During last year's College World Series, I wrote a couple of posts discussing the extreme geographical imbalance in college baseball strength. The bottom line is that the western part of the United States has dominated college baseball both historically and recently.
The NCAA's response to this obvious situation has been simultaneously consistent and duplicitous. A difficult combination to achieve, but the NCAA is an amazing organization. The consistent part comes from their ongoing efforts to do everything possible to help northern and eastern teams become more competitive. The start date for the season was moved later in the year to remove some of the theorized advantage gained by southern teams from better weather. And when it comes to the postseason tournament the NCAA annually delivers regionals and supers to eastern teams in an effort to make those stages truly regional in nature. It sounds like a decent argument until you realize that the basketball committee makes no such effort. The possible explanation that it's related to travel costs may have some credibility at the lower levels (the Division II tournament, for example, is even more strictly regionalized), but I don't buy it for Division I.
The duplicitous part? Part of the explanation and argument given for the season start change was the interest of fairness. The NCAA wanted things to be more fair to the northern teams because of the inherent disadvantage they faced that they could do nothing to overcome. And yet when it comes to rewarding teams for actual season performance, fairness goes out the window. I can understand wanting to build the sport into a more national commodity. There are ways to do that, though, that don't involve ripping off the better teams in the current setup. They could continue with season start changes, invest in facilities in the north and northeast, or even require baseball coverage levels as part of the basketball tournament TV package. Clearly they couldn't require massive coverage, but something. Knowingly preventing deserving teams from access to, or proper seeding and hosting opportunities in, the postseason tournament is just wrong. And it's not just the historical numbers that tell me they should know better, the selection committee also seems to choose one or maybe two western teams every year that get to be seeded where they belong isntead of in line with their RPI.
Speaking of the RPI, that's part of the problem. College baseball does not have nearly enough interregional play to make that formula work. The NCAA almost has to know this as well, both because of the evidence in tournament seeding mentioned above, and also because it's been written about extensively by Boyd Nation and NCAA committee chairmen have discussed the RPI's questionability. In fact, the NCAA's director of statistics once said "Every year we hear complaints, but nobody ever comes up with a better system." That's a fairly flabbergasting statement even if it was made a decade ago. That being said, it's been a decade so maybe the NCAA is interested in a better system. Do I really think so? No, continuing to use the RPI helps the NCAA in their efforts to prop up northern and eastern teams. Boyd Nation's ISR is one of the better systems out there, and I've also developed Adjusted Winning Percentage. Neither of these systems use margin of victory, which of course would enable more accurate rankings but runs counter to NCAA mandate. Both the ISR and AWP systems are straightforward and require no information beyond who played and who won. So let's see how bad the RPI really is, and if it really favors northern and eastern teams as much as I'm implying. Keep in mind, there's no coded bias based on location, it's simply that the RPI will necessarily overvalue teams from the weaker region in an overall league with little interregional play.
Click here for a table listing all NCAA Division I baseball teams along with their ISR, AWP, and RPI ratings. Also shown are conferences and regions, the latter being defined by the NCAA. More relevant to this conversation, though, is this page showing the ratings broken down by region and conference. The table is sortable by clicking on an appropriate column heading, and it becomes immediately obvious how incredibly ripped off the western teams are by the RPI. Lest anyone think that this analysis is being posted due to homerism, the Big 12 is right in the middle of the RPI's bias and is relatively unaffected as is Texas specifically. But let's take a look at what effect this discrepancy can cause when tournament time comes.
For the rest of this post I'm going to assume that each conference's ISR leader will get an automatic bid to the tournament. So automatic bids go to the following teams (numbers in parentheses are ISR/AWP/RPI ratings):
America East - Maine (180/190/191)
Atlantic 10 - Charlotte (112/108/100)
ACC - Virginia (9/9/3)
Atlantic Sun - Florida Gulf Coast (62/67/42)
Big 12 - Texas (3/3/4)
Big East - Louisville (11/11/5)
Big South - Coastal Carolina (8/8/7)
Big Ten - Michigan (76/83/68)
Big West - Cal St.-Fullerton (4/4/10)
Colonial - James Madison (117/113/70)
Conference USA - Rice (27/28/28)
Horizon - Wright St. (148/179/176)
Ivy - Dartmouth (169/175/129)
Metro Atlantic - Manhattan (133/154/88)
Mid-American - Kent St. (113/128/138)
Mid-Eastern - Bethune-Cookman (145/139/112)
Missouri Valley - Wichita St. (57/58/61)
Mountain West - TCU (5/7/13)
Northeast - Bryant (215/218/201)
Ohio Valley - Jacksonville St. (128/117/118)
Pac-10 - Arizona St. (1/1/1)
Patriot - Army (174/178/136)
SEC - Arkansas (6/6/6)
Southern - Charleston (52/41/37)
Southland - Texas St. (37/39/38)
SWAC - Southern (237/232/247)
Summit - South Dakota St. (102/116/145)
Sun Belt - Western Kentucky (41/47/40)
West Coast - San Diego (21/20/22)
WAC - Fresno St. (34/32/64).
That leaves 34 at-large bids into the tournament. I sorted the rankings by each of the three ratings and then filled in the bracket going by strict ranking order until I had 64 teams. I called the last 8 teams in "Bubble In" teams and the first 8 out "Bubble Out" teams in each scenario.
First, the teams that were unaffected in any way and were solid non-bubble at-large teams in all three ratings systems were:
Arizona, Auburn, California, Clemson, Connecticut, Florida, Florida State, Georgia Tech, LSU, Miami (FL), Mississippi, New Mexico, Oklahoma, Oregon, Oregon State, South Carolina, Texas A&M, UCLA, Vanderbilt, and Washington State.
Alabama, Louisiana-Lafayette, North Carolina, and Virginia Tech were Bubble In teams in the ISR and AWP systems but solidly in with the RPI. Stanford and UC-Irvine were RPI Bubble In teams but solidly in with the ISR and AWP. Kansas St. was a Bubble In team in all three systems. That makes for 27 of the 34 at-large teams that all three systems agreed should be in the field of 64. But what about the last seven spots? Here are each of the three systems' selections for those slots as well as their bubble out teams in parentheses. Bubble In teams are shown with an asterisk.
ISR - Long Beach St.*, Pacific, Pittsburgh*, Portland, Southern Cal*, UC-Riverside, Washington (Baylor, Florida Atlantic, Kansas, Kentucky, Nevada, Southeastern Louisiana, Southern Miss, Tennessee)
AWP - Florida Atlantic*, Pacific, Portland, Southern Cal*, Tennessee*, UC-Riverside, Washington (Baylor, Kansas, Kentucky, Long Beach St., Nevada, Southeastern Louisiana, Southern Miss, Texas Tech)
RPI - Florida Atlantic, Kentucky, Liberty*, Pittsburgh*, Tennessee*, Texas Tech*, The Citadel* (Baylor, Boston College, Elon, Georgia Southern, Kansas, South Alabama, Southeastern Louisiana, Washington)
Notice the geographical breakdown of the changes. Pacific, Portland, and UC-Riverside go from being solidly in the field in both the ISR and AWP systems to not even being on the bubble with the RPI. Long Beach State goes from Bubble In and solidly in with the ISR and AWP, respectively, to completely out with the RPI. Washington goes from solidly in with the ISR and AWP to Bubble Out with the RPI. Nevada goes from Bubble Out in both to nowhere near in the RPI. Southern Miss does the same.
Which teams improved their lot with the RPI? Tennessee was Bubble Out with the ISR but Bubble In with the AWP and RPI. Texas Tech was completely out with the ISR, Bubble Out with the AWP, and Bubble In with the RPI. Florida Atlantic was Bubble Out, Bubble In, solidly in. Kentucky was Bubble Out in both the ISR and AWP but solidly in with the RPI. Pittsburgh was solidly out with the AWP but Bubble In with the ISR and RPI. Liberty and The Citadel were both completely out with the the ISR and AWP but got in on the bubble in the RPI. Boston College, Elon, Georgia Southern, and South Alabama were nowhere near with either the ISR or AWP but managed to get on the bubble in the RPI.
No matter how you look at it, the RPI overstates the strength of eastern teams at the expense of the western squads. And it seems the NCAA not only recognizes this but is in fact thankful for it. Being able to cite RPI rankings helps justify the unfairness we see every year when tournament selection time rolls around. At a higher level, here are some selected conference and region comparisons of the three systems:
ACC - In both the ISR and AWP systems, the ACC is given 7 bids and is in line for 3 regional hosts but zero national Top 8 seeds (7/3/0). In the RPI, the ACC also received 7 bids but is in line for 5 regional hosts and 2 national Top 8 seeds (7/5/2).
Big West - Using the bids/hosts/seeds format, in the ISR the Big West received 5/1/1, in the AWP 4/1/1, but in the RPI only 2/1/0.
Pac-10 - ISR 10/5/2, AWP 10/4/2, RPI 8/2/1.
Atlantic Region - ISR 10/4/1, AWP 10/4/1, RPI, 12/6/3
South Region - ISR 11/3/2, AWP 13/4/2, RPI 14/4/2
West Region - ISR 20/7/4, AWP 19/6/4, RPI 14/4/1
I'm completely in favor of the NCAA's goal of making college baseball a more national sport. The problem is that it's not going to happen in the foreseeable future. The rule changes and biased selection process haven't spread the game's popularity yet, so let's get rid of the postseason chicanery and get back to making the tournament about picking the best teams and crowning a deserving champion. Then we can work on the tournament format itself.