Mid-Majors In the NCAA Tournament

I wish I had gotten to this a few days ago when people actually were arguing about the NCAA brackets, for now focus is on the games themselves, but better late than never.

All-around jerk-off and awful college basketball commentator Billy Packer started a controvery last Sunday during the NCAA tournament selection show, along with partner Jim Nantz, by rudely criticizing the committee chariman (to his face) for their various selctions, notably a few teams from various mid and low-level conferences over squads from the BCS conferences:

Nance and Packer repeatedly pressed Craig Littlepage, chairman of the selection committee, on why the Valley received the same number of bids (4) as the Pac 10, Big 12 and ACC.

The CBS pair cited tournament records over the last few years and poor non-conference schedules by "mid-majors," specifically Bradley, as to why several power conferences should have been given a fifth or sixth bid.

The obvious retort is that you can't compare teams who, on average, are given low seeds, to team who usually are seeded higher. Packer's fallacy has been pointed out time and time again, but nobody, to my knowledge, has actually looked at the data. I don't care how the Missouri Valley Conference's collective tournament record looks against the ACC's - I care about how MVC teams do when seeded 10th in their regions compared to ACC teams seeded 10th. If everyone is right (which I find likely) and Packer is wrong (which is always likely), then the records of mid- and low-major schools will look simialr, seed-for-seed, as major conference schools. Here's what I found.

The most important question is: how do the two types of teams compare when they are on the edge of the tournament? This is where the argument usually resides. The second place school in the MVC, Mountain West, etc. who hadn't played a really good team all year is being evaluated against the 6th place team from the SEC who hadn't beat a really good team all year. Here are the first-round winning percentages in tournaments since 1985 of schools seeded 7-12 (the spots occupied by teams on the bubble at any given point), separated into those schools currently affiliated with BCS conferences, and those not:


Seed BCS Schools Non-BCS
7 .585 (31-22) .677 (21-10)
8 .467 (28-32) .435 (10-13)
9 .533 (24-21) .564 (22-17)
10 .390 (16-25) .395 (17-26)
11 .346 (9-17) .241 (14-44)
12 .300 (6-14) .313 (20-44)

Conlcusions: here we see that generally, over these seeds, there is no significant difference between how the two groups fare. Mid-majors actually have better winning percentages at seeds 7, 9, 10, and 12, with major conference schools faring better in seeds 8 and 11. I take this to mean that historically the committees have done a very good job of finding where to fit in the smaller schools among the big boys, given the fact that they rarely compete directly during the season.

For the record, major conference schools did signifcantly better at seed 13; BCS schools are historically 2-3 (.400) at this seed while non-BCS schools are 14-64 (.179). However, this is a very low sample of BCS schools at 13. If anything, this could mean that if the committe is picking an at-large team at the very last spot, they may want to go with a BCS school.

Next, I wanted to look at a similar comparison of teams seeded 1-6. Here, first-round records have less meaning, so: for seeds 1 and 2 I looked at percentage of advancement to the final four, and for seeds 3-6 I looked at percentage advancing to the Sweet 16.


Seed BCS Schools Non-BCS
1 .410 (32/78) .667 (4/6)
2 .224 (17/76) .125 (1/8)
3 .481 (37/77) .250 (2/8)
4 .449 (31/69) .400 (6/15)
5 .347 (26/75) .110 (1/9)
6 .352 (25/71) .400 (6/15)

Conclusions: here, an argument can be made that when the committees have seeded non-BCS teams in the top 5 lines, they have often overreached. It's possible that the numbers on the far-right column are too few to be meaningful; but, if they mean anything, it is this: beware of a top-five seeded team from outside of the major conferences - there's a good chance they've been over-seeded. This should give pause to those who are complaining about Gonzaga's 3-seed this year. Yes, they are a good team; no, they shoul not have been a serious #1-seed contender, if general history (and theirs specifically) is any guide.

A few question/problems with my data:

1. There may have been some minor mathematical mistakes in there. I had to gather it by hand from this website (thanks, Dave Anderson - great site). But, if one data point makes a difference in your conclusions, then your data are not very meaningful.

2. One could possibly find better measures to evaluate these groups relative to seeds. How have the two groups fared in sending 10-, 11-, and 12-seeds to the sweet sixteen? Is there a difference? (I doubt it, but it's worth checking.)

3. There is no evaluation of any change or trends over time. There has been a recent effort by the committees to place more non-BCS teams. Has this improved the relative performance, has it decreased, or no change? (I'm thinking no change - rhetoric about the "leveling" of college basketball aside, it seems to me from my looks at the brackets that mid-majors have always been fairly represented over the past 20 years.)

In the end, it's fair to conclude that Nantz and Packer were way off in their criticisms regarding this issue. Any fair comparison of the data, which I don't expect from innumerate commentators, yields the conclusion that at-large teams from lower conferences have just as much likelihood of succeeding as the mediocre major schools they are compared to.


Update: Todd Zywicki has more.
Follow-up: Mid-Majors in the NCAA Tournament, Continued
Follow-up: Sweet Sixteen Thoughts
Follow-up: Mid-Majors in the NCAA Tournament, One More Time

Share this

As a Bradley University

As a Bradley University alum, I was there in 1981 when Bradley was unjustly snubbed by the NCAA and Dave Gavitt, the Big East's cheerleader for life. The great combo of Mitchell Anderson and David Thirdkill were relagated to also-ran status in the NIT. Bradley easily trounced the competition and sent Purdue packing in the finals that year.
Nothing was sweeter than seeing a composed and talented Bradley Team dispatch Kansas last night with talent and poise, and not a last second prayer. Billy Packer can kiss my ass.
Go Bradley Braves! The Mid-Major alumni of the World will be watching and rooting for you against Pitt in the second round.
A victory over Pitt would be sweet vindication for us all!

(I posted this over at

(I posted this over at Volokh too, were I came across your post)
Trent,
Excellent post at your site - I was hoping someone less lazy than I would do this work. Two points - I agree about the non-BCS schools seeded five and higher and that's precisely why I say in my comment that it is clearly the case that the very best teams in college basketball are almost always from the power conferences. I'm glad Gonzaga managed to play to seed this year, but they haven't in the recent past, and Nevada was seeded too high this year.
Second, I think there is a more obvious explanation for the difference in the records of 13 seeds (and provides a good reason to omit the records from your overall data points). The vast majority of non-BCS 13 seeds are from conference tourney winners in one bid conferences. All of the Big Six 13 seeds are at-large teams. So in order to make the relevant comparison, you would have to weed out the at-large non-BCS 13 seeds and compare them to the BCS 13 seeds. This year, there were two automatic qualifier 13 seeds and two at-large 13s (all 4 were non-BCS). The at larges went 1-1; the automatic qualifiers went 0-2.

Conference calls Stephen

Conference calls
Stephen Karlson and Todd Zywicki note the big conference/little conference performance in the first round. Stephen notes that the Big 10 is down to one team, Ohio State. Add to that the Big 12, which is down to Texas. Here's the overall performance...

You may have already seen

You may have already seen the post from the ping above, since I also left a trackback on Volokh...but I updated it based on your numbers. Thanks for the doing the calculations! I think perhaps it would be better to break out things a little differently, but that's a quibble, since you took the trouble to do the work, and I was too lazy to do so.

If you have the still have the data in a convenient form it would be interesting to break it down by fraction of teams meeting or exceeding expectations based on seeding. In other words, a 9-16 winning round 1 exceeds expectations, while a 1-8 winning in round 1 meets expectations. 1-4 are supposed to make the Sweet 16, 1-2 the Elite 8, and 1 is supposed to make the Final Four.

Jim, the problem with that

Jim, the problem with that approach, I think, is that if you framed it as "percent playing to seed", the 8/9 games might really screw up the data. I have compiled BCS/non-BCS head-to-head in 7/10 and 8/9 games, and will post soon.

Mid-Majors: Over the past

Mid-Majors:
Over the past week or so we have heard much from two outspoken critics of the college basketball mid-majors--namely Billy Packer and Gary Williams.

A few years ago

See, someone was asking

See, someone was asking about basketball blogging way back when I was running my CFB Top 25, and now we know which of the CTLY folk has the marginal advantage in this regard :)

I really appreciate your

I really appreciate your research, and you are right about the small sample issues. Another issue might be that powerhouse teams probably do not schedule really strong mid-major teams because they are afraid of the stigma of losing to a team with a lesser name and they won't play on the road against those teams. (Schools with decent programs can't get or don't need the guarantee games that the Hartfords and New Hampshires of this world play.)

If media types want to make the NCAA's mostly an opportunity for the Big East, ACC, etc. to beat up on the conference champion of America East or the MEAC they're crazy. To end a great season for George Mason because of 1 loss would be a travesty. If the NCAA increases the weight of conference prestige as a selection factor, eliminates autobids for lower quality conferences or increases the number of play-in games, it must go to centralized non-conference scheduling.

[...] does the heavy lifting

[...] does the heavy lifting by beating the #1 seed before you have to play them. Prior post: Mid-Majors In the NCAA Tournament Prior post: Mid-Majors in the NCAA Tournament, Cont [...]