The Bracket Split

Submitted by Brian on February 16th, 2018 at 11:55 AM

38463333850_f309b0fbe5_z

[Marc-Gregor Campredon]

So here's a thing: there is a big ol' discrepancy between various bracketology attempts when it comes to Michigan. I don't think I've seen such wide splits during this pleasant recent era when the tourney is a given unless the basketball gods hew down your two best players. And I'm not talking about randoms on Bracket Matrix:

I think the main discrepancy here is between people looking at overall strength of record calculations and those looking at the NCAA's teamsheets. Michigan has a Quadrant One problem. Take it from Palm:

The Wolverines have a poor non-conference schedule and only two quadrant 1 wins, but neither of those things can be addressed tonight.

It doesn't take long to figure out why the first part of that might be the case:

image

Michigan's quadrant one games are

  • @ Kenpom #4 Purdue
  • @ #6 MSU
  • @ #10 UNC
  • @ #11 OSU
  • @ #52 Nebraska
  • @ #45 Texas
  • home vs #4 Purdue

IE, five games against top 10-ish competition, four of them on the road, and just two in the more winnable section of the bin. So if you're looking solely at the bin that doesn't look too great. It looks significantly better if you take each individual game and pile them into a Strength Of Record calculation, as ESPN does. ESPN ranks Michigan 17th in SOR, i.e. the first five seed. The AP poll, which is sort of a folk SOR calculation, is a seed more skeptical but more or less agrees.

So if you're looking at the worse and more arbitrary bins and eyeballing it, if you're looking at basic nonconference SOS measures that don't know that playing 200 is exactly like playing 350 if you're tourney-level, Michigan doesn't look too good. If you're trying to sum up Michigan's season without the bins, they look a bit better. Thus a couple of votes for 5-6 seeds and a general 8-9 malaise.

The good news is that Michigan has three relatively easy Q1 games to finish the year: home against OSU and on the road against Penn State and Maryland. Those are 58%, 38%, and 41% shots per Kenpom, much better than the large majority of their Q1 games to date. (They had a 15% shot at MSU, a 10% shot at Purdue, and a 20% shot at UNC, for instance.) If they could pick two of those off and end up 4-6 against Q1 teams, hopefully their high degree of difficulty in those games—7 of which would be on the road against top 50 teams—would matter to someone and they could get into that 5-6 range the optimistic folks currently have them in.

FWIW, Alex played around with Bart Torvik's Teamcast a bit. Findings:

  • Winning out (ie: also winning the BTT) gets Michigan a three seed(!).
  • Losing out puts them in Dayton as the last team in the field.
  • A reasonably optimistic end of season featuring two wins in Michigan's final three games and a 2-1 record in the BTT nets them that 6.
  • Going 1-2 down the stretch with reasonably optimistic BTT gets them a 9 seed.

So there is a great deal left to play for here.

Comments

Dodort

February 16th, 2018 at 12:11 PM ^

This reminds me of their resume in 2016, where if you just looked at the RPI buckets, Michigan only had 4 top 100 wins, but when you looked closer they were all top 25 (or close to it).  It all depends on how lazy the committee is.

TrueBlue2003

February 16th, 2018 at 5:46 PM ^

our other metrics were terrible. We were 55th in kenpom entering the tourney. And that was down significantly from the top 30ish team were were before Caris was hurt.

Given our 3-6 finish to the regular season and the fact that the committee does (supposedly) factor injuries into the equation, we were pretty lucky to make the tournament in 2016.

You're correct that a large proportion of our "top 50" (which is what was used at the time) gamees were top 25, but we lost a lot of those at home, which changes the difficulty sustantially.  Those top tier games weren't as difficult as our top tier games this year.

ijohnb

February 16th, 2018 at 12:17 PM ^

team is not on the bubble.  If by on the bubble he means "could be in trouble if they lose the rest of their games," OK, but the same can be said for 80% of the eventual field.

MNWolverine2

February 16th, 2018 at 12:20 PM ^

I think the difference is we have 3 very losable games down the stretch.  Look at what happened to OSU @ PSU (PSU only 1 spot behind us in KenPom) and MSU probably should have lost @ Maryland.

OSU is almost a must win, otherwise we might have to sweat through another game against Illinois or Iowa in the Big Ten Tourney.

wayneandgarth

February 16th, 2018 at 12:54 PM ^

I very much agree.  We are starting to play better on the backs of the vetererns (Mo, MAAR, Robinson), but the three game stretch is tough.  If we were to go 2-1 it would be a hell of a shot in the arm.  But the key is avoid 0-3 and the key to that is win at home vs. OSU. 

The biggest problem we have is point guard and that is not good come tournament time. 

taistreetsmyhero

February 16th, 2018 at 12:31 PM ^

To me, the bubble is for teams that may or may not be in if they only win the games the easy games remaining. Then, the big games they have left either put them on the right side or burst their bubble. It’s stupid to include teams that would have to collapse to be iffy in that discussion.

TrueBlue2003

February 16th, 2018 at 6:06 PM ^

though so these bubble watch's always mention anyone that couldn't literally lose all remaining games and still make it.  The more teams you include, the more fanbases click and share.

As winnable as our remaining three games, they're all very losable.  It's not totally crazy to think we could lose out and be facing a must-win BTT game against a team like Minnesota or Illinois, both of which we had some struggles with at home.

kjhager444

February 16th, 2018 at 12:19 PM ^

Man, Maui ended up being an underwhelming field, especially with the injury to Colson.  Granted a path with a win over LSU would have been either Q2, Q2 (ND) Q1 (WSU) or Q2, Q2 (ND), Q2 (Marquette) which would have been better than what Michigan ended up with (Q2, D2, Q3) but considering what that looked like at the beginning of the season, it's reaaaaallly hurt the non-conference schedule.

ak47

February 16th, 2018 at 12:32 PM ^

The guys with Michigan as a higher seed are also the guys with bad track records. Lunardi's bracket is worthless until he gets inside info the day before, Jerry Palm is 82nd on the bracket matrix leaderboards and Howie Schwab isn't even ranked. Most of the guys who get it right have Michigan around 8/9.

ak47

February 16th, 2018 at 1:46 PM ^

That's fair, probably rude to throw him in with Palm and Lunardi at this point. But overall I think the trend is the numbers people who have been more accurate the last few years see Michigan as an 8-9, and to be honest I agree. Sure it sucks a lot of Michigan's Q1 games were really difficult but at the end of the day they only won one (obviously can change over the next couple of weeks).

Against Michigan's schedule most decent teams would have about the same record. Michigan's performance in those games might suggest a higher ceiling but the results matter more and the results look like the resume of an 8/9 seed a 1 seed probably isn't thrilled about.

TrueBlue2003

February 16th, 2018 at 6:20 PM ^

strongly correlated with the Strength of Record metric which is probably how Shwab is doing his rankings.

Based on the first iteration of the top 16 though, it looked to be heavily based on Q1/Q2 wins/performance.

I have little faith in the committee recognizing that our Q1 games are really, really difficult Q1 games overall and instead think they'll simply say, eh, only 7 Q1+Q2 wins. You get a 9 seed.

bronxblue

February 16th, 2018 at 12:40 PM ^

This has always felt like an 8/9 team to me, mostly because they only have one "signature" win (at MSU), and at best will have one more against OSU before the BTT.  But the tournament committee, despite their many failings, aren't completely obtuse.  Michigan looks like one of the top 38-ish at-large teams in the country throughout the year, and I don't see that changing between now and selection sunday.  Now, if they got blown out by Maryland, PSU, and OSU then let's talk, but other than PSU being frisky nothing coming up makes me think that'll happen.

I also think the realistic ceiling for this team in terms of seeding is a 7 seed, and honestly all the likely #1 seeds (Nova, Duke, UVa, Cincy, MSU, etc.) don't look demonstrably harder than a #2 you'd run into.  

Indy Pete - Go Blue

February 16th, 2018 at 1:10 PM ^

Good stuff Brian.  NCAA tournament seeding is somehow even more annoying than the college football player committee. Every year, they choose a new thing to emphasize, and it is always a moving target. They set up these arbitrary categories that are inflexible and (for example) create a gigantic distinction between the 75th ranked team and the 76th ranked team. There is heavy reliance on an antiquated measure called RPI that means nothing to anybody, except when it comes to the tournament where it means nearly everything.

The thing that kills me about the college football playoff committee: it all comes down to your record against THEIR top 25 teams.  Since they are the ones choosing the rght teams, they have tremendous power to influence things by putting a borderline rank team into or out of their top 25 per their agenda.  If you beat #26, #27, and #28, that is nearly the same as beating the 3 worste teams and FBS - they don't make a distinction beyond Top 25 and they do not mention your wins over top 30 or top 40 teams.

Bottom line this year, though: we are a tournament team. And actually, we have a favorable end of the schedule where we will have our fair opportunity to vault up the seed line. I'll take it!

TrueBlue2003

February 16th, 2018 at 6:30 PM ^

teams with winning records this year, which expanded it outside of the top 25 (but created an even worse problem in which it favored conferences that only play 8 games).

In any case, the CFP has made three controversial decisions in four years: OSU over a big 12 team in 2014, Wash/OSU over PSU in 2017 and Bama over OSU in 2018.

In each scenario they took the "best" teams over the "most deserving" (although even that was debatable in 2017).  They were vindicated in 2014 and 2018 by the results. Not so much in 2017.

I actually don't like that they did it that way because actually playing good teams and winning those games becomes secondary to "have good players and win by a lot of whomever" such that teams are disincentized from playing tough games.

At least in bball, if they get it wrong, it's that they care too much about having played a good schedule and having won tough games.

Boner Stabone

February 16th, 2018 at 12:48 PM ^

Boner's Bracketology has Michigan as a #9 seed and MSU and OSU as a #4 seed.  According to my formula my top 4 seeds are:  Virginia, Villanova, Xavier, and Auburn

My last 4 in are:  UCLA, Temple, Texas, and Syracuse

My First 4 out are:  Virginia Tech, St. Bonaventure, Kansas State, and Boise State

Monocle Smile

February 16th, 2018 at 1:59 PM ^

That's some seriously selective memory. I was in school from 2007-2012. This board has overreacted like crazy to a disappointing 8-5 on the back of two 10-win seasons.

Take a break from mgoblog during the offseason. Front page content is still good, but less frequent, and and board becomes a plague-infested sewer. Maybe pick up a hobby or something. Watch the World Cup.

iloveyellow

February 17th, 2018 at 12:16 PM ^

Another issue is that the Big Ten isn’t actually any worse than it’s been the past 4 years—this year is actually stronger top to bottom than 2015 and 2016 (KenPom). It just seems like it’s bad because a. The SEC is strong and b. A lot of the B1G’s good teams are getting single plays against the other good teams.

Just another example of how from one angle things look good and from another, they look bad.

Z_Wolverista

February 17th, 2018 at 2:59 PM ^

It seems tomorrow's match, while characterized as a "relatively easy" Q1 game, is pretty critical, both wrt bracketology & perhaps just as importantly, in terms of momentum-- psychologically.

Can anyone sketch out an updated preview?

Want to be prepared. Would super appreciate it! Thx all.