The Atlantic 10’s descent from routinely a three-bid league to often a one-bid league receives an appropriate amount of attention. It’s a shift which changes the fabric of college basketball. It’s not a shift that tears college basketball apart. Coverage reflects this.
This season, seven A-10 teams made the NCAA Tournament and NIT. After five days and nine games, all but one of those teams have been eliminated. That team, Loyola, is an underdog tonight at San Francisco. In about twelve hours, it’ll more likely than not be over for the A-10. Seven NCAA Tournament and NIT teams is a lot. The A-10 is one of the nine best conferences in the sport. But the early exits are disappointing from those hoping for a resurrection from the league. Making matters worse, they’ve mostly been surprising. A-10 teams have underperformed mathematical expectations this March.
We’re tracking two things about final scores this month, and we’re using kenpom to track them. The first, provoked by Pomeroy’s own column explaining why his algorithm probably overrates the SEC, is how much each conference is over or underperforming its kenpom spreads in the NCAA Tournament and the NIT. The second, provoked by ongoing discussions over the basketball used in these tournaments, is how totals are performing. Is the ball too inflated? When? Where? Why?
On that first front, the spreads: Here’s what we’ve seen from leagues whose teams have played four games or more:
| Conference | KP Spread W’s | KP Spread L’s | KP pt. diff. |
| Big Ten | 9.5 | 2.5 | 8.67 |
| WCC | 3.5 | 1.5 | 8.60 |
| MVC | 3.5 | 1.5 | 5.60 |
| Big 12 | 8.5 | 2.5 | 4.73 |
| SoCon | 3 | 2 | 3.00 |
| Big East | 4 | 4 | 1.38 |
| ACC | 3.5 | 4.5 | -0.38 |
| CUSA | 2 | 2 | -0.50 |
| SEC | 9 | 9 | -1.94 |
| AAC | 1 | 3 | -3.50 |
| MWC | 2 | 3 | -7.00 |
| A-10 | 3 | 6 | -8.56 |
| Big West | 1.5 | 2.5 | -9.00 |
The Big Ten, WCC, MVC, and Big 12 all came down to earth a little from where they stood yesterday morning, as did the Big East. This regression should be expected. Kenpom is not underrating teams from a given conference by 11.5 or even 8.6 points. If there is any meaning to this data, it’s subtle. Similarly, the SEC regressed in a positive direction, closing its gap with zero.
One league did not trend towards the mean, though, and it’s the one who saw two of its three best teams lose as favorites. In Chattanooga, Dayton lost a road game as the higher seed. In Fairfax, George Mason was beaten by visiting Bradley. These were the third and fourth A-10 teams to lose as favorites this postseason, making the league’s teams 2–4 when expected to win (and 0–2 when expected to lose).
In A-10 circles, a lot’s been made of what the conference can do to help its teams’ NCAA Tournament at-large chances. There’s talk of dynamic scheduling—striving to get better teams more Q1 and Q2 games. There’s talk of strategic nonconference scheduling—trying to game the system with NET, KPI, and SOR, a harder feat than hoodwinking RPI but potentially not impossible. There’s talk about better TV strategies and better branding leading to better subjective impressions while also driving revenue and notoriety, in turn boosting NIL spending. This last piece is the only one which gets at the real problem: These teams need to get better.
If Atlantic 10 teams want to inhabit a three-bid league again, they need to go get the bids in question.
If Atlantic 10 teams want to inhabit a three-bid league again, they can’t be worse than Conference USA.
As for the totals:
It was only two games, but the NIT had a higher-scoring day than expected. The tournament’s averaging final scores 6.3 points higher than their kenpom projections, not including points scored in overtimes. From Day 2 onwards—excluding Day 1, on which teams had fewer than 48 hours to prepare for one another—games have averaged finals 1.9 points beyond where kenpom had them. A smaller number, and not necessarily significant, but worth tracking given the curiosity around the balls used in NCAA postseason competition. Overall, the median game has exceeded scoring expectations by 8.5 points. That number stays up at 8.0 points if you remove Day 1 from the sample.
In the NCAA Tournament, scoring was up. Games overperformed their kenpom-projected totals by 9.1 points on average, the highest single-day mark in that category since the First Four. What do the First Four and yesterday and NIT Day 1 have in common? That lack of time with which to prepare. There were other factors at play, I’m sure, but that’s one that jumps out.
Notably, today’s NCAA Tournament sites saw poorer scoring compared to first round expectations than yesterday’s did. The bump may be short-lived. But for now, the average NCAA Tournament game this year has underperformed its projected total by 0.5 points, and the median game has met its projected total on the nose. I don’t know whether it’s better to go off of mean or median when trying to predict today’s overs and unders. Personally, I’m hoping the spreads all look appealing instead.
**
