If you feel like it's getting harder and harder to pick the Final Four in your bracket, you're not wrong. Since the tournament field expanded to the "near-modern" 52 teams (now 68), fewer and fewer of the high-seeded teams are reaching the Final Four. This year, if you add up the seeds of each of the Final Four teams, which I will refer to going forward as the "seed aggregate", you will get 17. This comes from a 1 for UConn, 1 for Purdue, 4 for Alabama, and 11 for N.C. State. This is the second-highest seed aggregate since 2014, with last year being the highest at 23. Only 6 years have had higher seed aggregates in tournament history, with 5 of them coming since 2001. The least "chalky" Final Four was in 2011, when 8-seed Butler and 11-seed VCU shocked the college basketball world on incredible Cinderella runs.
The seed aggregate and the average seed of a team in the Final Four has, on average, steadily risen over time. This is especially true since the early 2000's, with the only outlier being in 2008 when four 1-seeds made it for the first time in the modern era.
Many will cite this as proof of an increase in parity across college basketball, with the quality gap between mid-major and high major programs narrowing. I wouldn't argue that the talent-level of mid-tier programs has improved through the years, and that NIL and the transfer portal have given mid-major schools more opportunities. But trends and numbers show the selection committee may play a significant role in these increasingly unpredictable Final Fours.
The question lies in whether or not the selection committee has been accurately evaluating teams and seeding them appropriately. To judge a team's quality, I referred to College Basketball Reference's Simple Rating System (SRS). This is essentially their holistic rating of a team in its entirety, you can read a more in-depth description of it here. The graph below shows the average and median rating (based on SRS) of teams in the Final Four in that same time span:
This is almost perfectly flat, leading us to believe that the quality level of teams in the Final Four have stayed pretty consistent over time. So, while teams seeded worse have been making the Final Four more frequently, the quality of the team has remained relatively constant.
Is this just a function of the talent gap shrinking in college basketball? Is the quality of the teams just more similar, making the ratings of teams closer together and less variable? The graph below shows the standard deviation in SRS, across all teams in college basketball, from 1983 to 2024. This serves as an indication of the variability in ratings within a season. A lower standard deviation would represent less differentiation in team ratings (and therefore less differentiation in quality) in a given year. So, more parity should be reflected by lower standard deviations.
We see a slight gradual decline, represented by the linear trendline, but to nowhere near the extent that Final Four seed aggregate has risen. In fact, the correlation coefficient (which measures the strength and direction of relationships) between Final Four aggregate seed and yearly standard deviation in SRS is just -0.20. A correlation coefficient (r) of 0 represents no relationship at all, while a value of 1 represent a perfectly positive relationship (as one variable increases the other also increases) and -1 a perfectly negative relationship (as one variable increases the other decreases). The general rule of thumb is a coefficient between +/-0.8 and +/-1 is considered very strong, between +/-0.6 and +/-0.8 is moderately strong, and between 0 and +/-0.6 is fairly weak. This coefficient is negative, indicating that an increase in seed aggregate corresponds to a decrease in SRS standard deviation, which makes sense as we would expect more parity in a season could increase the likelihood that lower seeded teams make the Final Four. However, the very low magnitude of the coefficient gives us little confidence there is ay meaningful strength to this relationship. In fact, when a t-test of statistical significance is performed on this data, we get a p-value of 0.0527, which exceeds the statistical significance threshold of 0.05. Therefore, this relationship between aggregate seed and the standard deviation in SRS is not statistically significant.
What can we take away from this? Well, there is some indication of increased parity in college basketball, but simply not to the extreme that we are seeing lower seeds unexpectedly make the Final Four. So, while we may reasonably expect some lower seeds to make a deep run once in a while, it should not be happening as frequently as it has been. The committee needs to do better in selecting and seeding teams. Maybe not have a small group of school athletic directors and conference commissioners making the decisions? Let's open it up to more sportswriters, analysts, broadcasters, and the like; people who actually study the games and have far better knowledge of the sport. Regardless, NCAA you need to do better.