The AP Poll was once the best tool we had to measure the best teams. Those days are long gone.
Without speaking for any sport other than men’s college basketball, I can confidently say I’m in favor of eliminating the AP Poll and replacing it with something else.
That first part is important: I concede the AP Poll may still be necessary for women’s college basketball, where advanced stats are still lagging (shoutout herhoopstats for making significant headway there). Maybe in other sports too, where I’m less knowledgeable.
But if you want to know if Team A is better than Team B in men’s college basketball, where are you going? KenPom? Torvik? The NET? BPI? Massey? Take your pick — they’re all computer models that do a better job of representing the sport than the AP Poll.
So do individual rankings, like Eamonn Brennan’s power rankings or Gary Parrish’s Top 25 and 1. Neither Brennan nor Parrish is an AP voter, but both reasonably could be — the AP has a number of national voters, including Seth Davis, who works with Brennan at The Athletic.
The problem is that the AP Poll is too heavily reliant on other writers who don’t put in the work to have a reasonable idea of the broader basketball landscape and are too married to however their poll looked in the previous week to be able to step back and re-evaluate teams before they vote.
It leads to polls like this week’s, which seems to capture the general hierarchy just fine but on further examination, makes no sense whatsoever.
Yes, the top six are tough to argue, the teams in the teens are all quality, and 20-25 deserve recognition.
But let’s take a closer look.
Charleston, at 21-1, comes in at No. 18, despite zero Quadrant 1 wins. The Cougars are 73rd in KenPom and 43rd in the NET.
Now drop down a couple of spots and look at 18-4 Saint Mary’s. The Gaels don’t have as strong a win-loss record, but they have a win over San Diego State (KenPom No. 27) and are ranked in the top 10 nationally in both KenPom and the NET. Providence, at 15-5, has two wins over KenPom top-10 teams and three Quadrant 1 wins. The Friars, however, lost one game each of the last two weeks. By the unwritten, nonsensical rules of AP voters, that means they must move below less deserving teams.
We can parse every week’s poll exactly like this, then look at individual ballots to see who is not paying attention. Or we could just find a better gold standard to measure college basketball teams.
Why the AP Poll Matters … For Now
People love to say the AP Poll doesn’t matter. In some ways, they’re right. You don’t get any extra advantages by earning a higher ranking. The Selection Committee doesn’t use the AP Poll to put teams in the NCAA Tournament or to seed them. You don’t get a trophy for being ranked No. 1.
It still matters.
Go to NCAA.com and look at their men’s college basketball scoreboard. Which teams populate up top?
Turn on ESPN and wait for college basketball to come up on the bottom line. Which scores do they give first?
This might not matter to programs that are always in the national spotlight — Kentucky, Kansas, Duke, etc. — but for Saint Mary’s, who just entered the Top 25 last week despite being top 25 in KenPom since early December? To have their name up there, showing they’re right on par with Gonzaga, that helps. It also helps Florida Atlantic from an awareness (and eventually fan interest, recruiting, and university enrollment) perspective as it looks to build a basketball identity.
The thing is, the AP Poll doesn’t have to be what websites and networks use to prioritize the best teams. It could be the NET, KenPom, or some metric that hasn’t been invented yet.
I’ve scoured the entire US Constitution and there is not a single amendment declaring the AP Poll law.
Something to Argue About
The other argument for the AP Poll is that it provides a topic of discussion for fans and media from November to February when college basketball has to compete alongside the NFL, college football, and the NBA. That’s true, and let the record show I am all for shit-talk among rivals before we even hit Thanksgiving.
But college basketball media has already created, monetized, and beaten into the ground a better version of this in the form of Bracketology. It started with Joe Lunardi in his Bracket Bunker and eventually, every major sports outlet, fan site, and brackets dot blogspot dot com had a resident bracketologist. Those early season brackets are exactly as speculative as the AP Poll, but in a way that lets the reader daydream about NCAA Tournament glory. It’s better.
We don’t need Alabama fans lording their AP Poll superiority over Kentucky fans when Lunardi has the Tide as the 1 seed in the Midwest and Kentucky as the 12 in that bracket, needing to get past Penn State in Dayton, UConn in the first round, and potentially Baylor after that. That’s way more fun — and it lets Alabama fans waste time at work planning a hypothetical trip to Kansas City to see their team play regional games.
As for Kentucky fans, well, I covered the First Four a few years ago. My DMs are open if you’d like travel tips for Dayton.
Where To Go From Here
We all know how averse the NCAA is to change. In order to alter one word of the rule book, there needs to be a committee to explore the idea of creating a committee to make a committee on changing it in the next five years. Or something like that.
The good news is this isn’t an NCAA issue — it’s more a need to change the general consciousness of the college basketball world (somehow, maybe an easier task than changing something in Indianapolis?). The AP has every right to put out its poll and will continue to do so, but that doesn’t mean it needs to be the standard measure of quality all season. We can even look to the NCAA to figure out a better option.
Once it became clear that the RPI was an outdated metric — ok, maybe a decade after that became clear — the NCAA convened analytics experts and college basketball heads to create something new. As a result, it introduced the NET in 2018 for the men and in 2019 for the women. The NET, as it stands now, isn’t an ideal replacement for the AP Poll, as it needs a couple of months of data before it becomes a reliable metric. That means there’d be no preseason NET rankings or anything substantive in November or December. KenPom and Torvik have the same issue, but less dramatically so — and certainly less dramatically than the AP Poll.
Take North Carolina for example. The Tar Heels were preseason No. 1 and four weeks into the year were ranked 18th despite a 5-2 record and not looking good in any of their seven games. UNC has since gotten it together, but at the time, there was no justification for ranking it at all, other than voters’ own preseason biases. At that point, UNC was 22 in KenPom — still too high, for similar reasons, but less so.
That’s all to say that I’m not claiming to have the answer here. Maybe we need to get a bunch of smart people in a room again and come up with weekly rankings equivalent of the NET that makes a degree of sense year-round. Maybe we just need a new crop of AP voters to reimagine how the poll works and to define criteria that voters should use. I don’t know.
What I know is that the current AP Poll just doesn’t serve the same purpose that it used to. It’s time for change.
- Are The Timberwolves Prepared For An Anthony Edwards Contract Extension?
- Boardroom NIL Report Card: Kentucky Quarterback Will Levis
- Jerick McKinnon Contract & Salary Breakdown
- How Postgame & Reebok Define the New Age of NIL Partnerships
- Imagining the Future of FanDuel With Company President Christian Genetski