Understanding How ESPN Builds Its Poll
ESPN's Top 100 college basketball rankings aren't generated by some magic algorithm that spits out numbers. They're compiled by a panel of voters—media members, analysts, and former coaches—who submit ballots each week during the season. The process is straightforward on paper but messy in practice. I've watched this thing cycle through March for years, and there are some things about it that nobody really talks about because they're boring or inconvenient. Here's how it actually works. ESPN publishes a preseason Top 100 list around October or November. That list is based entirely on projections—roster returns, offseason movement, strength of schedule, that sort of thing. Then starting in February, they switch to a weekly Associated Press-style poll format, but only top 25. The Top 100 name lives on mostly as a brand from the preseason piece. What you're really looking at when people reference it is the preseason ranking system and the way ESPN staff debates who belongs where. The voting panel for the preseason list tends to include people like Jon Scheyer, Jay Bilas when he's still active, Lisa Byington, and a rotating group of conference reporters. Each voter submits a ordered list. ESPN tallies the points—first place gets so many, second place gets fewer—and produces the composite ranking. It's not weighted by conference or reputation. A voter from a mid-major beat gets one vote just like the national analyst.
What Actually Moves the Needle
The biggest misconception about these rankings is that individual player performance drives them. It doesn't. Team context does. And I mean that literally—when a guy drops thirty points in a single game, it rarely shifts the Top 100 unless it happens against a team already ranked near the top. A team losing three key rotation players to injury before the season starts? That gets you dropped ten spots immediately, regardless of how talented your incoming class is. The real predictors are:
- Returning production percentage—how much of your previous season's scoring and rebounding comes back
- Schedule difficulty in the early non-conference window
- Coaching changes, because a new scheme creates real uncertainty even when the roster looks fine on paper
- Losses to transfer portal departures, which voters penalize more heavily than Grad Transfer rule exits
One thing I learned the hard way: ESPN's system overweights name recognition. Programs with historical success—Kentucky, Duke, Kansas, Gonzaga—start higher than they deserve in October and stay there longer than logic would suggest. Meanwhile, a team like Creighton or Iowa State might be ranked five spots lower than their actual talent level warrants simply because they don't carry the same historical baggage in voters' minds. A couple of years ago I was tracking a mid-major program that had inherited a loaded roster through the portal. On paper, the returning production was nearly one hundred percent. But ESPN had them outside the Top 100 because their schedule looked soft and nobody on the voter panel had actually watched them play. I spent about two hours pulling Synergy highlights and box score data to build a case that they were top forty talent, then sent it to a couple of the conference reporters on the ballot. Zero response. The rankings came out and they were ranked sixty-seventh. They finished the season ranked in the AP top twenty-five. The workaround I use now is simpler. I don't try to change the rankings. I track the gap between what ESPN projects and what the actual season outcomes end up being. That gap tells you where the system is blind. Right now, the biggest blind spot is G5 programs with portal-heavy rosters. The voters see "non-power conference" and downgrade them before a single game is played. The second blind spot is coaches in their first year at a new school. The system assumes scheme disruption. Sometimes that's true. Sometimes the incoming coordinator is just the outgoing coordinator with a different email address.
Get the Full Details
How to Use These Rankings Actually
If you're using the Top 100 for bracket forecasting, ignore the preseason list after mid-January. It has zero predictive power once conference play starts. What you should do instead is track the weekly AP poll and compare it to KenPom and BartTorvik. When ESPN's top twenty-five diverges from those analytics platforms by more than five spots on average, that's usually a correction coming. The voter panel catches up within two or three weeks. For betting or depth chart purposes, the preseason list is marginally useful if you focus on the fifty-to-one-hundred range. Those are the teams where information asymmetry is highest. Everyone's guessing. The actual edges come from knowing which voters lean program-loyal versus which ones actually watch games in their assigned conferences. The list itself is available on ESPN.com under the college basketball section. You can subscribe to weekly update alerts if you want to track movement without refreshing the page constantly. There's no API for it, so scraping isn't really an option unless you want to deal with JavaScript rendering. Most people just use the published PDF or the webpage directly.
Where the System Breaks Down
The biggest flaw is sample size. The preseason poll is based on maybe one or two summer league events, some exhibition games that don't count for anything, and a bunch of film from the previous season. You're ranking a team that hasn't actually played together yet. I've seen three different voters put the same team in positions ranging from tenth to eighty-fourth in the same year. That's not a small variance. That's a system that can't agree on its own methodology. Another issue: the Top 100 includes teams that have no realistic path to the NCAA tournament. Several non-Division I programs, invitationals, and teams that are rebuilding get shoehorned in to pad the number. If you're using this for bracket prep, filter out anything outside the top sixty. The rest is noise dressed up as data. The rankings also don't account for injuries that happen between the vote and tipoff. I've had this come up twice where a starting point guard tore his ACL in September and the voter who put that team in the top fifteen hadn't updated his ballot. The ranking stood for a full week before anyone corrected it. The internet moves faster than the voting panel.
Bottom line: treat the ESPN Top 100 as a conversation starter, not a forecast. It reflects the collective opinion of a small group of people who don't all watch the same games. Cross-reference it with actual performance metrics and you'll find more useful signals than the ranking itself ever gives you.
