Five Takeaways: Michigan at Purdue

Without the services of Caris LeVert, who missed his second straight game with a lower left leg injury because he was in pain during Thursday's shootaround, Michigan hung tight with #20/18 Purdue for the first 35 minutes before the Boilermakers steamed ahead at the end for an 87-70 victory in West Lafayette. Here are my five takeaways from Michigan's fourth loss of the season:

1. Purdue's inside-out offense shredded Michigan's defense.

UM Hoops' Dylan Burkhardt pointed out on Wednesday that no team in the nation runs more offense out of the post than Purdue, which was evident last night. Purdue made a concerted effort to get the ball down low to seven-foot A.J. Hammons, 7-foot-2 Isaac Haas, and 6-foot-9 Caleb Swanigan, knowing their size would overwhelm Michigan's four-guard lineup. When those big men had the ball in one-on-one situations on the block, they made quick moves for dunks and layups. Michigan knew this would happen, too, which is why its guards and wings often doubled down or scraped down to help.

However, that's when things fell apart for Michigan's defense. Purdue is not a selfish team. This season, the Boilermakers have assisted on 68 percent of their field goals, which is the eighth-best rate in the nation. They do an excellent job of moving the ball around to find the open shot, and that's what happened last night (86.7 ast%). When U-M doubled or scraped down, Purdue's bigs either hit slashers for shots at the rim, kicked the ball out to open shooters on the perimeter, or swung the ball as Michigan's defense was slow to rotate. In the first half, this didn't kill Michigan because Purdue was ice cold from the outside, missing all nine of its two-point jumpers and six of eight of its three-point jumpers. But, in the second half, Purdue stopped missing. The Boilermakers made 16-of-21 field goals (76.2 pct.) and 7-of-10 threes (70.0 pct.). Accordingly, they scored at a pace in the second period (1.68 PPP) that prevented Michigan the chance to complete a comeback. And, in the final 5:08, they outscored U-M, 21-10, to make it look like a rout.

So can Michigan fix this? It's unlikely in a man defense with this sort of personnel mismatch that puts that much stress on the defense. Michigan's centers couldn't handle Purdue's down low by themselves -- all four of U-M's center played and picked up a foul in the first six minutes -- so the Wolverines took their chance by doubling. That's understandable. The problem was that their perimeter defenders were lost off the ball and had sloppy rotations. That's something upon which Michigan must improve, but this wasn't the first time it's happened and how much improvement can be expected after the midway point of the season? This defensive issue should plague U-M much of this year.

2. Purdue's two-point defense was as advertised.

As MLive's Brendan Quinn noted, John Beilein had the perfect summation after the game for why Purdue pulled away from the Wolverines at the end of the second half:

"And they're better defensively than us -- there's no question about that. We were getting tough baskets, they were getting easy baskets. We answered with a tough, they were getting an easy. Pretty soon, you're not going to make the tough and they're going to make the easy."

Purdue is first in adjusted defensive efficiency (87.4), shooting defense (39.5 eFG%), and two-point defense (37.5 pct.), and, boy, did it show on Thursday. Not only do the Boilermakers boast a pair of seven-footers, one of which always is on the floor, to protect the rim, they have a handful of perimeter defenders that are pests, including reigning Big Ten Defensive Player of the Year Rapheal Davis. This usually forces offenses to settle for mid-range jumpers because layups get sent to the third row and threes are contested.

Though Michigan did what it needed to do from three, drilling 11-of-27 (40.7 pct.), U-M couldn't execute inside the arc. The Wolverines made only 12-of-34 twos (35.3 pct.) because they couldn't find clear lanes to the rim or feared testing A.J. Hammons, who had four blocks in 29 minutes. An example of this was Michigan's screen sets. One reason why Mark Donnal was so successful against Illinois and Penn State was he would roll to the basket after his screens. This gave defenses a decision: help in the lane or stick with Michigan's deadly shooters on the perimeter. Usually, they opted to stick with the shooters, and Donnal punished them for it. Last night, though, Michigan's bigs weren't rolling, particularly D.J. Wilson, who played most of the first half while Donnal sat with foul trouble. They were popping for threes and mid-range shots -- Donnal and Wilson fired 11 shots, seven of which were threes, and neither made a single two. This hurt Michigan's spacing as five Wolverines consistently hovered around the perimeter.

With things clogged and clustered, Michigan couldn't get easy two-point looks. Instead, the Wolverines were forced to put up difficult mid-range jumpers and off-balanced runners near the rim. Most Wolverines struggled with this, which is why Michigan faltered down the stretch at the end. One did not, however, and had a career game.

3. MAAR's the only one that solved the riddle that's Purdue's defense.

Muhammad-Ali Abdur-Rahkman, who filled in for Caris LeVert in the starting lineup for the second straight game, was the only reason why this game was competitive for the first 35 minutes. He shattered his career high with 25 points on 10-of-16 shooting and tallied four rebounds, two assists, and two steals. What was so impressive was his ability to score inside the three-point line and doing it in a variety of ways. He completed a eurostep with a reverse layup, he threw up high-arching floaters over the outstretched hands of Purdue's centers off the glass and into the net as he faded away, and he finished step-back, mid-range jumpers. Overall, he connected on 8-of-11 twos (72.7 pct.), while the rest of his teammates converted only 4-of-23 twos (17.4 pct.). That pretty much says it all.

LeVert is Michigan's best player, but Abdur-Rahkman has stepped up phenomenally in LeVert's stead. In the last two games, Abdur-Rahkman has averaged 19.5 PPG with a 73.9 eFG% and demonstrated that he and LeVert are the only two Wolverines that can penetrate and score off ISOs. Though he still has a tendency to have tunnel vision when he drives and not see open shooters outside, he's becoming more comfortable as a scorer.

4. Michigan needed to put more pressure on Purdue's ball-handlers.

There were many keys for Michigan in last night's game, but two were that Michigan needed to neutralize Purdue on the glass and harass the Boilermakers into turnovers. The first one is obvious as Purdue is one of the tallest teams in the country and a dominant rebounding outfit on both ends. The second one was revealed last weekend when Iowa used a trapping press to force Purdue into 10 second-half turnovers and overcome a 19-point deficit for an incredible comeback victory. As discussed in the first takeaway, Michigan faced a size mismatch down low on defense. If Michigan could win the turnover battle, it could hinder Purdue's offense enough for U-M to pull off the upset.

Though Michigan stunningly earned a stalemate on the boards -- Purdue's OR% was 24.0, while Michigan's was 23.1 -- the turnover battle was a stalemate, too. Purdue had nine turnovers to Michigan's eight, which essentially was a win for the Boilermakers because they are 165th in turnover rate and Michigan is 11th. It was surprising that John Beilein didn't opt to use more zone as Purdue frequently shredded Michigan's man defense. Though the zones may have permitted Purdue to be more active on the offensive glass, the 2-3 and particularly the 1-3-1 could have forced Purdue to throw more errant passes. Same time, Michigan's zones haven't been great. Purdue could have shredded them, too.

Either way, it's clear: Michigan didn't cause enough turnovers to win this game.

5. Michigan fans shouldn't get too down after this loss.

On social media after the final buzzer sounded, Michigan fans were pretty dejected about the loss. It's an understandable instant reaction because they just witnessed Purdue burn the Wolverines' defense again and again in the second half. Yes, Michigan has its problems and doesn't appear to be a Big Ten title contender -- at least not until Caris LeVert returns -- but it's important to remember that this wasn't a game that Michigan was supposed to win. Not only was this a road game against the nation's best defense without its star player, KenPom considered it the most difficult game remaining on Michigan's schedule. And U-M scrapped with Purdue until the final five minutes.

Also, the result doesn't change anything. Michigan still is in the top 30 on KenPom and projected to finish with a 21-10 (11-7 B1G) record. Assuming that LeVert's injury isn't more serious than to be believed, that would make Michigan an NCAA Tournament team. And this was only the first of a brutal three-game span against KenPom top-15 teams. If Michigan can snag one of the next two, that'd be a successful stretch for U-M.