Let's talk about the 20/20 Club and the power of a good narrative

Choose your fighter: accuracy or convenience

A man in business casual frowning and looking frumpy at a spreadsheet, impressionism

Four years ago, prior to an NCAA Tournament that didn’t happen, I wrote about my disdain for a narrative developing on national broadcasts. Late in that season, I heard the same refrain over and over: “the national champion is almost always ranked in the top 20 of offensive and defensive efficiency at KenPom.” From the original article, here’s why this annoyed me so much:

On its face, this is a correct stat. If you were to click on kenpom.com right now, you would see that 17 of the last 18 champions, minus 2014 UConn, did indeed rank in the top 20 of both categories.

However: we have a clear issue that seemingly no one at ESPN, CBS, or the variety of networks that broadcast college basketball seem to be discussing. The KenPom rankings referenced are end-of-season rankings, not pre-tournament rankings. So, yeah, no wonder every champion ended up in the top 20!

The best part of this stat is that even if you did still believe in it for whatever reason and incorrectly read the stats, 2020-21 Baylor broke it in the very first NCAAT after I wrote this. They finished #22 defensively, which is not top 20. Still, I see this narrative everywhere, notably here or here or here or maybe here or here or here or here or here or here or here or definitely here, from the NCAA’s own dadgummed site.

The problem is that the disease has spread from the 20/20 Club to something even more confounding: the 40/22 Club. This refers to the other end-of-season rankings fun: that every national champion in KenPom history has finished inside the top 40 in offensive efficiency and inside the top 22 in defensive efficiency. (Guess which recent champion is the reason it’s top 22 and not top 20!)

Please note that this is not my screenshot, because I am not an 11-year-old child and do not have TikTok. NOT MY SCREENSHOT. Stop saying it’s my screenshot!

The reason people have pivoted to the 40/22 Club is, presumably, convenience. Go to KenPom dot com, click through every year available (now back to 1999), and you’ll see that every team finished with a top-40 offense (take a wild guess as to which 7-seed champion broke this) and a top-22 defense. After the 20/20 Club failed them three straight years, many have begun to pivot to this.

The only list with Houston, Purdue, Arizona, and Michigan State on it but NOT UConn? Plausibly.

I don’t feel the need to link to everyone on the 40/22 Club like I did the 20/20, because 40/22 is such a silly split that it only seems used with any frequency by the truest of engagement-loving accounts. Every time I see it, though, I wonder: do the people who shove these stats in your face disqualify teams who have a top-10 defense, but the #41 offense? That team could never win a title. Or maybe a top-5 offense with a 30th-place defense. Never seen that in my life!

The problem I have is less with a specific Club and more with how this has splintered into its own thing. You have the 20/20 Club and the 40/22 Club, but I’ve started to see the 25/25 Club pop up. (At least that one makes reasonable sense because of the existence of all the 25-team polls.) Here’s the 40/25 Club, which is like the 40/22 Club but with 0s and 5s. The 40/20 Club, the 12/12 Club (!), the 20/40 Club, the 21/37 Club, the 22/22 Club, the 10/15 Club, the 25/45 Club, and the 45/20 Club all have founding members. I even saw the 114/94 Club for the first time the other day, so we may be expanding into new states of mind that ancient philosophers never deemed possible.

I think it’s worth debunking this two-fold. First, the below are the rankings - overall, offense, and defense - of every champion on KenPom from 1997 (as far back as the database goes) to 2023. That’s 26 years of data. Most importantly: everything you see below is data BEFORE the NCAA Tournament began. So, those rankings. Every offensive/defensive outlier, AKA those outside of the top 20, is in bold. Teams outside the top 10 are in bold too.

• 1997: Arizona (#11 overall, #16 offense, #18 defense)

• 1998: Kentucky (#4 overall, #6 offense, #8 defense)

• 1999: UConn (#4 overall, #7 offense, #1 defense)

• 2000: Michigan State (#3 overall, #3 offense, #7 defense)

• 2001: Duke (#1 overall, #2 offense, #3 defense)

• 2002: Maryland (#4 overall, #5 offense, #12 defense)

• 2003: Syracuse (#20 overall, #16 offense, #31 defense)

• 2004: UConn (#5 overall, #11 offense, #8 defense)

• 2005: North Carolina (#2 overall, #4 offense, #3 defense)

• 2006: Florida (#5 overall, #7 offense, #23 defense)

• 2007: Florida (#4 overall, #2 offense, #15 defense)

• 2008: Kansas (#1 overall, #2 offense, #3 defense)

• 2009: North Carolina (#3 overall, #1 offense, #37 defense)

• 2010: Duke (#2 overall, #4 offense, #5 defense)

• 2011: UConn (#16 overall, #22 offense, #27 defense)

• 2012: Kentucky (#1 overall, #2 offense, #4 defense)

• 2013: Louisville (#2 overall, #18 offense, #1 defense)

• 2014: UConn (#25 overall, #57 offense, #12 defense)

• 2015: Duke (#6 overall, #3 offense, #37 defense)

• 2016: Villanova (#5 overall, #15 offense, #7 defense)

• 2017: North Carolina (#3 overall, #4 offense, #25 defense)

• 2018: Villanova (#2 overall, #1 offense, #23 defense)

• 2019: Virginia (#1 overall, #2 offense, #5 defense)

• 2021: Baylor (#4 overall, #3 offense, #44 defense)

• 2022: Kansas (#6 overall, #6 offense, #29 defense)

• 2023: UConn (#4 overall, #6 offense, #19 defense)

Look above and you’ll notice a few things. For one, no, not every champion is in the top 20 of offensive and defensive efficiency. Six of the last nine didn’t meet both requirements. Not every champion is top 25 in both, either. Four of the last nine and six of the last 15 didn’t get there. The 40/22 Club fails terribly here, too, because of the defense function. Only 10 of the last 20 champions actually qualify for the arbitrary 40/22 barriers. Even that goofy 114/94 barrier, which I really did have hopes for because it’s just a delight to read out loud, only worked for half of the last 12 champions using pre-tournament data.

A TI-83 calculator with a tornado spinning out of it, watercolor

The other thing you’ll notice is precisely what John Gasaway said two years back. You have a far better shot at finding your future national champion by narrowing down your search to six teams: the top six in KenPom entering the NCAA Tournament? Why not seven? Well, there hasn’t been a #7 to win it yet. Why not five? Then you’d only be 20 for 26. You could also say top 10 in KenPom and be just as right, but top six has the exact same hit rate. That has a better hit rate (22 for 26) than the 20/20 Club (16 for 26), the 40/22 Club (also 16 for 26), or picking a top-4 team (17 for 26) does. (Though the latter does a good job of narrowing things down!)

So, on one hand, you could read this data and join a club that already exists: the 20/40 Club. 24 of the last 26 national champions have had a top-20 offense and a top-40 defense on Selection Sunday at KenPom. If you want to fudge it and say top-45 to include Baylor, then you’re at 25 for 26. Then, I can go on X dot com and say that 25 of the last 26 national champions, excluding the ultimate outlier - 2014 UConn - were top 20 in offense and top 45 in defense. That list this year, for the engagement-inclined, would go:

• Purdue (2, 21)

• UConn (3, 11)

• Arizona (6, 12)

• Duke (10, 25)

• Houston (15, 1)

• Tennessee (16, 5)

• North Carolina (17, 14)

• Creighton (18, 24)

• Wisconsin (19, 39)

It even generates more engagement when people don’t see Auburn (22, 4), Marquette (23, 18), Alabama (1, 74), or even Illinois (5, 48) on your list. Now, you’ve given people nine teams to watch for. It would have given you nine teams last year, too. You would’ve had eight in 2021, 11 in 2019, and 9 in 2018. That narrows down the field, and you can let everyone know that this is the solution to your bracket woes.

Or you could just pick your favorite of the six best teams in KenPom, a system that’s worked eight Tournaments in a row and 22 of 26. Come to think of it, that sounds a little like telling you to pick the team you like the most out of the best group of teams in a given season. Maybe that would be easier than attempting to adhere to very vague guidelines that require arbitrary cutoffs, no?