Replace the committee with basketball games

storm

This is a piece about postseason bids, and this is not a picture of athletic directors in a conference room. It is instead a picture of people happy about their team. (startribune.com)

Selection committees are college basketball’s original sin.

The first modern postseason tournament was arguably the eight-team National Intercollegiate Basketball Championship Tournament in Kansas City in 1937. It had a selection committee. The following year, the inaugural National Invitation Tournament was held with six teams at Madison Square Garden. It had a selection committee.

Finally, in 1939, due largely to pushing and cajoling by Ohio State head coach Harold Olsen, the NCAA held its first tournament, an eight-team affair that culminated in a championship game in Evanston, Illinois. It had a selection committee.

At least the creators of the NIT had the decency to foreground the subjective nature of the endeavor in their event’s very title. The NCAA tournament has been an invitational now for decades, albeit one with 32 spots reserved for automatic entrants certified by their conferences.

We should learn from and follow through on the example set by these automatic bids. We should make each tournament spot an outcome to be won through unmediated basketball performance instead of a favor to be granted through jury deliberation.

People in the 1930s needed committees to put on these tournaments. We no longer do.  

Not only do we not need the committee, continuing to have one at the dawn of the 2020s carries significant opportunity cost in terms of lost energy and excitement during the regular season. We could be having the equivalent of 36 four-month pennant races for each at-large bid. It would be riveting sports theater, and it could start with the first game of the season. Instead, we spend the regular season in a speculative fog until, at last, we’re granted the privilege of sitting passively outside the committee room to await 36 verdicts. Why?

Whether your preferred metric is wins above bubble (WAB), strength of record (SOR), or something similar, any team’s record can be evaluated according to how difficult it would be for other teams to achieve the same result against the same opponents in the same venues. Wins are the key metric, as they should be, and teams can excel whether their schedules turn out to be grueling or unexpectedly soft.

For illustrative purposes, let’s take WAB out for a retrospective spin, courtesy of Bart Torvik’s handy archive. Here’s what one wins-loaded metric would have done with the 2019 NCAA tournament field:

                                    WAB seed   Actual seed
1   Duke                 Auto           1           1
2   Virginia             At-large       1           1
3   Michigan State       Auto           1           2
4   North Carolina       At-large       1           1
5   Tennessee            At-large       2           2
6   Michigan             At-large       2           2
7   Kentucky             At-large       2           2
8   Kansas               At-large       2           4
9   Houston              At-large       3           3
10  Florida State        At-large       3           4
11  LSU                  At-large       3           3
12  Texas Tech           At-large       3           3
13  Gonzaga              At-large       4           1
14  Kansas State         At-large       4           4
15  Wisconsin            At-large       4           5
16  Purdue               At-large       4           3
17  Auburn               Auto           5           5
18  Cincinnati           Auto           5           7
19  Buffalo              Auto           5           6
20  Virginia Tech        At-large       5           4
21  Mississippi State    At-large       6           5
22  Iowa State           Auto           6           6
23  Maryland             At-large       6           6
24  Wofford              Auto           6           7
25  Nevada               At-large       7           7
26  Villanova            Auto           7           6
27  Iowa                 At-large       7           10
28  Washington           At-large       7           9
29  Marquette            At-large       8           5
30  Minnesota            At-large       8           10
31  Oklahoma             At-large       8           9
32  Louisville           At-large       8           7
33  UCF                  At-large       9           9
34  Utah State           Auto           9           8
35  NC State             At-large       9           N/A
36  UNC Greensboro       At-large       9           N/A
37  Temple               At-large       10          11
38  New Mexico State     Auto           10          12
39  Belmont              At-large       10          11
40  Murray State         Auto           10          12
41  VCU                  At-large       11          8
42  TCU                  At-large       11          N/A
43  Ole Miss             At-large       11          8
44  Baylor               At-large       11          9
45  Ohio State           At-large       11          11
46  UC Irvine            Auto           11          13
47  Florida              At-large       12          10
48  Syracuse             At-large       12          8

49  Liberty              Auto           12          12
50  Saint Mary's         Auto           12          11
51  Yale                 Auto           13          14
52  Vermont              Auto           13          13
53  Oregon               Auto           13          12
54  Georgia State        Auto           13          14
55  Old Dominion         Auto           14          14
56  Northern Kentucky    Auto           14          14
57  Abilene Christian    Auto           14          15
58  Northeastern         Auto           14          13
59  Saint Louis          Auto           15          13
60  Montana              Auto           15          15
61  Colgate              Auto           15          15
62  Gardner Webb         Auto           15          16
63  Prairie View A&M     Auto           16          16
64  Bradley              Auto           16          15
65  North Dakota State   Auto	        16          16
66  Fairleigh Dickinson  Auto           16          16
67  Iona                 Auto           16          16
68  NC Central           Auto           16          16

Before we pick this apart like we do with any bracket from anywhere, pause to consider what could be gained from abandoning the jury model entirely and instead using a straight proxy for wins.

The NFL doesn’t have “selection criteria” for its playoffs. It has tiebreakers. We could do that, too. We could have the field of 68 selected the instant the final buzzer sounds in the title game of the last conference tournament. This could be settled on the court instead of in the committee room.

As for our sample 2019 bracket, actual participants Arizona State, Seton Hall, and St. John’s are nowhere to be found. In a WAB world they would have been bumped to make room for NC State, UNC Greensboro, and TCU.

No championing of one team ranked 40-something in the nation over another will ever be winningly conclusive, of course, whether that case is made by the actual committee, a nifty metric, or by a given team’s fans. The best we can hope for, perhaps, is transparency, consistency, and, at long last, real-time standings showing who’s actually in and out from November right through the last game of the last conference tournament.

Note additionally the rough treatment handed out to Gonzaga, bumped all the way down to the No. 4 line. The 2019 Bulldogs were indeed a tough evaluative nut to crack, historically dominant, truly, against the non-SMC WCC but 5-4 against the best that KenPom throws at you (Duke, Tennessee, North Carolina, Baylor, Florida State, Texas Tech, and three games against the Gaels).

Perhaps the larger point raised by Mark Few’s men, however, is that, in straight-playoff no-jury settings, we make room in our psyches every day of the sports week for the team we know is great but that has underperformed slightly during the regular season. Part of why we watch is to see that particular drama unfold in the postseason.

Nor is this some kind of structural measurement feature born of trying to assess the equivalent of Kentucky in a mid-major conference. In fact, if WAB had its way it would have bumped Gonzaga’s actual 2018 No. 4 seed up to a 3.

Indeed, a proxy for wins is kinder to mid-majors than the committee’s been, at least at the level of the bubble. Over the last five years, the method proposed here and actual reality have disagreed on 18 at-large selections. Seeing this glass as half-full, that’s a high degree of agreement (162 of 180 bids). But, if we must focus on contested terrain, 14 of the committee’s 18 preferred teams hailed from the six major conferences. The rest are VCU, Temple, and Tulsa in 2016, and Cincinnati in 2015. (The Golden Hurricane that season were perhaps a particularly adventurous choice.)

A straight win proxy has nothing against major conferences and, in fact, said method would have given the nod to the likes of the Wolfpack and Horned Frogs last year, Nebraska, Marquette, and Louisville in 2018, South Carolina, Florida State, and Florida in 2016, and Miami in 2015. Nevertheless, the four-fifths of Division I below the Pac-12 does fare better with WAB than it has with reality of late. In the alternate reality outlined here, UNC Greensboro last year, Saint Mary’s in both 2018 and 2016, Middle Tennessee in 2018, Illinois State and Monmouth in 2017, Valparaiso in 2016, and Colorado State and Murray State in 2015 would have all gone dancing.

To be sure, the committee’s done a pretty good job all things considered, particularly (go figure) in recent odd-numbered years. Moreover, the straight wins proxy trotted out here agrees with present custom on no fewer than 95 percent of bids and on fully 90 percent of at-large selections over the last five tournaments.

Zeroing in further, the seeding that would result from this type of approach would be within one line of what has actually happened 74 percent of the time. Purely in terms of selection and seeding, moving to a WAB- or SOR-type method would not constitute a revolutionary change.

It would, however, make its impact felt on every aspect of the season before Dayton. Selection would become real-time, and each game’s true impact on the bubble or on the race for the four No. 1 seeds would be known day by day. Conference tournaments in particular would be turbo-charged as we’d know the exact stakes in play and as bids change hands hour by hour.

Then, with the field already selected, Selection Sunday could transcend the mere reading of a static list and instead be as entertaining and unpredictable as an NBA draft. The bracket could build itself before our eyes according to choices made in real time by 68 programs.

Is it better for No. 3 seed Michigan to play in a bracket of death close to home or to head out West to face a more favorable No. 6 seed? Why is a third party making this calculation? Let Juwan Howard decide, and let him do so live on CBS and on the clock.

The NCAA tournament used to be referred to unselfconsciously as “the playoffs,” a turn of phrase that made sense before the mid-1970s when a bid typically entailed winning your conference. This linguistic habit lingered on into the 1980s before falling into desuetude. Perhaps new practices could awaken old phrases.

This week the NET had a bit of a messy rollout, though the problems were mere reporting bugs downstream from an otherwise sound (if needlessly shrouded) algorithm. The technical hiccups of December will be long and properly forgotten come Selection Sunday. Still, NET rollout day, only in its second year, has already become something of an analytic Festivus, and the quadrant system was again roundly abused in the more numerically-inclined corners of the internet.

The abrupt round-number-divisible-by-five cutoffs merit comment, naturally, but, as it happens, I’m done trafficking in this vein of critique. Approaching this as an analytic issue feels like a category error. The NET is fine analytically, and even the bolted-on quadrant chassis is mere superstructure. Conversely, the committee is base. The salient point about a committee selecting teams based on good analytics is that somehow there’s a committee on the loose selecting teams.

At their best, sports are an escape from committees. The NCAA tournament should bid adieu to its committee era and instead serve as the playoffs of college basketball.