Don't Expand the NCAA Tournament

Last weekend, in what has become an annual and completely unnecessary new tradition, the NCAA Men's Basketball Tournament Selection Committee announced their current list of top four seeds for upcoming 2025 basketball tournament. On its face, this announcement was interesting, but otherwise forgettable.

But as a part of revealing of this potential tournament bracket, a side discussion of sorts was started. Members of the ESPN College Gameday crew provided an update on current discussions between the NCAA and conference leadership on the potential to expand the NCAA basketball tournament beyond its current scope of 68 teams to 72 or possibly 76 teams.

When the topic of tournament expansion was raised during the live broadcast from Alabama's Coleman Coliseum, the fans in attendance booed loudly and in unison. Had I been present in Tuscaloosa on that day, I would have booed right along with them.

[https://x.com/PeteThamel/status/1890803483688775838]

I have personally made my opinion on the matter very clear over the year: the current NCAA Tournament is virtually perfect in its current form. It is the single best annual sporting event in the history of time. Messing with the format is not just a bad idea. It's a terrible idea. 

While the discussion of tournament expansion has come up from time to time, this time feels different. The talk seem more serious, and this is a very, very bad thing. 

Most concerning is that prominent members of media are suddenly joining the call to expand the tournament. Specially, CBS Sports analysis Seth Davis wrote an article advocating for the expansion of the tournament. In the past week, Seth has been active on social media defending his position.

I have always respected Mr. Davis as a basketball analyst. I believe that he is one of the best in the business. But on this topic, his opinions are misguided and not substantiated by the facts. Today, I will explain, in detail and using data, exactly why.

College basketball and the NCAA Tournament is a passion of mine that I have had since I was a child. As a grew older, I began to apply some of the skills that I has acquired as a engineer and research scientist to the analysis of tournament results and trends. 

This has provided me with a deep understanding of the statistical nature of the Tournament, which I have written about extensively in the past. I will use some of that insight to respond specifically to Seth's main points.

Overview of Seth Davis' Four Main Points

In Seth's article, he specifically refers to three reasons why the tournament should expand. Those points are:

• There are a lot more good teams
• It would spice up Tuesday and Wednesday
• The power conferences want it

Based on the introduction to the article, I would add a fourth point that Seth is using to support his thesis:

• The Tournament has stayed at 64 or 68 teams for too long and "The time has come for the tournament to expand again"

Let's address all four of these points, but in a slightly different order.

"The power conferences want it"

Seth buried this reason at the bottom of his article, but for me, this is the key point of the entire discussion and likely the most serious issue. The logic behind this argument also reveals how any possible expansion is likely to proceed.

The key point is that power conferences are pushing for expansion, not for money (which Seth mentions several times is not the primary driver of expansion talk) but because "because they want more of their teams to play in the tournament."

Moreover, Seth emphasizes the seriousness of this point by explaining out that the most powerful high-major conferences could "blow up March Madness and create a tournament of their own." This comment is in reference to recent news that the power conference leadership is pushing to have more control over all Division I post season tournaments.

In other words, leaders of conferences such as the Big Ten, SEC, ACC, Big 12, and Big East are demanding more teams, and if they don't get what they want, they may take their ball and go home.

First, this is not a rational argument for why expansion is a good thing. It is simply an explanation of why the discussions are happening in the first place. While the true motivation may not be money, the underlying motivation is the same: greed.

This attitude from the power conference is reminiscent of two very different scenarios. First, it sounds like a threat from a blackmailer or other criminal. The appropriate response is simple: don't negotiate with terrorists.

But the second scenario is likely more accurate. This position reminds me of a toddler who is demanding a second piece of cake for dessert. I can easily the youngster screaming, "BUT I WANT IT NOW!" to his or her poor mother. 

"I want it," is also simply not a good reason. It is barely a reason at all. The resolution to this problem is simple. The adults in the room need to explain to toddler that that can't always get what they want, especially if the thing that they want is bad for them (or others) in the long run.

But the other key point from this part of the discussion is what is most likely to happen if the tournament does expand: more teams from the power conference are going to get in. We might see one or two additional mid major teams in a 72- or 76-team field, but the general result is that the rich are going to get richer.

That is not OK.

"The time has come for the tournament to expand again"

Seth began his article with a history lesson regarding the size of the NCAA Tournament, from the humble beginning including only conference champions, to the expansion from 40 teams in 1979, to 64 teams just six years later.

The tournament added one more team in 2001 and then four in 2011 to reach the current number of 68. "It has been 40 years the NCAA Tournament has undergone a significant expansion," Seth wrote. "The time has come for the tournament to expand again."

But this statement alone has no merit. Why is a larger number better? There is obviously both a lower and an upper bound on the optimum number of teams in the field. 

Too few teams does not create enough access to ensure a legitimate champion. Look no further than the BCS era of college football for an example of the problems when a "tournament" is too small.

But tournaments can be too large as well. From logistics point of view, there is only so much time in the schedule to accommodate so many games. Furthermore, a tournament including all 364 Division I basketball teams would be logistically ridiculous and make the regular season meaningless. 

Any post-season tournament should have a sense of exclusiveness. Only teams who are "worthy" should be invited to participate. That, in itself, is a philosophical question too deep to address here, but the implication is that there exists an optimum size for tournament. 

That size should be justified based on reason and not based on an idea that it has expanded in the past and therefore should expand in the future. That notion is nonsense.

Therefore, the key question to answer is "what is the optimal size of the NCAA Basketball tournament." This is the question that Seth started to explore is his third point.

"There are a lot more good teams"

This is the section of Seth's article that forms the logical crux of his argument. It is the section where we uses data and analogies to defend his position. It is also the section of his article where we went the most astray. Let's break down each point of his argument. Buckle up, folks. This is the part where I brake out my slide rule.

His first point in this section is that since 1985, the number of Division I teams in college basketball has expanded from 282 to 364. He argues that it is reasonable for the number of teams in the tournament to expand accordingly.

The problem with this argument is that the expansion of Division I has exclusively occurred at the low-major level with small schools making the jump up from lower divisions. 

But we have already established that the additional four to eight teams added to the field would almost exclusively come from the ranks of the high-majors or at least the mid-majors. Adding teams only increases access for the existing "haves" and not the newly added "have nots" from more recent Division I expansion. This argument is not at all compelling.

Seth goes on to cite the results from a report from the NCAA’s Division I Transformation Committee in 2023 that called for a participation rate of 25% in NCAA Championship of active Division I members for team sports sponsored by more than 200 schools such that the tournaments would "provide national-level competition among the best eligible student-athletes and teams."

All I will say here is why is a value of 25% reasonable? Why not 20% or 30%? Furthermore, why is this value specified only for sports with more than 200 schools? If it were applied to football, the college football playoff would have to expand to over 30 teams.

None if made sense until I read that the report in question was co-authored by none other than SEC commissioner Greg Sankey, one of the loudest voices behind Tournament expansion. It seems that Mr. Sankey has attempted to legislate his agenda into NCAA policy. The only reasonable response to such a foolish gambit is to simply ignore it. It has no logical foundation and therefore holds zero value.

But the heart of Seth's argument in this section revolves the title of this section. His hypothesis is that there are more worthy teams now than there were 20 years ago and that "parity that has overwhelmed the tournament the last 20 years."

Unfortunately, neither statement is true based on a through evaluation of the data. His assumption of increased parity appears to be based on a combination of recency bias, a misunderstanding of statistics, and some cherry picking of the data.

Direct evaluation whether teams, especially bubble teams, are better in 2024 than they were 20 years is challenging, but there are several ways to test this hypothesis.

In what will be a running theme in this section, I often utilization predictive metrics, such as the tempo adjusted efficiency margin data tabulated by Ken Pomeroy, to compare teams. My justification is that "Kenpom" efficiency data correlates strongly to Vegas spread data, and Vegas spread data has a very strong correlation to the probability that a team wins or loses. I have written of this topic extensively in the past.

If it were true that bubble teams have gotten stronger over the years, a trend should be visible in the Kenpom data for teams ranked roughly No. 40 to No. 60. One would expect those teams to be closer ranked to each other in 2024 than they were in 2004. Such as trend does not exist.

As a more relatable example, it is possible to use Kenpom data to estimate the point spread for an average No. 1 seed against an average No. 11 seed as a function of time from 2002 (when Kenpom data is readily available) to the present. If the hypothesis is true that the bubble teams (the No. 11 seeds) got stronger over time, this estimated point spread should show a downward trend over time.

A ran these calculations and the result is shown below in Figure 1.

Figure 1: Estimated point spread for an average No. 1 seed versus an average No. 11 seed from 2002 to 2024 using historical Kenpom efficiency margin data.

As we can see, there is no obvious trend in this data, but there is a lot of year-to-year variance. There is no evidence that a bubble team in 2002 is any better than one in 2024.

Seth goes on to provide several anecdotal pieces of evidence to support his hypothesis that parity has increased significantly over the last 20 years. There are three main examples that he cites for more parity

1. Mid major teams making the Final at a higher rate
2. Double-digit seeds making the Final Four as a higher rate
3. A higher rate of big upsets, such as No. 16 and No. 15 seeds winning first round games and beyond

As for the number of mid major teams making the Final Four, Seth states "before 2006, the 1996 UMass team was the only mid-major team to make the Final Four since Indiana State in 1979." He then lists eight mid major teams who made the Final Four between 2006 and the present.

If we ignore the fact that evoking the success of mid-major teams in the context of a proposal that is mostly aimed to help high-major teams, there is still a major issue with this statement: it is not very accurate.

As I look through the list of Final Four participants from 1979 on, I spotted several teams that I believe qualify as a mid-major, at least at the time. Pennsylvania (1979), UNLV (1987 and 1990), Cincinnati (1992), Utah (1998), and Marquette (2003) would all seem to quality. Mid major participation in the Final Four has not changing significantly since 1979.

As for the number of double-digit seeds making the Final Four, in order to understand if these value are usual or unusual, one must have a way to estimate the odds of these teams advancing that far in the Tournament. Fortunately, I also have developed the statistical tools needed to answer this question. 

Once again, I used historical Kenpom efficiency data to project potential point spreads for all possible tournament match-ups. If I assume that each seed is as strong as a historically average team of that seed, it is possible to calculate the odds that each seed will win a region and advance to the Final Four. 

In Figure 2 below, I compare the results of this calculation to the actual number of Final Four teams of each possible seed from two different time periods: 1985 to 2003 and 2004 to 2024. 

Figure 2: Odds for each NCAA Tournament Seeded team to make the Final in two time frames compared to simulated results.

Note that in general, the results of my calculation agree very well with actual results of the Tournament, with some notable exceptions. For examples, No. 3 seeds and No. 6 have both underperformed significantly in the past 20 years.

As Seth points out, No. 11 seeds have had more recent success reaching the Final Four than expected compared to the first 20 years of the 64-team era and the simulation results. In addition, No. 1 seeds had more success than expected in reaching the Final Four from 1986 to 2004. Both of these observations tend to support the idea of more parity in recent years.

That said, Figure 1 has already demonstrated that there is nothing special about the No. 11 seeds over the past 20 years. There is no evidence that teams that far down the seed list are more talented or capable of beating a No. 1 seed than they were in 2002. 

But what Figure 1 does highlight is the potential variance in the relative strength of different seeds over the years. This is not a trend, it is simply the result of the nature randomness of the Tournament and sports in general.

From this point of view, the overrepresentation of No. 11 seeds in recent Final Fours is simply a minor statistical anomaly. It is no more or less significant than the fact that a No. 6 seed hasn't made the Final Four since 1992.

But just making the Final Four in just one parameter. As an additional counterpoint to the success of No. 11 seeds in reaching the Final Four recently, I would be remiss if I did not point out that the lowest seed to ever win the National Title since the introduction of seeding in 1979 was No. 8 seed Villanova, who did so in 1985.

Furthermore, two of the next three lowest seeds to ever win the Tournament were a pair of No. 6 seeds in 1983 (N.C. State) and in 1988 (Kansas). No. 7 UConn in 2014 in the only team seeded lower than No. 4 to win the Tournament since the late 1980s.

Seth's final point about the apparent sudden appearance of shocking upsets such as a pair of No. 16 teams beating No. 1 seeds twice in the last seven years lead me to another important insight that I have gained in my extensive study of the NCAA Tournament. 

The frequency of upsets in the NCAA Tournament is completely predictable based on an analysis of typical point spreads in any arbitrary college basketball game, whether it occurs in March or not.

For example, the average point spread for a first round game between a No. 3 seed and a No. 14 seed is about 11 points. College basketball teams which are 11-point favorites, no matter when the game is played, win straight up about 86% of time. In NCAA Tournament play, No. 3 seeds have a record of 133-23 (85.3%) against No. 14 seeds.

The math checks out basically every time. This is why it is possible to simulate the possible results of the NCAA Tournament in way that provides statistics and trends that match reality, such as Figure 2.

So when it comes to No. 1 seeds versus a No. 16 seed, the average point spread is around 24.5 points. This corresponds to a winning percentage of almost exactly 99%. Thus, in NCAA Tournament play, we should expect to see a No. 16 seed beat a No. 1 seed roughly once in 100 tries, or once every 25 years. 

With No. 16 seeds first making an appearance in 1985, the first No. 16 seed upset was not to be expected until around 2010. It arrived a few years late in 2018. The second No. 16 upset was most likely should occur sometime between 2010 and 2035. It arrived a little ahead of schedule in 2023.

When it comes to low probability events, this type of occurrence is not a huge surprise. It is well within normal variance.

As for No. 2 seeds versus No. 15 seeds, the point spreads suggest that a No. 15 seed should prevail about 5% of the time or once every five years or so. The first win by a No. 15 seed happened in 1991,  seven years after the introduction of No. 15 seeds and therefore just two years later than expected. The second No. 15 seed upset happened two years later in 1993, and the third happened in 1997. Over the first 20 years of the expanded tournament, a No. 2 seed lost four times, exactly as expected.

Extending this analysis beyond a simple exploration of No. 15 and No. 16 seeds, it is possible to conduct a more comprehensive analysis on the frequency of common tournament upsets as a function of time and as a function of the expected frequency. Figure 3 below shows this comparison for the 20 most common seed pairing in Tournament history, divided in the first and second 20-years periods of the expanded NCAA Tournament.

Figure 3: Comparison of the actual upset rates for the 1985-2003 timeframe and the 2004-2023 timeframe to the expected upset rate extracted from the projected point spreads.

In this figure the blue circles represent that data from 1985 to 2003 while the green diamonds represent the data from the last 20 years. 

If the hypothesis were true that there has been more parity recently, we would expect to see the green diamonds mostly appearing in the bottom half of the graph and the blue circles mostly appearing towards that top. While that is true for some of the seed pairings, it does not appear to be a general trend.

For example, it is true that the frequency of No. 16 and No. 15 upsets has modestly increased in the last 20 years. No. 12 seeds and No. 11 seeds have also faired better recently in first-round action. In second-round action, No. 7 seeds have more frequently beaten No. 2 seeds and No. 11 seeds have racked up more wins over No. 3 seeds in the last 20 years than in the previous 20 years.

But for all of those examples of lower seeds performing better in the near term, there is an almost equal set of counter examples where underdogs performed better prior to 2003. For example, No. 14 seeds had more upset wins in the first 20 years of the expanded tournament, as did No. 9 seeds over No. 8 seeds.

This pattern is more prevalent in later tournament rounds. For example, No. 10 seeds used to fare much better in second round games against No. 10 seeds, as did No. 12 seeds against No. 4 seeds. The first 20 years of the Tournament also saw far more upsets of No. 6 seeds over No. 3 seeds and No. 4 seeds over No. 1 seeds in the Sweet 16.

In general, Figure 3 shows no clear signs of a "overwhelming amount of parity" over the past 20 years. There are signs of statistical variation, which is completely expected, but not a overall trend of more upsets or parity in recent year.

That said, there is a potential explanation for the modest increase in upset rates for the No. 15 and No. 16 seeds since 2011. When the NCAA expanded the tournament to 68 teams and instituted the First Four, it resulted in a slight increase in the difficultly of the first round games for roughly the top four or five seeds in each region.

Starting in 2011, the lowest four teams in the entire tournament played in the First Four. This means that the remaining two No. 16 seeds would have been placed on the No. 15 seed line in 2010. Similarly, half of the No. 15 seeds would have been No. 14 seeds in previous tournaments. Practically, this resulted in a subtle tightening of the point spreads (and therefore an increase in the odds of an upset) for all of the top seeds.

This effect certainly could give the appearance of more parity, but it is not because "there are a lot more good teams." It is because the structure of the tournament fundamentally changed in a way that impacted the probabilities.

In my final rebuttal of this third point, I believe that there is a way to mathematically calculate one parameter that captures the overall parity of a given tournament. This value is equal to the geometric mean of the winner's odds in each individual tournament game. 

I have also written about this extensively elsewhere. This "parity parameter" is lower if a tournament has a lot of low probability results (such as big upsets) and it is higher when the favored teams generally win. The parameter correlates to the total number of upsets in a given tournament, weighed by the "severity" of those upsets. (It is also mathematically related to the odds to correctly predict a prefect bracket, but that is a story for another day.)

If it is true that the parity of March Madness has increased over the past 20 years, my parity parameter should show a downward trend. Figure 4 shows the results of this calculation, going back to 1979 where point spreads are approximated in the time prior to prior to 2002.

Figure 4: The Parity Parameter tabulated for every NCAA Tournament from 1979 to 2024. Lower values correspond to more upsets and more parity

In general, Figure 4 does not support the the hypothesis of a clear trend of increasing parity over the past 20 years. The 40 trend is best explained as follows. 

Parity and chaos reigned in the first few years of the tournament, but the madness gradual subsided over the first 15 years to reach a relative low point in parity in the mid-1990s. 

Parity increased slightly and then was relatively stable until 2007, which was the first of three consecutive years of some of the lowest parity observed in Tournament history. This includes the only year when all four No. 1 seeds advanced to the Final Four (in 2008).

The Tournament got a bit more chaotic starting in 2010 and parity increased through the 2014 tournament, punctuated with No. 7 seed UConn claiming the National Title. The next five years alternated between high parity and low parity all the way up to the lost COVID year of 2020.

The three post-COVID tournaments of 2021, 2022, and 2023 all had relatively high levels big upsets, which likely in the cause of the some of the recency bias present in Seth's analysis. But in 2024, a relative low level of parity returned.

Based on all of the data presented above, I am confident that we can reject the hypothesis that the last 20 years has been filled with parity due to in influx of a lot more better teams. This idea is not a justification in any way for tournament expansion.

"It would spice up Tuesday and Wednesday"

The final point in Seth's article is that an increased number of games in a supersized First Four, mostly likely on the Tuesday and Wednesday after Selection Sunday, would be more fun. If some basketball is good, then more basketball has to be better.

I beg to differ.

To expand on Seth's food analogy, a delicious dish, more than anything, has balance. The perfect chili recipe has the optimal blend of meat, beans, tomatoes, garlic, and chili pepper. While there is no accounting for taste, adding a half dozen ghost peppers to the pot will likely make it inedible.

The NCAA Tournament, in its current form, also has perfect balance. The first two days of the Tournament is wall-to-wall action, often filled buzzer beaters and a few unexpected upsets as David takes down Goliath. As the Tournament rolls on, Cinderella is eventually shown the door, and the stronger teams that survive battle for the ultimate prize over the next five rounds and three weekends.

The NCAA Tournament, in its current form, provides access to all Division I teams (through conference tournaments) but also includes pretty much any teams with a reasonable statistical chance to win it all through the at-large big process.

From a simple scheduling point of view, there is no need to "spice up" Tuesday and Wednesday. Quite honestly, us fans could use the break.

As a passionate college basketball fan, I watch many of the conference tournament games in the week prior to Selection Sunday. The build up to the release of the bracket is perhaps the most exciting time of the entire year for me. 

As for the Tournament itself, I have taken paid leave every single year of my professional career, without exception, to watch as much of the first round action on Thursday and Friday as possible. Unless absolutely necessary, I don't leave the couch until late Sunday night.

It's glorious, and I am far from alone in this tradition.

But to be honest, I do not carve out time to consistently watch the First Four. If an expanded tournament calls for games to be played during the work day on Tuesday and Wednesday, that is even too much for me. It would especially be too much if those games mostly include a pair of low-major "Davids" fighting just to reach the "real" Tournament or mediocre high-major conference teams that don't deserve to be in the tournament anyway.

If this is my stance and feeling about an expanded First Four, the response from more casual fans is likely to be even worse. It will be apathy.

After the drama and excitement of the conference tournaments and Selection Sunday, I prefer to spend the next few days in peaceful reflection with the bracket. It would also be great it that bracket contained a minimal number of the "either/or" options that the play-ins games create. Those only cause confusion and frustration in the ubiquitous cultural phenomenon known as the "office pool." 

No, Seth, we don't need more basketball between Selection Sunday and Thursday. Those three days provide a much needed calm before the exhilarating storm that arrives on Thursday at noon. If anything, the Tournament needs fewer play-in games, not more.

What is the Optimal Size of the Tournament?

Based on the thorough analysis above, I have presented what I believe to be a convincing argument as to why there is no need to expand the NCAA Tournament. But this begs the question as to what is the best number of teams to include in the bracket?

As stated above, the ideal tournament should provide broad access to the full Division I field and it should also contain all the teams with a reasonable shot to win the tournament. The automatic bid process through conference tournament champions handles the first issue easily. This currently accounts for 31 automatic bids. But how many at-large bids are needed?

We can look at this from a couple points of view. As a general rule, I propose that if a team cannot finish over .500 in conference play, they likely are not worthy to play in the NCAA Tournament. This is the very definition of high-major conference mediocrity.

There are certainly exceptions to this rule. This year conferences such as the SEC are strong enough that perhaps 13 of the full 18-team membership are deserving of a bid. But conferences such as the  ACC may only deserve five teams. 

As a whole, a reasonable rule of thumb is that the Tournament should contain no more than half of the at-large candidates from the five major conferences (the Big Ten, Big 12, SEC, ACC, and Big East). The current combined membership of those five conferences is 79 teams. Five of those teams receive automatic bids. 

This leaves 74 at-large candidates. If a maximum of 50% of those team received at-large bid, that is 37 teams. If we add that the number automatic qualifiers, we arrive at the maximum reasonable size of the NCAA Tournament field:

68 teams. 

Bingo.

My proposal is that this number of at-large high-major teams (37) becomes a hard cap the for selection committee. The Committee can and certainly should select fewer than 37 high-major at large teams in order to make space for worthy mid-major teams that did not win their conference tournament.

Furthermore, any further expansion of the NCAA Tournament should keep this concept of a hard cap in place. If the Tournament were to expand in the future (which we have already established in an objectively terrible idea) that expansion should only include mid-major teams. 

That is the only even remotely acceptable option. It is in line with the idea that an expanded field of Division I teams has exclusively occurred outside of the high-majors conferences. It is also in alignment with the what should be the spirit  NCAA’s Division I Transformation Committee's 25% guideline, even if the true intensions of its co-author were far from pure.

That said, there are also logical reasons why even a 68 team field is too large. As stated above, any at-large team added to the field should have a reasonable probability to actually win the Tournament. Despite the apparent magic of the No. 11 First Four teams over the past few years, all data suggests that the farther one goes down the seed list, the lower the odds to win the National Title.

But what is a reasonable probability? I would proposed the cutoff should be the point on the seed line where there is a 99% chance that a team above the line will win the tournament. Using the same methodology outlined above, I have calculated the cutoff line in each year back to 2002.

That number fluctuates over the years, but the average brackets suggests that it includes only the top 36 teams in the field or in other words, the top nine seeds only. If we expand the criteria to include only teams with a combined 99.5% chance to win the Tournament, the cutoff line moves down four slots to include the No. 10 seeds as well. 

Despite the fact that five No. 11 seeds have made the Final Four four times since 2006, the odds suggest that around 200 Tournaments would have to be played before an No. 11 or lower would actually win the whole thing.

This analysis suggests that the hard cap of 37 at-large high major teams is already too high. A number in the low 30s would provide plenty of access to any high-major teams with a mathematically reasonable chance to win the National Title.

If I put all of this analysis together, the actual perfect size of the NCAA Tournament has been staring us in the face this whole time. The leadership of the NCAA figured it out back in 1985. 

The perfect size of the tournament is exactly 64 teams, with a potential hard cap of the number high major teams to encourage the Selection Committee to include a few strong mid-major teams that failed to win their conference tournament.

Some might argue that there is nothing special about the number 64, but that is simply not true. The number 64 is a multiple of two. A 64-team tournament has no play-in games such that the potential path of each team contains the same number of games. 

Eliminating the play-in games would once again allow all conference champions to play in the real Big Dance instead of sending two of them home on Tuesday and Wednesday night. Those kids deserve a chance to play in the real first round. Taking that chance away from them in 2011 was disrespectful at best and despicable at worst.

A 64-team bracket has the same symmetric beauty as the face of a super model. It's the perfect answer.

Some might argue that I simply fear change and want things to go back to the "good old days." But I just spent 6,000 words telling why I am right. Besides, the NCAA has attempted to make changes to the Tournament before that have not worked out. 

Does anyone remember the period of time when the NCAA named the four Regions after cities instead of the traditional cardinal directions? How about the era after the introduction of the First Four where the NCAA insisted on calling the games on Thursday and Friday "the second round?"

Those were both bad, unpopular ideas and NCAA eventually acquiesced to put things back the way that they were. Both were simply window dressing. Tournament expansion is far more serious, far more unpopular, and an objectively far worse idea.

The Single Biggest Reason Not to Expand

With all of those words spent on why the Tournament should not expand (and if anything should contract) I have only touched on the signal biggest reason not to expand the NCAA Tournament.

Fans absolutely hate the idea.

The ESPN GameDay Crew got a taste in Tuscaloosa of the vitriol that fans such as myself have for expansion. The response to the proposal through the week on social media has been overwhelming negative.

The argument has been made that fans will simply "get over it" and will still watch the games. Maybe. I have already explained how more games on Tuesday and Wednesday are unlikely to draw that much interest from either hard-core or casual fans. 

If an expanded tournament were to push more mid-major teams into the play-in games, that could reduce the magic in the first two rounds of the tournament as well. If we combine a more exhausting schedule, a less interesting opening weekend, and a more complicated office pool bracket with fans growing disenchantment with concepts such as NIL and the transfer portal, it could be a very, very dangerous combination. 

Like it or not, all of these billions of dollars that the NCAA Tournament generates come from advertising dollars. Advertising dollar are dependent on the number of eyeballs watching these games. 

College basketball fans are the customer in this particular endeavor. Fans hate the concept of expansion. The customer is always right.

The Adults in the Room

So what happens next? Over the next few months, my understanding is that members of NCAA leadership and the media broadcast partners (CBS/Paramount) will hash out a deal to either expand the tournament or not.

I have presented my case as to why no changes are needed to the Tournament. 

At the beginning of this piece, I mentioned how there needs to be an adult in the room to tell the Greg Sankey and his cronies that just because they want something does not make it a good idea. Those adults in the room need to be the executives of the media broadcast companies. To close, I would like to speak directly to them.

The NCAA Basketball Tournament in its current form is close to perfect. There is no logical reason to expand the tournament field other than the greed of the high-major power conferences. Expansion of the Tournament add little or no value but it does presents a real risk to the ongoing popularity of the event. 

When it comes to the NCAA Tournament it is the goose that lays the golden egg. It seems like some people are advocating that the goose should be fattened up if not force fed because duck meat and pate might be tasty. We all know how that story ends.

The correct choice could not be more clear. Just say no to NCAA Tournament expansion. America will thank you. 

Dear @SethDavisHoops. About a week ago, I promised to share my thoughts and data regarding your article advocating for the expansion of the NCAA Tournament. I have summarized those thoughts in the attached article and in this thread.

1/16 Your core hypothesis seems to be that the a larger tournament is warranted because there are both more Division I teams and more "good teams" then there were 20 years ago. I dug into this core question from multiple angles.

2/16 First, the fact that D-I has expanded over the past 20 years is irrelevant, as the main beneficiaries of tournament expansion are clearly mediocre high major teams and not the low-majors teams added through D-I expansion

3/16 Second, I used Kenpom data over the last 20 years as a proxy for projected point spreads. These data show no trend in the improvement of potential bubble teams since 2002. There is significant year-to-year variance, but no trend.

4/16 Third, while it is true that No. 11 seeds have made the Final Four at a higher frequency than expected, the data above shows that this is due to simple statistical variance, similar to the odd lack of No. 6 seeds in the Final Four since 1992.

5/16 Fourth, a comprehensive analysis of common pairings (such as No. 3/14 seed match-ups, 2/15, 1/16, etc.) shows no trend in the first 20 years vs. the 2nd 20 years of the 64+ team bracket. It only shows normal statistical variance, which is easy to cherry pick.

6/16 Fifth, the frequency of No. 16 seed upsets (and all other upsets) is completely predictable based on point spreads alone. The observed pattern is consistent with the expected rate of one occurrence per roughly every 25 tournaments (1% odds).

7/16 Sixth, I have developed a single metric (the "parity parameter") that measures the number and intensity/severity of upsets in a full tournament. This parameter shows no trend in the increase in parity over 20 years or even 40 years.

8/16 Seventh, several analogies in the piece were anecdotal and failed to mention key counter examples, such as other mid-major teams making the Final Four in the 80s and 90s, and the three of the four lowest seeded champions occurring in the 1980s.

9/16 Based on the seven points above, the hypothesis that "there are more good teams now" can be firmly rejected as false. Moving on to the other key points of the article which are all subjective

10/16 The Tournament does not need more spice. Serious fans need a break between conference tournaments, Selection Sunday, and the start of the Madness on Thursday. Casual fans don't care about the First Four.

11/16 The fact the conferences such as the SEC and Big Ten may want expansion is meaningless. Someone (namely CBS) needs to tell them "no" because "it's for your own good." Fans ABSOLULTELY HATE this idea, and fans are the customers.

12/16 The size of the "ideal" bracket can be estimated using a few reasonable assumptions. The Tournament should provide access to each conference (through its champion) and it should include enough "worthy" at-large teams.

13/16 One way to define "worthy" is to finish over .500 in high major conference play. A related metric is that no more than half of the power five teams should receive at-large bids. More than that dilutes the field. This should be instituted as a hard cap.

14/16 A second method is to defines a line on the seed line where there is a 99% or even a 99.5% chance that a team above the line will win the tournament. This falls around the No. 10 seed line, very close to the current "bubble."

15/16 This analysis suggests that the current Tournament size (68) is ideal, if not slightly too large. The folks in 1985 got it right. 64 is the "correct" size of the tournament, as every team starts on Thurs/Fri and every team has the same path length.

16/16 All this and a lot more is found in my full article. The conclusion is clear. There is absolutely no reason to expand the tournament and if they do there is risk of damaging the best sporting event even conceived. Just say no to expansion.

Rumors are circulating that the NCAA Basketball Tournament is on the brink of expansion and certain media members have joined the chorus. Dr. Green and White (@paulfanson) is here to explain, using the scientific method, why we should just say "no."