It’s time to jettison the AP (press) and USAToday (coaches) polls in favor of a more valid way of ranking college basketball teams. The NCAA’s NET system does just that.
Why isn’t SDSU ranked #1 in the polls? Crazy question, huh? The Aztecs of the Mountain West Conference are undefeated, yes. But come on! They aren’t a Duke, Kansas, Kentucky, or Michigan State.
I don’t know anybody who believes SDSU is the BEST team in college basketball–even though no team seems to want to be #1 this year.
That goes to show you the competitive nature of today’s college basketball. At least that’s one conclusion. Another conclusion is this: only one, a few, or perhaps none of those four teams should have been #1 in the first place. That’s crazy?! Not really. Here’s why.
It’s how teams are ranked. Voters do it—the media for the Associated Press poll and basketball coaches for the USA Today poll. Voters rank the teams individually and, then, all of their rankings are compiled into a composite set of final rankings. In other words, humans use a variety of criteria (not a standard set) to rank the teams. The team that ends up being ranked #1 is #1 because that’s the consensus opinion.
Here’s a better way. The media and coaches gather to pick a set of data-based performance criteria by which to evaluate college teams. Then, data from actual games—game after game, week by week—are fed into a computer system. The system converts the outcomes across the criteria and creates a composite ranking–from #1 on.
I’d much rather have a team ranked #1 that way than based on opinion.
The NET is an excellent example of the system I prefer. NET, the successor to the RPI, is the NCAA’s newest system of ranking teams based on pre-selected criteria and on-floor performance. NET relies on “game results, the strength of schedule, game location, scoring margin, net offensive, and defensive efficiency, and the quality of wins and losses.” For those who’d like to have more info, Sports Illustrated offers a detailed breakdown of the system. Read it here. And here’s a recent and short video about NET.
Let’s look at how the two systems compare.
This week, Gonzaga is ranked #1 in both polls and #5 in NET. Who’s #1 in NET? It’s San Diego State. But how did SDSU do in the polls? The Aztecs aren’t a top-ten team in either poll. They are #14 and #15, respectively. That’s a big difference!
Another reason I prefer a system, like NET, is that it removes the politics of choice-making that are endemic in any system based on human selection. Let’s face it: certain teams benefit from a halo effect, while other teams don’t get the benefit of the doubt. Halo does not affect the NET, and it shows in the most recent NET rankings.
Consider these results: Butler (NET #3, #12 AP), West Virginia (NET #10, #22 AP), Wichita State (NET #11, unranked AP), and Dayton (NET #13, #18 AP).
A pattern is evident: human evaluators rank all of those teams lower than the NET. Just the opposite happens in the case of blueblood Michigan State, the pre-season #1 team. This week the AP ranks the Spartans #14, but the NET puts MSU at #22.
The results of the two systems aren’t always far apart. Ohio State, for example, is ranked #2 this week in NET and by AP voters. But it’s not differences in the rankings that bother me the most. I trust the NET. I don’t believe the polls.
Here’s how I see it.
If you can measure something well, measure it, and then use the results in your work. Gobs of rich game-related data are routinely collected in basketball, and NET is one example of how those data can be used. I say this: use a quantitative measurement system instead of human guesswork.
If San Diego State keeps winning, they’ll undoubtedly move up in the AP/UPI polls. But sustained winning won’t necessarily keep SDSU #1 in NET. That’s because the system isn’t based on just winning. It’s more complicated than that.
And, fans, that’s the way it should be.
