Picture a gym in Oxford, Ohio, sometime in late February. Miami of Ohio has just held its conference tournament opponents under 41% effective field goal percentage across every single game. The coach, Travis Steele, who a few years ago couldn't get a return call from power-conference programs that used to pick up on the first ring, is watching his team do something specific and repeatable: take air out of the ball, force hard 2s, and trust a defensive scheme built over years with the same group of guys. The bracketologists look at Miami's conference affiliation and put a 12-seed line next to the name. Their models say upset. Their hearts say chalk.
The numbers from 2025 are not subtle. Only 13 outright underdog wins, tied for the fewest since the field expanded to 64 teams back in 1985. The Elite Eight was all top-3 seeds, the chalkiest bracket since 2007. The average spread in tournament games hit 8.7 points, a 10-year high. Games got decided by 15.1 points on average, a 5-year high. Power programs played 50 fewer inter-conference games against mid-majors in 2025-26 compared to 2023-24, and the point spreads in the games they did play ballooned from 16.3 to 22.9 points. Small-conference teams beating power opponents dropped nearly 59% from 2021-22. The gap is real, it is documented, and the NIL economy built it intentionally.
What the Spreadsheet Cannot Hold
Here is where I'll give Jax Moreno his yard: the regular-season data does reflect something genuine about talent concentration. When a mid-major loses its 3 best players to the portal every spring because a power program can now write a check that looks like a car dealership's quarterly earnings, the talent pool visibly thins. That is a structural change, not a blip.
But bracketology lives in regular-season data, and the tournament is not the regular season. The tournament is 40 minutes in a neutral gym where the 1-seed's freshmen play their 4th high-stakes game in 6 days while Miami of Ohio's 5th-year point guard has run this exact defense 31 times. Fatigue compounds differently. A team that has played together for 3 years has communication habits that no portal roster assembled in October can replicate on command. Familiarity is a weapon, and it shows up late in games when legs are gone and the margin for error collapses to nothing.
The reason bracketologists consistently underweight upsets is not that they are bad at math. It is that their models measure what happened between October and March, then assume the same forces apply when you strip away home courts, strength of schedule padding, and the emotional insulation that comes from having 30 games left to correct a loss. Remove all of that, and you have 1 game, 40 minutes, and a team from the MAC that has held opponents under 41% effective field goal percentage for the last 2 weeks. That is a different calculation entirely.
The Game Dean Oliver Called Amazing Theater
Analyst Dean Oliver put it plainly: the basketball at the top is better than it has ever been, and it is single-elimination basketball. He meant it as a compliment to the spectacle. I hear it as the whole argument. Better top teams still lose in single-elimination formats because single-elimination is a pressure test, not a quality assessment. George Mason went to the Final Four in 2006. Hampton knocked Iowa State out in 2001. These were not accidents; they were what happens when preparation meets a specific opponent on a specific night with no margin for correction.
The money widened the gap. The format kept the door open. Bracketologists trust the gap and forget the door. Every March, that costs them.
