Discussion Question: Is the new RPI as bad as new Coke, and, if so, why aren’t college hoops fans making the same kind of fuss?

To be fair, the “new” RPI is really not all that new. This is the fifth season that the new formula has been in place. Before the 2004-05 season, the NCAA changed its RPI formula to give greater emphasis to road wins and more punishment for home losses. The basic idea was to encourage the big teams to go out on the road and also to reward those mid-majors who had to go on the road, since none of the big boys would come to their gyms.

The idea was noble; the implementation has been less so. The NCAA decided that the fair weight to be given to a road win or home loss was 1.4 and that a home win or road loss would be worth 0.6. So, basically, one road win is worth more than two home wins. This seems a bit extreme in theory, and it is in practice as well.

The main fault of the new system is not so much the credit it gives for road wins but the penalty it gives for home losses. A home loss against any team is worth 1.4 losses, meaning that even a team with a good home record in a tough conference — say, 6-3 — gets killed in comparison to a team in a mediocre conference going 7-1.

(As an aside, what follows is not intended to be an endorsement of Providence or an indictment of Florida but merely a discussion of the faults of the current RPI using those two schools as an example, since — among major-conference teams — they are most hurt and most benefited, respectively, by the formula.)

The main reason why this whole issue came up was because some Providence fans had inquired as to what made Florida such a better candidate for the NCAA Tournament than Providence. The Friars have more top-25 wins, more top-50 wins, no losses to teams outside of the top 100 (as Florida has to Georgia) and a strong record (9-7) in a conference rated much better by the RPI — and by anyone with a pulse — than the SEC. Florida has feasted with 15 wins against teams outside the RPI’s top 100. The Gators are tied with fellow SEC teams LSU, Auburn and Mississippi State among all 73 major-conference teams with 16 games against teams outside the RPI’s top 100. Providence has 12 such games and — to repeat — is undefeated in them.

The biggest difference is that ugly RPI. The Friars’ was 70 through Thursday’s action compared to Florida’s 40. Even after defeating the RPI’s No. 1 team, Pittsburgh, Providence’s RPI only improved by six spots. After all, it was just a home win. Florida also has 21 overall wins to PC’s 17.

Left woefully short on a convincing explanation of what made Florida’s résumé that much superior to Providence’s led me to investigate just how much the changes in the RPI formula affect how teams like Providence and Florida are viewed. As it turns out, the impact may be huge.

Here’s a list of the teams in the current RPI top 80 whose RPI rank is at least four spots better thanks to the new formula:

It’s not a surprise that most of the teams benefiting from the new RPI are mid-majors who go on the road a lot — that’s what the changes were intended to do. There are, however, two major-conference teams in there — Florida and South Carolina. More on those two in a bit.

Here’s a list of the teams in the current RPI top 80 whose RPI is at least four spots worse:

New Mexico and Davidson are the only two non-BCS teams in this list.

The question is why the advantage and disadvantage, and the answer to the latter is, in large part, home losses. Georgetown leads the way among teams on this list with five home losses. Eight others — including Providence, Nebraska and Northwestern — have four losses at home, and four other teams have three home losses. Florida, on the other hand, is just one of three major-conference programs without a home loss this season (Missouri and Pittsburgh are the others), and South Carolina has just one. The undefeated home record more than makes up for the Gators’ 2-6 road record, because road losses are almost seen as a virtue under the current system.

It may sound like I’m a biased Providence fan sticking up for my team, which is hovering near the at-large cut line. And that may be true, but the Friars and Gators are simply object lessons in the discussion of the greater problem with the RPI. PC is 13-4 at home with losses to Northeastern, Marquette, Villanova and Notre Dame. Certainly, the Friars would have preferred to lose fewer than four home games, and the Northeastern loss is the most distressing, but why should the outcome of that one home game be worth 13 points in the RPI rankings when a win against the No. 1 team is worth only six?

The answer to why is two-fold: 1) it shouldn’t, and 2) the home/road weighting is too extreme. Florida’s 15-0 home record includes exactly one win against an RPI top-80 team — South Carolina; Providence has three such wins, all against the top 50 and two against the top 25. Florida is being rewarded primarily for having a conference that lacks strong teams to send to Gainesville, though Tennessee and Kentucky will soon make appearances there.

My point is not to insist the Friars should be in the field and the Gators shouldn’t. My point is to identify the flaw of a ratings system that is so integral to the selection process. When my partner, Lukas, and I were re-building our at-large selection model this winter, we discovered that we were consistently overrating teams from mid-major conferences at the expense of major-conferences. Finally, Lukas decided to just add five points to the composite total of each “BCS” team, and — voila! — the model got much more accurate. Is this because the committee automatically favors the major schools? Maybe. But, in large part, it’s because the committee realizes that the new RPI formula inflates the track record of teams from the mid-majors at the expense of majors, because it gives too much credit for just going on the road and too much penalty for losing any game at home. Mid-majors — and Florida very much has a mid-major profile this year and is in the weakest major conference — are less likely to lose games at home, because the competition isn’t as stiff.

The problem with the committee automatically inflating the credential of the major teams — even subconsciously — because of the shortcomings of the RPI is that the new RPI formula doesn’t punish or reward major-conference teams and mid-majors equally. As we’ve seen, the “major-conference bonus” and the RPI are both benefiting Florida’s rating, so much so that the Gators moved just a few spots lower after a terrible loss at Georgia. Again, losses on the road barely injure you in the new RPI formula.

This is a message to the NCAA. Scrap the new RPI formula or adjust it to narrow the differences between road and home results — I suggest 1.2 and 0.8. The current formula may reward teams for traveling, but it’s creating an index that is far less accurate in judging the relative strength of teams. And isn’t distinguishing the relative strength of teams the whole point of the RPI?

But, Brendon, the RPI’s been fazed out a bit, hasn’t it? Isn’t it less important than it used to be? Everything I read tells me the RPI doesn’t matter as much, that the committee doesn’t use it, right? The problems is that the numbers don’t lie — there is nothing that does more to determine a teams inclusion or its seeding than the RPI. It’s true that the committee has been a bit less dependent on the RPI of late, in large part because of the flaws of the new formula, but it’s impact is still large, even if it’s implicit. Moreover, teams are judged by their records versus certain subsets of teams — 1-25, 26-50, 51-100 — which is determined, of course, by the RPI.

This years’s SEC — particularly Florida — will be a test case for exactly how much the committee values the RPI. Of course, the Gators might lose to Tennessee and Kentucky in the next 10 days and find out just how much it hurts to fall at home.

A lot of the problems for teams like Providence can be traced to ill-conceived or unfortunate scheduling. Here are some pointers for major-conference teams hoping to fill out a non-conference schedule.

Tricks to “beating” the RPI:

1) Go on the road against bad teams out of conference. That’s right — it pays to travel to small teams that you’re confident you can beat. Don’t go to Butler or Creighton, but take road trips to Dartmouth or Canisius or Bryant. Put together two- and three-for-ones with schools who usually don’t get the offer. The loss of income at the gate will be more than made up for by the gain in RPI points that might get a team into the Tourney.

2) Schedule teams from small conference who are likely to have good records. American, Cornell, Robert Morris, Weber State — these would have all been good teams to fill out a schedule. The key part of how they’ll affect your RPI is through their overall record, which should be good, and a team fighting for an at-large berth should still be able to beat each team.

3) I know fans love to see their home team take on another power in its arena, but it really doesn’t pay. Put together a home schedule out of conference that gives your team a very good chance to finish undefeated at home. If you’re in a good conference, you’ll have more than enough chances to lose to very good teams at home, so there’s no reason to invite a loss against a tough team out of conference. If you have to play a dangerous opponent out of conference, play that team on the road. It may sound illogical, but part of the “greatness” of the new RPI formula is its dismissal of logic.

4) Play in early-season neutral tournaments. If your roster can hold up against the competition, play in one of those made-for-TV events around Thanksgiving or Christmas. Losses won’t hurt as much as they would at home, and wins will help more.

5) Play in a weak major conference. This is obviously beyond a team’s control, but it would help if your team played in the fifth or sixth best conference. That way, it would still get credit for being a major-conference team without having to run the gauntlet of a truly testing schedule.