College Basketball Watchability Above Replacement (CBBWAR): A new-old way to find the NCAA’s most entertaining teams

I think anyone who’s read my writing on Tennessee basketball can tell that over the course of the last month, the writing itself has grown more…negative? Cynical? Downward-looking? Over it? I’m not sure what the phrase is, but I’ve had a couple of people reach out with suggestions like “watch other teams.” I appreciate said suggestion, as I already do watch other teams, but that idea became rooted in my head as something I’d like to explore.

Long ago on Fangraphs, a baseball statistics site, writer Carson Cistulli created NERD, “an attempt to summarize in one number (on a scale of 0-10) the likely aesthetic appeal or watchability, for the learned fan, of a player or team or game.” I like baseball, but I realized about ten years ago that what I really love is postseason baseball. Regular season baseball…it’s 162 games, man. But instead of giving up on 162 games entirely, I loved reading Cistulli’s NERD reports every day, pinpointing the most interesting games of the week.

Cistulli left NERD (and Fangraphs) behind a few years ago, but the formula remains out there. I copied it to run my own sort of NERD for MLB this season, with a few tweaks (higher emphasis on homers, because homers) and new calculations. It selected the Los Angeles Dodgers, Toronto Blue Jays, and Tampa Bay Rays as the three most watchable baseball teams of 2021, which, yeah. (Braves fans: you were fifth. I imagine that if Weighted NERD existed, you would probably be top four or top three.)

Anyway, this is a long way of getting to the point: I’ve workshopped a similar idea for college basketball. Ensuring that all 359 team ratings are 100% accurate is borderline impossible, because 1) I have a day job and 2) As such, I’m unable to watch a lot of teams until they either play the team I cover or they’re on a network that everyone can agree on. I watch a lot of college hoops, but Sean Paul (not the singer) and others are whooping me in this regard.

To make up for this, I’ve devised a metric I’m loosely calling College BasketBall Watchability Above Replacement (CBBWAR). The name could be changed, but it’s a name that describes the point of the project and the acronym makes it sound like something that will get $41 billion dollars in military funding somehow. Here are the components involved, all sourced from either KenPom or Hoop-Math:

  • Tempo (alternately possessions per game)
  • KenPom Adjusted Offensive Efficiency (Adj. OE)
  • Effective Field Goal Percentage (eFG%)
  • Percentage of Shots That Are Long Twos (Mid%)
  • FG% at the Rim (Rim%)
  • Three-Point Efficiency (3PT%)
  • Defensive Block Rate (Block%)
  • Defensive Steal Rate (Steal%)
  • Three-Point Attempt Rate (3PA%)

Here’s how these components are currently weighted in my Excel sheet:

Tempo*0.5 + OE + eFG*1.5 – Mid*0.5 + Rim*0.5 + 3PT%*.75 + Block*0.5 + Steal*0.5 + 3PA*0.25 + Constant (currently 2.2, a completely arbitrary number to provide us with higher-rated teams)

What this roughly comes out to is an equation that values offense at 75%, defense about 18%, and tempo 7%. I like fast games, but I like seeing shots go in the net more. I do not like watching a bunch of mid-range twos by teams that generally cannot hit them. (Consider that the NBA average on a mid-range shot this season is 40.3%; in college basketball, it’s 36.9%, and only 21% of teams in America shoot at 40.3% or better. Thanks, but no thanks.) I like when teams hit threes. Also, I like when teams create havoc on defense by blocking a bunch of shots and forcing turnovers.

Unsurprisingly, eFG% and Adj. OE have the strongest correlation to a higher CBBWAR score at +0.94 and +0.86. Of the defensive stats, blocks (+0.30) have a higher correlation to watchability than steals (+0.10), which makes sense, because really good blocks are freaking cool. Tempo makes very little of a difference, which is ideal.

Your personal formula is probably different, which is fine. No two watchability metrics would ever be the same. But for me, this makes sense, and maybe it will for you, too. This metric is a work in progress, so don’t be surprised if/when it changes. For now, CBBWAR is what I’m using going forward to figure out which games and which teams are the ones I want to watch the most.

The initial CBBWAR rankings can be scrolled at the sheet below. A version where you can see the individual Z-scores is linked here:

As seen above, here’s the initial top 10 teams:

  1. Gonzaga (+14.19)
  2. Purdue (+12.5)
  3. Arizona (+11.48)
  4. Kansas (+10.56)
  5. South Dakota State (+10.15)
  6. Auburn (+10.08)
  7. Duke (+9.83)
  8. Iowa (+9.68)
  9. Davidson (+9.62)
  10. Colorado State (+9.46)

I genuinely like and would approve of this initial draft. To my eyes, I don’t see many teams missing from this top 10 (or top 20, to extend) that would be huge misses. Some of the selections will certainly appear strange, such as a 9-7 Memphis team being 18th overall or Santa Clara, the fifth-best WCC team, being in the top 12. What I would offer is this: no metric is perfect, and these two are early outliers. Still: as someone who isn’t a Memphis fan, this Memphis team is hilarious and amazing to watch. Santa Clara plays fast, scores efficiently, and is one of the best shooting teams in America. I kind of get it.

There are improvements to be made, certainly. I’ll expand on CBBWAR in coming weeks, with more changes after further testing and additional analysis with fair frequency. Hopefully, this gets us closer as a college basketball community to some sort of tool that combines team quality and subjective enjoyment. It will never be perfect or fully satisfactory, but I think it’s a decent start to expand upon.

Lastly, here’s a sheet that will be updated daily with the day’s most watchable games, per CBBWAR.