College Confidential

The mysterious power of college basketball's most clandestine clashes.

As the dog days of the college basketball offseason give way to the first glimmers of approaching conflicts, murmurs begin circulating of covert gatherings between the brightest minds and largest leapers that university basketball teams have to offer, testing and tinkering far away from the prying eyes of the media and the shrewd scouts of their enemies. What mysteries lurk beyond the ironclad gates of the secret scrimmages?

It's not quite so cloak-and-dagger. Before each season officially begins, each team is permitted by the NCAA to play one closed-doors scrimmage at any location agreed upon by the two teams. Sometimes this is a home arena, sometimes a practice facility, sometimes a high school gym halfway between the two colleges. “Away” scrimmages allow staffs a mock run through their road game routines and logistics, while home arena scrimmages allow players to familiarize themselves with what will become their shooting backdrop for the upcoming season. As promised, some level of secrecy is maintained, but scores, stats, and even Zapruder footage has been known to exist, usually buried in the bellows of message boards that retain a 1990s charm and populace.

I’ve cropped out the ensuing and entirely unrelated schizophrenic argument as to whether this was, or was not, in fact, the real Seton Hall fan forum.

Of course, the cooler heads among us scoff at such mindless conjectures based on unrealistic and informal events. True analysts know that conference games in late January will be far more reflective of a team’s true strength than these mere workshops in October.

But are they?

Rare photograph of Dr. J. Robert Oppenheimer explaining that teams who score more points aren’t necessarily better at offense, as their games may have a higher number of total possessions than a more efficient, yet slower team.

In order to test the viability of secret scrimmage outcomes as a predictor of future success, I set up two simple questions to answer for each of last year’s bouts. Who won? Who finished higher in the KenPom rankings at the end of the season? Shockingly, or perhaps not, the first question was far more difficult to answer in many cases than the latter. Of the 29 “official” secret scrimmages cataloged by Stadium, at least five managed to keep such a seal of confidence that even a score could not leak into the musty slats of message board rumor mills. I was able to source a score for a few more of these through some rather dubious channels, but even so, a select few remain mysteries to the outside world.

Despite transferring to North Carolina this offseason, Cormac Ryan’s silence has made clear that secret scrimmage results remain classified even after one leaves a program. (Or perhaps he just doesn’t respond to random writers badgering him through Instagram DMs)

Of course, the rate at which secret scrimmage outcomes match final KenPom order is useless without some sort of context, so I also gathered a random assortment of conference games to serve as the baseline against which the success rate can be measured. One can expect that the serious and competitive environment of conference play would generate more predictive outcomes than the inchoate skirmishes of tinkering teams. The chances of secret scrimmages matching final KenPom rate closer than real basketball games are near zero.

What else can you expect from theory alone?

In 2022, power conference secret scrimmage outcomes matched KenPom ranking order at an equivalent, if not higher rate than a random selection of conference games over an equal sample size.

Does this mean that the message board overreactors are correct? Should we heed the whispers of outcomes that leak out of Los Alamoses with the same bated breaths we hold during mid-February clashes?

Yes, but the problem lies with the analysis of an average conference game, not with these Trinity tests. Secret scrimmages are not reliable indicators of which team will be better at the end of the year, not because of any intrinsic unreliability due to the unique conditions of the matchup, but simply because a secret scrimmage suffers from the same low sample size that any other singular data point presents. A single conference game is also unlikely to be representative of the entire season, as trends take many games to bear out. When paired with the substantial effects of home court advantage, travel times, and injuries, any one matchup in the Big Ten may be steeped in just as many zany variables as a laboratory scrimmage played on a high school floor in Davenport.

So go ahead, tigerbear7878. Allow Clemson’s secret scrimmage victory (when you add up the two independent segments, of course) to fuel your wildest speculations of the Tigers’ upcoming run to NCAA Tournament revelry. You’re no worse off than any analyst attempting to forecast the future from Clemson’s March 3^rd win over Notre Dame.

Just don’t leak the results to the Soviets, you pinko.

Did you, or did you not, start an “overrated” chant at 18-year-old G.G. Jackson when South Carolina played George Washington in 2022?