Three-Man-Weave

View Original

Forecasting The Leap: How Good Are We At Predicting Team Jumps Into The Top 10?

-Jim Root

Preseason rankings are about as subjective a measuring stick as possible. At this point, we have no legitimate evidence of how teams will perform in the coming season. Prior year statistics, recruiting rankings, and copious film study are valuable tools, but even then, no one knows the whole picture. As such, groupthink often lurks in the ranking landscape.

Falling prey to groupthink is often simply a defense mechanism. If almost everyone agrees a team should be ranked at/around a certain spot, being the metaphorical fly in the ointment will inevitably draw plenty of flak. And considering we’re still five months from games being played, the upside to zagging is limited – even if you’re right, will people remember (or care)?

In my view, at this point in the 2022 offseason, the biggest groupthink squad is Creighton. The Bluejays were an extremely young team last year who left a positive final impression in everyone’s minds by competing with eventual national champion Kansas for 40 minutes while exceedingly short-handed. As a result, suggesting that Creighton is not an ironclad lock to be a top five squad in 2022-23 has become blasphemous.

I find myself skeptical of that notion. Two huge reasons for that Kansas effort – Ryan Hawkins, arguably the Bluejays’ best and most consistent player, and Alex O’Connell – are gone. Baylor Scheierman is an outstanding replacement, but even with him joining the flock, I am not fully convinced the baby Bluejays immediately mature into a nightly juggernaut. Most of my reservations center around the fact that, by objective measurements, Creighton simply was not that good last season.

Rather than argue opinions on this, though, I wanted to arm myself with some data. For teams in the past like Creighton where the accepted notion is that they will take a significant leap, how often are we correct?

More simply: how good is “the crowd” at predicting major jumps?

To investigate this, I went through the last 15 seasons of AP preseason polls and pulled out all of the preseason top 10 teams. Note: I don’t love the AP poll, but it’s a solid proxy for “consensus opinion.”

For each preseason top 10 squad, I looked at the team’s ending KenPom rank from the previous season to narrow down the “leapers” – the teams everyone seems to think will climb from a middling finish the prior year into the nation’s elite (my cutoff: outside the KenPom top 25). That resulted in a sample of 28 teams over 15 seasons.

I then checked those teams’ final KenPom rank from that season. Essentially, did they become elite like everyone expected?

Note 2: All end of season Pom rankings are pre-tournament.

Don’t shoot the messenger, but those teams rarely became the steamrollers most envisioned. Of the 28 “Leapers” who fit the sample, only eight – just 28.6% - finished in the KenPom top 10. That’s pretty bad!

It’s especially poor when compared to teams with proven track records of success. I broke out the other 122 top 10 teams into two categories: finished previous season in the KenPom top 10, finished previous season from 11-25. Both of those categories proved far more stable in projecting future powerhouses:

It gets better if we expand the results into who at least managed to finish in the top 20, but again, the “off the radar” teams have the worst combined success rate:

Put another way: of the 28 teams expected to take “The Leap”, more of them (9) finished outside of the KenPom top 20 than those that finished inside the top 10 (8).

Maybe these conclusions hit you with the dull thud of an obvious “duh!” And to be VERY clear: I am not disproving that Creighton is a top 10 team — they might be!

Instead, the takeaway here: college basketball prognosticators are pretty bad at identifying which teams will go from “decent” to “elite.” Thus, we should not demand that groupthink dictate how we rank in the preseason — especially for someone like Creighton.   


Before you think I’m just bullying Creighton (I’m not! This is an objective data study!!!), remember that the 2022-23 Bluejays are not the only squad that qualifies. Last year’s national runner-up, North Carolina, was also outside the KenPom top 25 before the NCAA Tournament, meaning the Heels fit the bill.

2023 UNC has been compared ad nauseum to 2022 UCLA: coming off a torrid March run, placed firmly in the top 2-3 of all preseason lists. Those Bruins landed 8th in KenPom before last year’s tournament, so in this context, they were a success story. But it’s also fair to argue they were not the towering terror many assumed they’d be after pushing previously undefeated Gonzaga to the brink at the Final Four.

If only 30% of all teams expected to make this leap succeed, it would be fairly surprising if both UNC and Creighton managed it in the same year. Which one do you trust more? For me, it’s the Tar Heels.


Ok for fun, here’s some other quick hitter nuggets from the sample:

-The most common teams in this archetype are traditional powerhouses that underachieved. Of the 28 teams in the sample, nearly half (13) of them are UConn, Duke, Michigan St., Kentucky, UNC, Texas, or Villanova. It’s far less common for it to be a team like Creighton – traditionally on the outside looking in, but now firmly expected to crash the party.

-Of the eight teams that successfully vaulted up into the top 10, 2022 Kentucky is the only one to go from outside of the top 40 to inside the top five – aka what is universally expected of Creighton. And the ’22 Wildcats added the consensus National Player of the Year (among many other key newcomers).

-The best blueprint for Creighton? That would be 2018 Michigan State. The Spartans drew a ton of hype after freshman-laden group (Cassius Winston, Miles Bridges, Nick Ward, Josh Langford) earned a 9-seed in the NCAA Tournament and lost in the second round to 1-seed Kansas. Sound familiar? That ’18 edition brought back a potential lottery pick in Bridges, added star recruit Jaren Jackson Jr., and ended up 6th in KenPom and a 3-seed in the Big Dance (I guess I’ll mention they got upset by 11-seed Syracuse).

-Some of the recent failures who did not deliver on their hype and should serve as cautionary tales:

  • 2016 Maryland (Added Diamond Stone, Rasheed Sulaimon, and Robert Carter to a ballyhooed core of Melo Trimble, Jake Layman, and Jared Nickens)

  • 2018 USC (brought back four starters from an 11-seed that made the Sweet 16, soph leaps from Nick Rakocevic and Jonah Matthews)

  • 2020 Florida (soph core of Andrew Nembhard, Noah Locke, Keyontae Johnson while adding Kerry Blackshear, Scottie Lewis, and Tre Mann)


To tie a bow on this, I want to reiterate: I am not concluding that Creighton cannot be a top 10 team. I personally will rank them around 12th or so, but it is possible that “the crowd” gets this one right. To me, though, placing the Bluejays in the the top 5 assumes an extremely high-end outcome for a team that will still be fairly young and unproven as a unit. As indicated by the data above, the median outcomes for the Bluejays are probably in the 11-20 range.

UNC has some similar concerns. Brady Manek was vital to that team’s success (the on-off numbers are beyond jarring), so maybe we’re all a little too presumptive in the Tar Heels’ own jump.

Remember: preseason rankings are often wrong. Mine, KenPom, the AP Poll, and especially Jeff Goodman will all surely have some laughable placements when we look back. But for such an inexact science, do not fall prey to groupthink — particularly for teams expected to make a significant leap.