchup. The odds for various tournament outcomes can then be compared between any of the past and proposed tournament scenarios.
This set of calculations produces a ton of data that reveals a great deal about the impact of potential NCAA Tournament expansion. Figure 2 shows the impact of expansion on first-round upset rate when five different tournament structures are considered. For reference, the data points show the observed upset rate for the 1985 to 2025 tournaments.
Figure 2: Calculated first round upset rate for six different tournament structures (solid lines). The data points (circles) represent the actual upset rate from 1985 through 2025.
The Figure above highlights some basic rules of thumb that hardcore "March Madness" nerds know well. No. 1 seeds and No. 2 seeds both win over 90% of the time. No. 3 and No. 4 seeds win over 80% of the time. No. 5 seeds and higher experience upsets much more frequently.
Note that the actual upset frequencies agree well with the model predictions, which some variation No. 5 seed and No. 8 seed get upset more frequently than expected. No. 7 seeds have historically avoided upsets more than expected.
This type of variance is completely expected. The model assumes a highly idealized version of the real tournament bracket. But the predicted trends in upset frequencies for the expanded tournament are interesting.
As Figure 2 shows, the modest expansion from 64 to 65 and finally to 68 teams is projected to have a small impact on upset rates. This impact is on the order of two or three percentage points at the most and it is fairly consistent from No. 1 seeds down to No. 6 seeds.
If the tournament were expand to 72 or 76 team is expected to result in a significant increase in first round upsets. The magnitude of this impact is five to 10 percentage points. It is especially significant for the No. 3 and the No. 4 seeds.
Expansion will not impact seeds No. 7 to No. 10. Teams in this range will essentially play exactly the same teams in a 72- or 76-team tournament than they would in a 64-team field.
Another way to look at the data presented in Figure 2 is from the point of view of the number of years between first-round upsets of various seeds. Figure 3 gives the predicted number of years in between "major" upsets involving No. 1 to No. 5 seeds.
Figure 3: The predicted number of years between major first round upsets (seeds No. 1 to No. 5) in each of the six highlighted NCAA Tournament configurations.
One interesting note is that when the NCAA Tournament expanded to 68 teams in 2011, there was a significant jump in the expected upset rate of the top seeds. No. 1 seeds went from an expected upset rate of once about every 21 years to once in 14.5 years.
In the proposed expansions beyond the current 68-team format, that rate is expected to increase to once in 12.6 years. This is due to completely to the fact that adding additional play-in games to the bottom of the bracket increases the average strength of the No. 16 seeds.
The impact to the No. 2 seeds is even more significant. In the 64- and 68-team brackets, No. 2 seeds are expected to lose about once in 4.5 to six years. If the NCAA Tournament is expanded to 72 or 75 teams, that rate will almost double to once every 2.5 to 3.5 years.
For No. 3 seeds, the upset rate is expected to increase to more often than every other year. An expanded tournament would also result in about one No. 4 seed upset each year and more than one No. 5 seed upset per year.
This change in first-round upset rates has a longer range impact. If No. 2, No. 3, No. 4 and No. 5 seeds are upset more often, teams on those seed lines will generally have less success in March than in a smaller tournament.
But the impact is not just limited to team on those seed line. Other teams on the bracket will actually have better odds to advance. If more No. 2 or No. 3 seeds get upset in the first round, the paths for No. 6 and No. 7 seeds (which would face those highly-ranked team in the second round) get easier.
Figure 4 below visualized the predicted change in Final Four participation for seeds No. 1 to No. 10 under a 72- and the two versions of 76-team expanded tournaments.
Figure 4: Change in the number of Final Four teams on each seed line for the proposed 72- and 76-team tournaments compared to the 64-team baseline.
Tournament expansion to 72 or more teams will very likely result is a subtle shift in Final Four participation and overall March success across the board in ways that are not expected.
No. 1 seeds will actually benefit from expansion. While the upset rate of No. 1 seeds will tick up, the tops seeds are still expected to win 98% of all first-round games going forward. This small increase in attrition is offset by a larger upset rate for the No. 2 to No. 5 seeds.
As Figure 4 shows, Final Four participation among No. 2 to No 5 seeds is expected to drop by up to 10%. No. 3 seeds will be hit the hardest. When fewer No. 2 and No. 3 seeds advance, the path to the Final Four for No. 1 seeds gets easier.
Teams located in the middle of the bracket and seeded No. 6 to No. 10 will also see a small benefit as the higher seeded teams are upset more frequently. No. 7 seeds are expected to make the Final Four 10% more often in the extreme case.
That said, the odds of a No. 7 seed (for example) making the Final Four will remain slim. In the 64-team bracket, a No. 7 seed has a 2.4% chance to make the Final Four (i.e. about one appearance every 10.6 years). In the 76-team brackets, those odds increase to 2.6% (i.e. one appearance every 9.4 years).
Add it Up
If the powers that be make the awful decision to expand the NCAA Tournament, fans can expect a few changes, both overt and subtle to the way "March Madness" plays out.
More obviously, a larger percentage of low- and mid-major teams will be pushed into play-in games. Half of these potential Cinderella teams will be eliminated before the start of the first round on Thursday afternoon.
But those teams that do survive the play-in games will be a bigger test for the high seeds that they do end up facing in the first round. The total number of first-round upsets will increase by about one per year (7.6 to 8.7 or slightly higher in the Big Ten proposal).
This increase in upsets will mostly impact teams seeded No. 2 to No. 5 and as a result, these teams will see a slight decrease in Final Four participation and overall Tournament success.
Conversely, No. 1 seeds and teams seeded No. 6 and lower will benefit from this subtle change and see a slight increase in Final Four participation and overall NCAA Tournament success.
