Yesterday, Joel wrote an article about how difficult it is to tell where a team really ranks in college basketball. He focused on the differences between how the RPI and the Pomeroy rankings see results. This matters, obviously, when it comes time to seed a tournament. Who should really be in and who should be out? This sparked quite a conversation on Twitter.
To dig more deeply into the problem, different sources see teams vastly differently. While for now this seems like a minor foible, it will make all the difference in the world come March 11th. Some teams illustrate the problem facing the selection committee.
Take Rhode Island, for instance. In the RPI, with it’s weight so firmly on only result, they are ninth. That’s well in the range for an at large bid. Actually, it’s basically a shoo-in. The Pomeroy rankings value strength of schedule and margin of victory more. There, the Rams are 28th, which is right on the cut line for an at large. The BPI (from ESPN) weights winning as a skill and emphasizes strength of record, has the Rams at 29th.
A quick layman’s glance at Rhode Island shows home win by seven over Providence as their best win [edit: a one point win over Seton Hall should also be listed here], and away to Alabama by four as their worst loss. The Rams also lost a reasonably close game on a neutral floor to Virginia and by seven at Nevada. URI’s schedule isn’t that tough, though, with their next best win being St. Bonaventure and then the dreck of the A10. They are 15-3 (7-0).
Ah, the Bearcats. The RPI has them at 24th, KenPom has them at sixth, same as the BPI. RPI has them probably in with an at-large, while KenPom and the BPI see them as something akin to a two seed. The RPI is unimpressed with the Bearcats strength of schedule (who isn’t?) and docks them for it. KP and BPI see a team that is blowing out all comers and rewards them for it.
Reality shows that UC is going to hang its hat on a road win against UCLA and a home hiding of SMU. Beyond that, they’ve lost to the only other two decent teams they’ve played. For what it is worth, ESPN projects the Bearcats as a four seed, which means they don’t entirely believe their own system or the RPI.
Here’s one that the committee will really wrestle with. According to the RPI, the Bulls are 28th. That very much merits serious consideration for an at large bid. In fact, 28th would be the second highest RPI ranking to ever miss the tournament. Buffalo is winning, too, they are 14-5 and undefeated in the MAC. What does that earn them from the metric systems? 70th from KenPom and 95th from the BPI. A team that RPI essentially has in the tournament, the computers think is mediocre at very best.
Looking at it, Buffalo hasn’t beaten a decent team yet. Their best win is likely UAB. Beyond that, they are playing poor teams and sometimes just squeaking by. Their win over a very bad Delaware team was by three, and they only beat Jacksonville St at home by five.
That’s a quick three team look at the task set out for the committee. Their own metric thinks Buffalo and UC are basically the same team and thinks Rhode Island is better than Purdue. The computers don’t help though, because they think UC is in contention for a one seed and Gonzaga (RPI: 56th) is in the top 10.
There is one tool, though, for a college basketball fan lost in all this mess. The Massey Composite takes over 100 ranking systems and aggregates them. For the teams we’ve looked at, it has UC at 10th, Rhode Island at 25th, and Buffalo at 61st. Gonzaga, something of a late addition to this article, comes in at 16th. Maybe the wisdom of the crowd is not to be discounted.