less dunking on the MEAC, but still some [Marc-Gregor Campredon]

20 Game Conference Schedules Are Good Comment Count

Brian December 4th, 2018 at 12:44 PM

The most recent Ken Pomeroy podcast had a brief discussion of the Big Ten's 20-game conference schedule, which got a thumbs down because Pomeroy prefers nonconference games. Nonconference games connect various conferences and are required for ranking systems to make sense, so Pomeroy's got a point.

But what kind of games are being excised by the expanded conference schedule? I looked at everyone's schedules this year and last and divided them into approximate major and non-major categories. There's some wobble in these distinctions. The A-10, Mountain West, and Big East count. Certain programs (Gonzaga and the top of the American) outside of the top 7 conferences also count. I ended up grudgingly including DePaul because they're in a major conference but left out some incidentally top 100 opponents like Montana and South Dakota State on the assumption that these were buy games that were accidentally good opponents. FWIW, if you were to do it the other way and drop out programs like Pitt while including good mid- and low-majors the number of games worth playing would remain essentially equal.

Anyway, the conclusion is that the extra conferences games have almost universally replaced bad buy games:

  2018 2019
  Major-ish Mid-low Major-ish Mid-low
Michigan 5 (LSU, VCU, UNC, UCLA, Texas) 8 5 (Nova, GW, Providence, UNC, SoCar) 6
MSU 5 (Duke, UNC, ND, DePaul, UConn) 8 5 (KU, UCLA, Texas, Florida, Louisville 6
Wisconsin 6 (Xavier, BU, UCLA, UVA, Temple, Marq) 7 6 (Xavier, Stanford, OK, UVA, NCST, Marq) 5
Nebraska 4 (St John's, BC, Creighton, Kansas) 9 5 (Seton Hall, TTech, Clemson, Creighton, OkieSt) 6
OSU 5(Gonzaga, Stan, Butler, Clemson, UNC) 8 4 (Cinci, Creighton, Cuse, UCLA) 7
Indiana 4 (Seton Hall, Duke, UL, ND) 8 5(Marq, Ark, Duke, UL, Butler) 6
Maryland 4(Butler, Bonnies, UNM, Cuse) 9 2(UVA, Seton Hall) 9
Purdue 5(Marq, Tenn, AZ, UL, Butler) 8 5 (Davidson, FSU, VT, Texas, ND) 6
PSU 4(NCST, GW, Pitt, A&M) 9 4(DePaul, VT, NCST, Bama) 7
Northwestern 6(Creighton, La Salle, TTech, GT, DePaul, OK) 7 6(Fresno, La Salle, UT, GT, DePaul, OK) 5
Minnesota 4(Prov, Bama, Miami, Ark) 9 5(Utah, A&M, UW, BC, OkieSt) 6
Illinois 5( DePaul, Wake, UNLV, NMSU, Mizzou) 8 7(GTown, Zags, ISU, Xavier, ND, UNLV, Mizzou) 4
Rutgers 2 (FSU, Seton Hall) 11 3(St John's, Miami, Seton Hall) 7
Iowa 3(VT, ISU, CU) 9 4(Oregon, UConn, Pitt, ISU) 7
    118   87

There are 28 fewer slots for mid- to low-major buy games and 31 fewer mid-to-low-major matchups. Only two teams (OSU and Maryland) are playing fewer major opponents this year.

There's an argument that the increased slate of conference games reduces opportunities for teams like Bucknell, which was three points away from a win at Maryland last year, to establish their tournament bonafides. That may be happening to some degree but teams like Bucknell, Marshall, Belmont, Wright State, Loyola-Chicago, and Bradley still speckle Big Ten schedules.

It's undeniable that most of the games that aren't being played as a result of the 20-game schedule aren't really worth playing.



December 4th, 2018 at 12:49 PM ^

I think it just puts a lot of importance being in a quality early season tournament. Maryland was in a shitty one this year so they didn't get a good slate of teams. 


December 4th, 2018 at 12:49 PM ^

I was listening to the podcast last night and had the same thought. My thought was to look at non-con SOS this year compared to previous years for B1G teams and see how the numbers look. Are they stagnant, lower or higher? If teams have much harder non-con SOS's, then it looks like we're just cutting out the body bag games and vice-versa. This is a good way of proving that point now.

Side note: I really don't like Pomeroy's co-host for the pod. He doesn't provide anything of substance and is just a yes man who agrees with whatever Pomeroy says.

Arb lover

December 4th, 2018 at 12:57 PM ^

So this anti-sec model of more conference games and fewer cupcakes will cause conference opponents to have more end of season losses. Does that hurt our record for tournament seeds i.e quad 3 v 2, or can/will rankings like kenpom account for this?

While wins against worthless teams are seen as just that, if Iowa/Wisconsin  have an extra 4 losses from the increased B1G schedule, does that bump them out of a quality quadrant win for Michigan?


December 4th, 2018 at 1:14 PM ^

I'm always in favor of more accurate rankings within a conference, which I think is accomplished by a more conference heavy schedule.

To me it seems that more conference play will provide for a better conference tournament both in terms of team preparation and proper seeding, which in term is better preparation for the national tournament.


December 4th, 2018 at 1:28 PM ^

While it's good that strength of schedule isn't being hurt, a team's record will be hurt. Those two games were usually two wins for almost every Big Ten team. Now, there's 14 more losses to be spread across the conference because there is a loser in each conference game.

Now, hopefully, the analytics give the Big Ten teams the benefit of the doubt because of the strength of schedule, but losses are never good in analytics either.


December 4th, 2018 at 1:52 PM ^

Not necessarily the case.  True, NET is sort of a fancified RPI that is based on W/L, but stuff like KenPom actually doesn't care an iota about the outcome of a game.  Only performance relative to the opponent.  Losing good is better than winning bad.  Reason being that NET is trying to quantify a team's results, while KenPom is trying to be predictive.

NET and RPI have a reward component for beating Coppin State by 1 and a ding component for losing to Duke by 1.  Other analytics, KenPom specifically, would ding you for the former and reward you for the latter.  Simplistically speaking.


December 4th, 2018 at 2:29 PM ^

It also makes the conference race a bit more equitable.  Now we play seven teams twice and six once, which is still far from perfect but an improvement over the ridiculous situation the last few years (when we had five double-plays and eight single-plays), leading to situations like MSU's last year where they beat one tournament team in league play and still won the title.

A 22-game schedule (nine double-plays and four single-plays) would be better still.

El Jeffe

December 4th, 2018 at 8:07 PM ^

I'd kind of like to see the "Mid-low" category split into "Mid" and "low." I recall much pearl clutching in some quarters (including on Brian's own fainting couch) about how UM had all these boat anchor games and smart guys like Richard Pitino were gaming the system to get higher than deserved bids.

So was that not true in 2018, or is it that among UM's 8 "Mid-low" games there were a larger proportion of said boat anchors* than among Minny's 9 "Mid-low" games?

* Why do we use this term? Is there another kind of anchor?