The 2012 season's very vocal complaints about the Ratings Percentage Index began with Scott Van Pelt's legendary rant and continued throughout the mock selection committee process, during which the NCAA tried to downplay the use of the metric, even as its prominence was plain. While maybe a bit louder than usual, beef with the RPI is nothing new. It feels like a mid-January rite of passage. We bubble watchers and bracketologists all say, "Well, here comes the RPI. We agree that it's stupid. Don't blame us. We're just trying to predict the field. And now back to your regularly scheduled programming."
When we complain about the RPI, we do so because we feel the metric is inaccurate. Even as the data set becomes larger, it doesn't adequately encapsulate the nuances of how good teams really are. The best way to do that, we all agree, are metrics that take into account scoring margin, or injuries, like Ken Pomeroy's adjusted efficiency and ESPN's BPI.
Smart coaches can game the RPI too easily, we say. But because this complaint is not new, and the means for exploiting the RPI (like playing top non-BCS teams, or scheduling brutal non-conference road games, and so on) are likewise old news, we rarely take a step back to explain to readers exactly how a really smart basketball program can manipulate the most important factor for NCAA tournament inclusion.
SI's Luke Winn decided to do just that this week. On Thursday, he explained what we mean when we say the RPI can be exploited -- Winn coins this the "exploitable gap" -- by coaches who study the system to find its cracks.
One such coach, Nebraska's Tim Miles, did exactly that in 2012. His Colorado State team had a bafflingly high RPI throughout much of the bubble season, even as it finished a respectable-but-hardly-mindblowing fourth in the Mountain West Conference. How? Luke got Miles and CSU athletic director to outline their strategy. It includes all the usual hallmarks of RPI smarts: The Rams avoided games against really high RPI teams (even preferring to schedule against Division II opponents to avoid a knock against the RPI) went to play Duke at Cameron Indoor in an "all-reward, no-risk" affair, and scheduled as many non-conference games as possible against teams that would finish well in non-BCS leagues, like Montana, Southern Miss, and Northern Iowa. There is also luck involved -- it helps to play teams that play difficult schedules, after all.
Winn also discusses Pittsburgh, where, per his study, coach Jamie Dixon has been the most consistent coach in the country at utilizing the makeup of the RPI to give his teams their best possible chances at tournament seed strength:
What Dixon likes to do for his home guarantee games, he says, "is play the teams that we think are the best picks to win the non-BCS conferences." These are the best "gap" teams, because they're beatable despite having high RPI returns. In 2010, Dixon beat five of them in Wofford (69 RPI), Wichita State (43), Kent State (47), Ohio (95) and Robert Morris (129). He only had one 250-plus RPI opponent (Youngstown State, at 271), either, and so it didn't matter that he played just one marquee game (against Texas) and lost it; the Panthers were in good standing due to their choices of non-BCS opponents. Despite their efficiency profile suggesting they were the quality of a 7-8 seed, they were a No. 3 on the strength of their RPI.
And there you have it. Go read Winn's piece for excellent details and a host of numbers; it's well worth your time.
Again, this is not necessarily new stuff; savvy coaches (John Calipari was always adept at this, especially at Memphis) have been using techniques like these to goose their RPIs for years now. Would that happen even if another metric was used in place of RPI? Probably. But the crude nature of the RPI, and the NCAA's over-reliance on it, makes Winn's "exploitable gap" possible.
That's why a selection process that takes into account a wide range of metrics -- that doesn't organize its team sheets and baseline rankings according to a system invented long ago -- would be vastly preferable to the current system. We still get good brackets out of the RPI-dominated system. But it's impossible to argue they couldn't be better.