Bubble Watch: Art plus science

March has arrived, and so have clear bubble stakes

What the NCAA Tournament selection committee will look like in approximately four years (Image Credit: Meta)

OK, fine, let’s talk about it.

A lot of people have reached out about ESPN’s new Bubble Watch. They are mad about it. They want to know what we think of it — what we make of the way ESPN is leaning on its own metric, the Basketball Power Index (BPI), to predict and structure the portrait ESPN is providing of the prospective NCAA Tournament field. They want to know if we are also mad about it. Have we noticed? Do we care? What do we think? Are we also mad?!

Official response: Guys, it’s fine.

Bubble Watch does not have to be one thing. It can be different. This Bubble Watch, for example, spends most of its time every season only tangentially focusing on the actual “bubble,” such that said bubble even exists in January. When we begin, we don’t care too much about the nitty-gritty details, whether or how teams are going to make the NCAA Tournament, or how many games they need to win to get in because, let’s be real: Who cares? Lot of season left! We committed, long ago, when we realized we were going to be regularly writing this many words about college kids playing a sport, to make this something more. Bubble Watch is a useful skeleton, a vessel for expression, a way for us to structure thoughts about every relevant team in college hoops throughout the best part of the season. It is — until right about now every year — art over science by design.

You may like that. (You’re here, so let’s assume you do. Thanks.) Other people want more science: percentages, 10,000-run simulations, hard probabilities attached. That’s also cool! Folks can dislike BPI — it has some profoundly weird aspects that consistently throw off its rankings relative to the rest of the trusted predictive metrics, altitude (formerly?) chief among them, to the point that we kind of don’t look at it much anymore and kind of hope the committee doesn’t either — but you can understand why the idea of a purely BPI-based Bubble Watch is appealing, particularly (obviously) for ESPN. In a world full of bracket projections, it’s an interesting, different idea.

The only real problem (besides maybe giving unsophisticated readers the wrong impression): It ignores the nature of the committee.

This is a group of people in a room just arguing. Bracketologists are like (most) judges: They crave precedent. The committee are jurors, each of them deciding which evidence matters most. The whole bracket-prediction idea, from Joe Lunardi’s founding, is to analyze what these committees have done in the past to form educated guesses on the future. What do these folks care about? What does this group of college athletics administrators — the selection of which is slightly different every season — really want?

In the RPI era, when analytics were ignored, this was simple. There was one number; solve for that. Now that the NCAA keeps (correctly!) adding new metrics to team sheets, it has become more difficult than ever. Who knows what these people value? How much do big wins matter? How much do predictive analytics matter? What about record metrics? Who knows how the last-team-in bubble argument between one random conference commissioner and the other is really going to go? It has all changed pretty quickly.

Bracket predictions are closer to pure science than ever before. But they are also, not inconsequentially, still very much art. As long as there is a human committee doing this work, humanistic analysis will remain vital, and then eventually we will hand all of this over to a large language model that can accurately predict what the NCAA’s own LLM will decide on Selection Sunday. Imagine the efficiency. Imagine the accuracy! Won’t it be fun.

Anyway: This is exactly the time of the year that this Bubble Watch starts to trend more toward the science. The bubble is pretty well-established by now. There remains much to discuss, and much still up for grabs — but now, as you’ll see below, is when the stakes get clear.

Automatic bids from non-Bubble Watch (i.e. one-bid) leagues: 20
Locks: 28
Should be in: 5
Work to do: 21

The usual housekeeping:

• Some email providers impose limits on email length. This is a long file, so if your email cuts off, be sure to click through to the site itself (where you can use the navigation on the left side of your screen to scroll to conference sections quickly!) to read it all.

• There are almost certainly typos in the below copy. We are our only editor; this is a one-man show. If you spot factual mistakes or just think we should consider a team not on the page, get in touch in the comments or shoot me a note.

• NET and SOR are always current as of the previous day. Records are always up to date. Thanks as ever to Warren Nolan for his immensely helpful site.

• We wrote all of this, by hand, as nature intended. If you would like to support these efforts (and be able to comment, and join the live chats, and — at the founding tier — do a regular Zoom call with us to talk hoops) please consider a paid subscription today. Thanks!

• (Neil Paine does great work. Folks should check out his Substack here.)

ACC

When the 2025 Bubble Watch debuted, both Clemson and Louisville could have gone either way. Yes, sure, both teams had established themselves as a tier above the scant few bubble contenders in the middle of the league, but they were caught in a tricky position — not an obvious lock or No. 1 seed like Duke, while vulnerable to bad defeats against the robust collection of ACC teams with no chance of making th`e field. You could imagine a world in which either, or both, got dragged into the muck. Instead, admirably, both teams have made their separation real. Louisville’s overhauled roster is 16-2 in league play — it has lost exactly once since Dec. 14! — while the Tigers have matched them, losing at the Yum! Center but toppling Duke in Littlejohn along the way. Both will be dancing, and both will ensure yet another sorry edition of the ACC doesn’t face the added insult of one-bid-league talk. It’s something!

Lock: Duke, Clemson, Louisville
Work to do: SMU, North Carolina, Wake Forest

SMU (22-8, 13-6; NET: 41, SOR: 48): The Mustangs live! Sort of! Barely! Look: There is no real shame in losing at Stanford; Stanford is good, well-coached, and a road trip to the Bay Area, even from Dallas, is tough. (Also note: SMU barely survived Cal last Wednesday.) But if that loss called into question SMU’s chances of even remaining in this bubble conversation having made it to March with zero quad 1 wins, a home loss to Syracuse Tuesday night would have ended it once and for all. SMU trailed for most of the game; it took until the final minute of regulation for the Mustangs to finally overcome a dead Syracuse team. But they did, and so they’re still here, even if we think this is just about the longest shot on the page. SMU hasn’t beaten anyone. It’s March 5. Are you guys really sure you want to expand the tournament? Really?