statistics are sometimes bad

A picture of the conference. Michigan's defense isn't that much of an issue so far:

It's pretty good, and then the offense is off the charts. It's only in the context of the super-elite teams vying for a national title that it seems deficient. And with that offense… well… Gasaway's latest Tuesday Truths puts it in perspective:

It may turn out to be the case that Michigan is not in fact excellent at defense, that they're merely very good at it. But that needs to be seen in the proper context. First, this isn't a case like, say, Missouri last season, where a good many people chose to overlook the Tigers' vulnerability on D. (There was a push to give that team a No. 1 seed. I still shudder at the memory.) John Beilein's defense this season is day-and-night better, thus far, than Frank Haith's was last season.

Second, whatever Michigan's level of performance has been on defense, the Wolverines have been able to plug that in as one half of an equation whose result has been outscoring the best conference in the country by nearly a quarter of a point per possession. The Wolverines' only loss this season has come not to an offensive juggernaut that was able to exploit UM's worrisome deficiencies on defense, but to the hapless-on-offense Ohio State Buckeyes, who shut down Michigan's offense beautifully.

Lastly, the past 10 years can be ransacked profitably not only for prerequisites (and I'll be joining Luke on this beat soon -- watch for it!) but also for weirdness. I've seen a team rank No. 8 in its 12-team league in two-point accuracy and then go on to win a national championship. I've seen a team rank No. 103 in the nation in offense and then go on to make the Final Four. And do I even need to drag Gordon Hayward into this?

The most likely outcome of March is that Michigan will indeed lose to some other team in the tournament, because they are only amongst a leading group of teams. If and when that happens, people will point to the defense; I'll just be like "Michigan was the Vegas favorite and still 5 to 1 against."

Slightly more favored in the league. Michigan's huge scoring margin in the league sees them favored to win the Big Ten in SpartanDan's Bradley-Terry projection system*, which may not be a huge surprise. What is surprising is how much they're favored by.

Dan's basic system that does not take margin of victory into account says Michigan has a 69% chance of an outright title and an 85% chance of sharing. The margin-aware numbers are 80%(!!!) and 92%(!!!).

Those numbers are probably too high since Michigan is likely to have outperformed its real level of skill significantly in the opening third of the conference schedule, but… wow.

BONUS: Penn State has a 30% shot at going winless in the margin-aware system.

*[College hockey fans: this is KRACH.]

Bullet of stats-enthusiasm-dissing hypocrisy incoming. While I'm generally a fan of Big Ten Geeks, their latest foray into stat assemblage is goofy to me. They use "stops," which is a Dean Oliver formula that crams steals and blocks and rebounding statistics into a number. As with all attempts to create a catch-all defensive statistic, it waves its hand at who is in fact responsible for team defensive rebounding and how replaceable they may or may not be. Also unaccounted for is a player's contribution to the opponent's shot quality.

But they've compiled the numbers and shown you the results:

Let’s look at Stops:

Player	Stops per 40 minutes
Adam Woodbury	11.35
Mitch McGary	10.94
Jordan Morgan	10.47
Branden Dawson	10.21
Trevor Mbakwe	10.20
Cody Zeller	10.20
Ryan Evans	9.45

Well, this is interesting—we have a couple of freshmen leading the way. Both Woodbury and McGary are tremendous rebounders (as is Jordan Morgan this season), which explains why they rate so high. And to those who complain that Stops unfairly rewards good rebounders, I think that’s about as valid a point as the complaint that offensive rating unfairly rewards efficient scorers. Rebounding is defense—a big part of it.

So this works if rebounding is, in fact, defense. It's not. It has an impact but the top ten teams in defensive rebounding are 54th, 144th, 162nd, 147th, 103rd, 171st, 240th, 64th, 18th, and 25th in defensive efficiency. As I mentioned when pooh-poohing Mason Plumlee's KPOY candidacy, rebounding is the least important of the four factors. It's only its trackability that makes it so prominent. It's easy to say who got a rebound. It's really hard to credit someone for an effective rotation.

This metric thinks Jordan Morgan is a lot better this year because the team is better at rebounding. His personal DREB rate is a tick better this year, but it's still just 257th. He gets credit that other players don't because Tim Hardaway is mansome this year.

Morgan is then declared the best defensive player in the league because he fouls less often than the other guys at the top of the list, with this capper:

I can’t believe I’m saying this, but Jordan Morgan has been the best defensive player in the Big Ten so far. Unbelievable. And in case you’re wondering, Oladipo fouls quite a bit—4.42 fouls per 40 minutes. Talk about the eye test all you want, but the numbers suggest he’s not the best defensive player (or even the best defensive perimeter player) in the Big Ten.

I accuse Big Ten Geeks of gross misuse of stats. Yes, it is unbelievable. Thus you should disbelieve it.

Victor Oladipo isn't a high-end defensive rebounder because he's frequently sticking his hand in the face of the highest-volume shooter the opposition has. His role defines his numbers. You can cram defensive rebounds into a slightly modified form all you want—notice that not one perimeter defender appears on this stops list—but all you get is a comparison between yourself and David Berri. Deployed.

Sometimes you have to go by the eye test because the stats compiled are inadequate, and until basketball stats get crazy detailed individual defensive performances are in that bucket.

BONUS WONKY STATS COMPLAINT. Most attempts to compile defensive numbers underrate the value of a steal, by the way. A defensive rebound is just the successful conclusion of a defensive possession ending in a missed shot. A steal ends a possession by itself—it's the miss and the rebound rolled into one—and frequently leads to a transition opportunity at the other end. That latter part is not well accounted for.

Morgan's ankle. Nothing broken, just a sprain, AP got a totally gross picture of it, if he can play basketball on Wednesday he will play basketball on Wednesday—I bet he cannot play basketball on Wednesday.

Zak Irvin continues rain of destruction. Last week: 26 points and 30 points in wins. One was over Arsenal Tech, both the best-named and top-ranked team in the state until Zak Irvin declared his school was now named Sharkfin Elfin 3000 and scored almost half of his team's points in a 64-59 win.

You want to watch the whole game, you say? You have free time.

If you are going to do this you probably want to start at halftime. Irvin scored 26 of his 30 after the break.

Zing. John Niyo on the Nobody Remembers #1 thing:

"It's Jan. 27," Beilein said after a 74-60 victory at Illinois last weekend, "and not one of you can remember who was No. 1 last Jan. 27."

Well, actually many of us can. It was a 20-1 Kentucky team that went on to win the Southeastern Conference and the SEC tournament and eventually the NCAA title.

But point taken.

5 to 1 against, 5 to 1 against, 5 to 1 against, repeat until you internalize the likely outcome of the season is not cutting down nets…

Etc.: You can be happy about being #1. Via UMHoops, the view on Bielfeldt from Peoria. Being back on top is nice and you should be happy. Here's an excellent primer on Beilein's 1-3-1 from the man himself.