12/6/2020 – Michigan 80, Central Florida 58 – 4-0

The first thing Terrance Williams did when he came in the game was run half a guy over and score. Michigan went on a 22-5 run immediately after his insertion.

Brandon Johns also came in at about the same time. Johns ended up a free throw away from an inverse trillion: he had something in every other column of a box score (1/1 from two, 1/2 from 3, an OREB, a DREB, two assists, a turnover, a block, a steal, and four fouls). At one point he dove in from the perimeter and ended up dumping a ball off to Hunter Dickinson from an angle that looked impossible until the moment it opened up.

use the force Johns [Campredon]

Then , they were gone. Johns got five minutes in the second half; Williams didn't get in until a minute before the Darko Milicic lineup hit the floor. Michigan didn't need to go back to them because Franz Wagner, who missed most of the first half with two early fouls, was swooping in for absurd reverse layups and Chaundee Brown was floating around the court with a Black Hole Sun smile on his face, pouring in long jumpers.

I love how happy he is on the court. It’s like he escaped from Shawshank

— Scratch BBQ and Catering (@ScratchBbq) December 6, 2020

Isaiah Livers didn't score until he got a putback with six minutes left. Michigan was up 26.

----------------------------------------------------

If you've been reading this site for at least a year you probably remember that Michigan melted without Isaiah Livers last year. We spent big chunks of the season slicing Bart Torvik's rankings into with and without Livers ribbons to demonstrate how indispensable he was to the operation. After he returned Michigan's three point luck went from horrendous to good and Michigan stormed through February, finishing as the #1 team in the country for the month.

This team is not going to melt if Livers isn't available, or is having a bad game.

This edition of Michigan has a lot of questions yet to be answered but we just got one about depth against a reasonably good team coming off a win against Auburn. Minus their established leader and projected breakout player, Michigan clawed back a 12-point deficit, and by the time that leader got on the board they'd turned a close game into a laugher.

There is a player that they'd melt without—Hunter Dickinson, about whom more in a minute—but last year it felt like virtually any player exiting for an extended period would result in a severely compromised team. This time around guys can emerge from the bench, thump dudes for a quarter of the game, and get a hat that says I PLAYED IN THE FIRST HALF upon return to the bench. That's going to be extremely valuable with a compressed schedule and COVID chaos on the horizon.

[After the JUMP: plus minus is useful at the extremes, and we've got extremes]

BULLETS

The eye test, aka feelingsball. We're going to have to ride with those two things if we're going to make sweeping pronouncements four games in. The numbers are currently being recalcitrant. I think that lineups featuring Austin Davis and Mike Smith are Michigan's weakest defensively. You probably do too. Hooplens:

Well… crap. That's actually a reasonable sample four games in and it doesn't look at all like you'd expect.

I'm still rolling with the eye test, which tells me that Davis can wall up a bit but is less mobile than Dickinson and that Dickinson ends up creating significantly tougher shots for the opposition. I'd imagine the differences get starker against higher-level opposition, which is going to be more capable of exploiting weak points than mid-majors.

There is one thing that I think is indicative.

Yes, this is a stat that is not in vogue but when the gaps are so big… Michigan's plus/minus says something about something. Seth collated plus/minus from statbroadcast and colored the boxes based on percentile on team:

We are most interested in the Davis/Dickinson gap, because it is huge.

+/- is not favored by folks looking for ways to summarize league- or conference-wide performance because lineup effects can lead to some nonsensical results. For example: Terrance Williams would have been +18 instead of +10 in the most recent game if he wasn't on the floor during Kenpom time.

Also a huge problem: noise. Pomeroy, a +/- skeptic, ran an experiment to demonstrate why he felt it was not a useful stat. He created a zero-impact player and simulated a bunch of seasons with a basic mathematical model:

Just for consistency, I decided to run my 20-game experiment 50 times. The results should give a more accurate assessment of the error associated with plus-minus. It wasn’t difficult to get a crazier result than the one described above. The third trial in this new run gave my player a -43 on the floor and a +52 off the floor. A few trials later, my guy was +48 on and -73 off. My dud had become a stud. I hope he doesn’t get hurt! For the 50 sets of 20 games, the average error in measuring my worthless player was 4.8 points per 40 minutes.

These are the reasons +/- is not in vogue.

And yet: here we have a straight one-for-one swap at center. Dickinson has a 19 point differential per 30, which is 25 per 40. That's five times the average error from the Kenpom thought experiment. This size of gap is extremely likely to be meaningful, and extremely likely to be a large effect.

Dickinson got 24 of the 36 minutes with rotation players on the floor; Davis got 12. Dickinson's currently getting 54% of Michigan's minutes. That should be more like 80%, fouls permitting.

Kenpom time buckets cause the bench to blow up [Campredon]

About Kenpom time. Since this is the first time this season I've invoked Kenpom time, a refresher: Kenpom time is the portion of a game after it is effectively decided. Lax defense, turnovers, and attempted 360 tomahawk dunks that scythe into the crowd and lodge themselves in a beehive hairdo are much more common during Kenpom time than competitive play. Also you will get deep bench lineups where walk-ons take spectacularly inadvisable shots with the bench going crazy on each attempt, let alone make.

Since AFAIK Kenpom does include garbage time—a search for its inclusion or omission doesn't turn much up other than Bart Torvik noting that his model does exclude it and that this is a difference between his and Pomeroy's—we are rooting for good results in garbage time because that way Michigan ranks higher and we feel better about Kenpom's season projections.

If you're shaking your head about this please remember it is a bit.

NC State in jeopardy. Michigan's lone high-major nonconference opponent just cancelled a game against UConn that was scheduled for the Mohegan Sun bubble. Wednesday's game is in doubt:

A lot of uncertainty surrounding Wednesday's game between Michigan and NC State. NC State has COVID issues and, I'm told, it'll be their decision whether to cancel or play. Appears to be in jeopardy but we'll see what the next 24 hours bring.

— Josh Henschke (@JoshHenschke) December 7, 2020

There is chatter about the NC State positive being a false one and a PCR test coming back negative but if that was the case I'd think the game would be already announced as on. Quinn:

Noteworthy that the language on Michigan-NC State will be "postponement," not cancellation. So might be rescheduled.

I'm told Michigan will try to fill the schedule with a game this week. Could be an in-state opponent or Toledo. Game could be Wednesday, but more likely Thursday.

— Brendan Quinn (@BFQuinn) December 7, 2020

Michigan has a 12-day gap from the 13th to the 25th but NC State has games scheduled for the 12th, 16th, 19th, and 22nd. Louisville is on the 16th and is also on pause. If that gets postponed or canceled that might be a makeup date. Otherwise it's impossible to find one. Also, if Michigan does schedule a replacement game that would hit their 25-game limit (no tourney, and insanely the tourney exemption still exists this year) and necessarily cancel NC State unless they got a waiver.

Michigan losing their sole high-major measuring stick would not be good, obviously. The Big Ten has done pretty well in the nonconference and I imagine the committee's tendency to stick it to teams with bad nonconference SOS rankings will be suspended this year. Even so the prospect of entering Big Ten with no nonconference buffer is slightly ominous.

Repeat after me: May Madness. Allow post-conference window for various bracket buster games and move the tourney back to a point where vaccines give it a better chance of successfully concluding.

[UPDATE: NC State is officially postponed. Matt Norlander also reports that Michigan was trying to line up a game with UConn and then UConn had to pause.]

This did not go down [Campredon]

OMFG. I had the same reaction Zeb Jackson did when his change of direction and explosion got him to the bucket in Kenpom time and then the ball touched every part of the rim before escaping. Jackson made a couple bad decisions and looks like a guy who's going to need a year to get situated but you can see the athletic potential every time he attacks the basket.

Midrange is grudgingly approved. I'm not sure about Chaundee Brown taking midrange jumpers, as he did twice in this game, just yet. For guys like Mike Smith, Eli Brooks, and Isaiah Livers I withdraw my usual complaints about two-point jumpers. Smith is 7/10 on other twos this year; Livers is 6/10; Brooks is a little behind but has had success in the past. These numbers include floaters and the like, which Brooks has historically been very bad at.

Pulling up from 15 feet is a more worthwhile activity than it has been in the past for various reasons.

One: Michigan is crashing the boards far more than they did in the Beilein era or even last year. Their OREB% has shot up from the 280s the last two years to the 50s. While opponents have an impact on that, Michigan sent two or three guys to the offensive boards for much of this game against an uncommonly long and athletic mid-major. That would not have happened under Beilein. When you have a better chance of grabbing a rebound the value of a jumper goes up relative to further offense that may result in a turnover.

Two: Michigan does not have a ball-dominant point guard who's going to finish top five in assist rate, and turnovers may be more of an issue. The value of a jumper goes up when you don't have a default decent late-clock shot generator and may turn the ball over more frequently. (I say "may" because Michigan started last year with a TO rate that ranked 150+ last year and ended up 11th.)

Three: they've got some guys who have not been very good at drawing fouls or finishing when they get contested looks at the rim. 

I'm not in favor of the sophomore Tim Hardaway Jr long two with 25 seconds on the shot clock, but if Michigan's facing drop coverage and they've only got time for one more action a jumper is just fine.

[Campredon]

Dickinson's stats are absurd. To date:

• 15% OREBs, 22% DREBS
• 6.7 fouls drawn per 40 and a FT rate of 67; he's hitting 70% from the line.
• 75% from two
• 24% usage
• a 10.6 TO rate, and a 12.6 assist rate

That is absolutely bonkers, particularly the part where a freshman big has an assist rate higher than his TO rate. Also the OREB rate. Also hitting 75% from two. Sledding is about to get a lot tougher but I don't think he's slowing down enough to not be one of the better centers in the Big Ten.

One more thing COVID has taken from us. Juwan Howard ref face isn't the same when you can't see the entire pained expression.

[Campredon]

Next year we will have a picture of Howard going full disappointed dad on someone after a block/charge call and it will feel so good.