Thursday, February 24, 2011

What If Middling Majors Played Mids Schedule (Part II)

Climbing back on that precarious bubble, it's time to follow up last week's post about Mids and Middling Majors by flipping the script and re-running the analysis in reverse.

This time, the Middling Majors get to join the Horizon League or Conference USA or The Valley and try their luck at the "cupcake" schedule of a Mid Major. As a refresher, the six Mid Majors selected for this analysis were picked on the basis of being on the Bracket Matrix's bubble as of February 16th (UAB, Butler, Duquesne, Colorado St., UTEP and Missouri St.). The nine Middling Majors were selected in the same way and include Michigan St., Boston College, Virginia Tech, Baylor, Oklahoma St., Alabama, Kansas St., Georgia and Marquette.

The methodology was quite simple - take each Mid Major team's offensive and defensive rating and the relevant variance and co-variance metrics and generate win probabilities against the various opponents on the substitute Middling Major team's schedule. Then, run 1,000 simulations to see what the Mid's profile would look like if it had played the Middling Major's schedule.

For all the detail, please read the previous post (linked in the opening sentence), but the findings seemed to indicate that Duquesne would have generated the best profile, even better than the Middling Majors, but likely would have been the only Mid Major in the top five. Three more, however, would have been in the next five (UAB, Butler and UTEP).

Now it's time to compare the results when we unleash the Middling Majors on those Mids' schedules.


Much like the last post, let's kick things off with a team from Alabama - the University of Alabama-Birmingham.

With three games remaining in its regular season, here is UAB's expected resume:

Overall Record 21 8
Conference Record 11 5
RPI Top 25 0 1
RPI Top 50 0 5
RPI Top 100 7 1
RPI 101-200 6 1
RPI 200+ 8 0

The biggest failing on the Blazers' resume is the 0-6 mark against the RPI Top 50. That must be balanced, however, with UAB's incredibly strong 7-1 record against teams ranked 51-100.

But how did the Middling Majors do with UAB's schedule?

Every Middling Major went an expected 0-1 against the Top 25 opponent (at Duke), but of the nine BCS teams, six went 3-2 against UAB's Top 50 opponents with just two (Boston College and Oklahoma St.) going 2-3 and one (Marquette) going 4-1.

Only Marquette matched the Blazers' 7-1 mark against teams ranked 51-100 with four more going 6-2, three going 5-3 and one (Oklahoma St.) going 4-4. Marquette and Alabama were the only two Middling Majors to record just one loss against RPI 101+ opposition with four of those BCS squads recording as many as three losses against the cupcake portion of UAB's schedule.

While the record against teams 26-50 probably sinks the Blazers relative to most of the Middling Majors, it has to be balanced against the bad losses most of those clubs incurred and which UAB managed to avoid.


Projected Resume:

Overall Record 21 9
Conference Record 13 5
RPI Top 25 0 3
RPI Top 50 3 0
RPI Top 100 2 1
RPI 101-200 9 4
RPI 200+ 6 1

The Bulldogs present almost the exact opposite case as the Blazers. Four BCS teams eclipse Butler's overall record (led by Marquette at 25-5) while two more match the Bulldogs' 21-9 mark. Only BC and Oklahoma St. failed to win a game against Butler's Top 25 foes, but no one could match the Bulldogs' 3-0 mark against the 26-50 teams.

Everyone except for Oklahoma St. matched Butler's 2-1 mark against the 51-100 ranked squads, leaving six of the nine BCS squads tied with Butler at 5-4 against the RPI Top 100.

The roads diverge at 101+, however. Marquette and Alabama lead the way, going 19-1 versus Butler's cupcakes, while Michigan State and Kansas State finish 17-3. Only one BCS team (Oklahoma State) falls to the Bulldogs' 15-5 mark. While the Top 100 record is very competitive, it is the plethora of bad losses that keeps Butler from posting a resume which would be competitive with most of the Middling Majors.


Projected Resume:

Overall Record 20 9
Conference Record 12 4
RPI Top 25 0 4
RPI Top 50 1 0
RPI Top 100 2 3
RPI 101-200 8 2
RPI 200+ 8 0

Despite its gaudy performance with the Middling Majors' schedules, the Dukes come nowhere close to matching up when those BCS squads switch roles.

Alabama and Marquette lead the way going 24-5, while Kansas St. (23-6), Virginia Tech and Michigan State (22-7) all finish at least two games ahead of the Dukes. Oklahoma St. and Boston College finished 1-3 against Duquesne's Top 25 opponents while the remainder of the BCS squads went 2-2. Seven of the nine Middling Majors went 4-2 against teams ranked 26-100. Overall, that means that Duquesne's 3-7 mark against RPI Top 100 opposition is tied for dead last and three games back of a majority of the BCS teams.

The 16-2 mark against RPI 101+ opposition was only eclipsed by Alabama and Marquette, while five of the Middling Majors finished at 15-3 or 14-4. It's nice that the Dukes have managed this portion of their schedule adeptly, but they just don't stack up at all in the quality win department.


Projected Resume:

Overall Record 20 10
Conference Record 10 6
RPI Top 25 0 5
RPI Top 50 2 1
RPI Top 100 4 2
RPI 101-200 8 2
RPI 200+ 5 0

The Rams vastly outperformed their showing from the first part in this series, recording an overall record that would tie for fifth (dead in the middle of the BCS sample) and only two games behind Alabama and Marquette.

The 0-5 mark against the RPI Top 25 ranks dead last among the Middling Majors - all of whom except for Virginia Tech (2-3) finished 1-4. The 2-1 mark against teams 26-50 was only matched by Virginia Tech, Alabama and Marquette and the 4-2 record versus the rest of the Top 100 was also exceeded by no team and matched by those previous three along with Michigan St. and Kansas St.

That leaves Colorado St.'s 6-8 mark against the RPI Top 100 tied for fourth, a game behind Alabama and Marquette and two games behind Virginia Tech.

Only Alabama, Kansas St. and Marquette post a better record (14-1) against the Rams' 101+ opponents, while Boston College, Virginia Tech, Baylor, Oklahoma St. and Georgia all posted worse ones.

The Rams might actually be the most competitive Mid thus far, despite being among the bottom of the group in part one. Colorado St. is the first team to balance a competitive record against top competition with a decent avoidance of black marks from bad losses.


Projected Resume:

Overall Record 23 8
Conference Record 11 5
RPI Top 25 0 1
RPI Top 50 1 2
RPI Top 100 4 2
RPI 101-200 9 3
RPI 200+ 8 0

UTEP's schedule provides a very interesting spread of average records from the top (Marquette - 27-4) to the bottom (Oklahoma St. - 21-10).

No one finished with an expected 1-0 mark against the Top 25, but five of the nine Middling Majors pulled off a 2-1 record against teams 26-50. The Miners' 4-2 mark against the remainder of the Top 100 is tied for fourth, behind Alabama, Kansas St. and Marquette (all 5-1).

It's the 101+ opponents where the teams start to divide. Alabama and Marquette went 19-1, while Oklahoma St. languished with a 15-5 mark. UTEP's own 17-3 record was tied for third with four Middling Majors.

The Miners clearly performed better than a few of the Middling Majors and about as well as a few more, but were well off the pace, as most Mids have been, established by Alabama and Marquette.


Projected Resume:

Overall Record 22 8
Conference Record 14 4
RPI Top 25 0 0
RPI Top 50 0 1
RPI Top 100 2 5
RPI 101-200 9 2
RPI 200+ 11 0

How did the whipping boy of Part I fare when the Middling Majors ran through its schedule?

Not surprisingly, not so well. Missouri St.'s record only topped Oklahoma St.'s (21-9) and tied Boston College, while finishing between two and four games behind five of the BCS squads. None of the Middling Majors went 1-0 against the RPI Top 50 opponents, but every BCS team went at least 3-4, with five going 4-3 and two more (Alabama and Marquette) going 5-2.

Missouri St.'s 20-2 mark against RPI 101+ competition was incredibly competitive - only Alabama and Marquette did better and six of the BCS teams did worse. The record against the RPI Top 100 is so poor, however, that Missouri St.'s ability to avoid bad losses isn't enough to make up ground on the Middling Majors which were much more successful against those top tier opponents.


After flipping the script and running the Middling Majors through the Mids' schedules, how would the teams all stack up on a relative S-Curve? As in Part I, all 15 teams (nine BCS teams and six Mids) will be ranked together with brief explanations.

1. Marquette
Outperformed every Mid and did as good or better with every Mid's schedule than the other Middling Majors did.
2. Alabama
Mentioned in the same breath as Marquette most of the way, the Crimson Tide consistently outperformed in this analysis.
3. Virginia Tech
Tended to do very well with the Mids' Top 100 opponents, but also posted questionable marks against the 101+ teams.
4. Colorado St.
The most balanced Mid, posting a competitive Top 100 mark, while keeping a solid record against 101+ competition.
5. Kansas St.
A step behind the top three BCS squads on this list, but still finished even with or eclipsed the Mids pretty consistently.
Much like Colorado St., the Miners posted decently competitive marks both against the top 100 and against the cupcakes.
7. Michigan St.
The Spartans didn't get too many mentions above, primarily because they avoided being both very bad and very good.
8. Butler
Would have been near the top of this list based on the Top 100 resume, but the bad losses knock the Bulldogs down a bunch of notches.
9. Georgia
The Bulldogs also stayed off the radar for the most part, but did pop up in some of the "worst" notes, dropping it to the third level of Middling Majors.

10. Baylor
Same as Georgia, the Bears stayed solidly in the middle with a few appearances on the negative side of the ledger.
11. UAB
Against teams 51+? The Blazers are easily a top five team on this list. But 0-6 against the Top 50 has relegated UAB to No. 11.
12. Boston College
The Eagles weren't terrible, but they were much closer to the worst Middling Major than the group ahead of them.
13. Duquesne
The Dukes looked great against the Middling Major schedules. When the Middling Majors took on Duquesne's schedule? The Dukes looked really weak against the Top 100.
14. Missouri St.
Much like the Dukes - the abomination at the Top 100 level is too much to ignore.
15. Oklahoma St.
The Cowboys were easily at the bottom each time and subsequently grab the bottom spot on the S-Curve.


After having performed this analysis both ways (BCS teams playing Mids' schedules and Mids playing BCS slates), there are two sets of summaries to compare.

Marquette, Virginia Tech and Kansas St. are the only three teams to perform consistently well in both runs. Aside from Missouri St., the Mids almost flipped top to bottom - Butler and UAB stayed in the middle, while UTEP and Colorado St. vaulted up and Duquesne tanked. Butler and UTEP were probably the top two Mids, because they were the most consistent, but neither performed very well in either analysis.

While the rankings are optically convenient, let's return to the original question. Are the relatively weak Mid Major profiles a function of true team quality or merely a derivative of the lack of opportunity?

The answer is, well, unclear, which may be the most interesting outcome. The Mid Majors didn't really knock the ball out of the park when they got more cracks at Top 100 competition and the Middling Majors didn't dispatch the Mids' opponents in decisive fashion.

The edge probably goes to the BCS squads, which had a higher portion of their group top the Mids in the Top 100 games than the Mids did when playing the Middling Major schedules. It's not a big edge, though, especially since those same BCS teams did a little more poorly on average against the 101+ opponents.

So, sorry college basketball fans. The numbers just won't let you cry foul in favor of those Mid Majors, and they won't let you exalt the virtue of a .500 record in a top league. The situation is just as cloudy as ever and unless there's a seismic shift in college basketball scheduling, a definitive answer in the Mid/BCS debate will be difficult to discern.

No comments:

Post a Comment