Examining lost production and projecting replacements

Dylan Burkhardt

Michigan lost three starters from last year’s roster and will have to replace 56.3% of last year’s minutes and 57.9% of last year’s scoring.

It’s no secret how important Derrick Walton, DJ Wilson and Zak Irvin were to the Wolverine rotation last season, but it isn’t always as clear what part of the offense will be most affected.

Using play type statistics from Synergy Sports, we can see what types of plays Michigan used the most last season, what portion of that production was lost, and who on the roster is best equipped to fill the void in each area of the game offensively.

Spot Up

% of Poss. Lost: 53%
% of Offense: 27.1%
Top Candidate: Muhammad-Ali Abdur-Rahkman (112 poss, 1.17 PPP, 90th percentile)

I expected to write up Duncan Robinson here, but Abdur-Rahkman is not only Michigan’s highest usage spot-up player returning, he was also more efficient than Robinson. Abdur-Rahkman scored 1.17 points per spot-up possession, using 112 possessions. Only Derrick Walton (1.5 PPP) was more efficient spotting up while Irvin (141 poss) and Wilson (119 poss) used more. Duncan Robinson wasn’t far behind at 92 possessions (and 1.13 points per), but Abdur-Rahkman’s efficiency surprised me.

Lost in the shuffle here is the fact that Abdur-Rahkman is actually a very good shooter off the catch. He had a 60 eFG% on catch-and-shoot jumpers last season and he’s more dynamic in spot-up situations because he has a nice array of moves to put the ball on the floor and attack closeouts. Right now a move like this isn’t in Duncan Robinson’s bag of tricks.

One other name to keep an eye on here is freshman wing guard Jordan Poole. If Poole is going to play a Zak Irvin-as-a-freshman type of role this year, then he could easily use 100 spot-up possessions off the bench. Irvin used 103 as a freshman with a 61.2 eFG% and Poole certainly has the range and a smooth stroke.

Pick & Roll Ball Handler

% of Poss. Lost: 80%
% of Offense: 15.1%
Top Candidate: Jaaron Simmons (228 possessions, .82 PPP, 65th percentile @ Ohio University)

This is the one that probably keeps John Beilein up at night. It’s also the reason that he went out and secured Jaaron Simmons as a graduate transfer from Ohio. Michigan lost 322 of 402 ball screen shooting possessions from a season ago — a staggering number when you account for the fact that the pick and roll is the most important element of the Wolverine offense.

Derrick Walton graded out in the 86th percentile as a ball screen scorer and, despite his ups and downs, Zak Irvin graded in the 63rd percentile. Muhammad-Ali Abdur-Rahkman used 54 ball screen possessions, but scored in just the 31st percentile nationally.

It’s tough to adjust for the upgrade in competition, but Simmons appears to be the perfect answer. He shows the ability to score in the ball screen game from every level and has a significant sample size (228 possessions). Here’s a quick reel of Simmons’ ability to score in the pick-and-roll game.


% of Poss. Lost: 59%
% of Offense: 13.4%
Top Candidate: Charles Matthews (25 possessions, .64 PPP, 5th percentile)

This is where Charles Matthews should make an impact. Matthews is an elite athlete who can finish above the rim in transition. The problem with that assumption is that Matthews’ offensive efficiency numbers were bad in almost every area other than putbacks during his one season at Kentucky.

If you believe in Michigan this year, then you have to believe that Matthews will be a far better offensive player than he showed at any time in Lexington. That doesn’t take a big leap of faith, but the question is just how much better? At a minimum, Matthews should have his fair share of transition dunks this season.


% of Poss. Lost: 63%
% of Offense: 8.7%
Top Candidate: Jaaron Simmons

Michigan is generally heavier on pick-and-roll action than isolation sets, but the Wolverines graded out in the 87th percentile in ISO offense last season and Derrick Walton and Zak Irvin led the way.

Muhammad-Ali Abdur-Rahkman is a great isolation scorer in his own right, scoring 40 points in 39 possessions last year (87th percentile), and Moritz Wagner was also effective (28 points on 30 possessions) — make sure to check out our look at his isolation offense in Crete — but the sheer volume of isolation offense that Jaaron Simmons ran last season is impressive.

Simmons scored 126 points on 132 isolation possessions last season and only 6 Division I players scored more points in ISO sets last season. This isolation ability will be critically important because it will give Michigan an option to attack switches in the ball screen game.


% of Poss. Lost: 57%
% of Offense: 6.4%
Top Candidate: Charles Matthews

DJ Wilson scored a third of Michigan’s points off of cuts last season. He gave the Wolverines the production as a baseline finisher and backdoor cutter that they hadn’t had since Glenn Robinson III was on the roster.

It’s hard to be an inefficient cutter (Mark Donnal was the worst on the roster last year, but still managed 1.25 points per cutting possession), but great cutters create opportunities and finish plays that others wouldn’t.

Moritz Wagner had 37 cutting possessions to Wilson’s 55 and I’d expect to see that number increase, but from the wing Charles Matthews seems like another worthy candidate with his length, athleticism and finishing ability.

Off Screens

% of Poss. Lost: 41%
% of Offense: 6.2%
Top Candidate: Duncan Robinson (55 possessions, 1.11 PPP, 80th percentile)

This is one of the play types where Michigan returns its top option from a season ago. Duncan Robinson had a 59.2 eFG% coming off of screens and used 55 possessions. However, shots off of screens accounted for only 6.2% of Michigan’s offense. If Robinson plays more minutes this year — a safe assumption with the graduation of Zak Irvin and early departure by DJ Wilson — then it is easy to imagine John Beilein trying to generate some extra offense by running Robinson off of more sets like this.

Pick & Roll Man

% of Poss. Lost: 39%
% of Offense: 5.8%
Top Candidate: Moritz Wagner (89 possessions, 1.34 PPP, 90th percentile)

The story here isn’t losing the primary option, it is losing a damn good number two. Only two players in the country scored more points rolling (or popping) in the ball screen game than Moritz Wagner’s 119 — St. Mary’s Jock Landale and Virginia Tech’s Zach LeDay. Wagner is an elite talent in the ball screen game and Michigan should be able to get him even more touches in 2017-18.

What the Wolverines need to find out is whether Jon Teske or Austin Davis are capable of catching a ball screen pass and finishing appropriately. It’s unlikely that either is a pick-and-pop threat — an option that Michigan had from both Mark Donnal and DJ Wilson at the five — but they’ll have to prove that they can finish consistently to make up for the loss of Donnal and Wilson.

Put Backs

% of Poss. Lost: 60%
% of Offense: 4.1%
Top Candidate: Moritz Wagner (23 possessions, 1.13 PPP, 56th percentile)

Michigan has ranked better than 275th nationally in offensive rebounding rate just three times under John Beilein and two of those rosters had Mitch McGary. The Wolverines lose their best put-back man in DJ Wilson, but if it comes down to winning and losing games with offensive rebounding than this roster is going to be in trouble.


% of Poss. Lost: 33%
% of Offense: 3.9%
Top Candidate: Moritz Wagner (63 possessions, .952 PPP, 81st percentile)

Post-up offense accounted for just 3.9% of Michigan’s offense and Moritz Wagner was responsible for 61.2% of those possessions. Wagner is very skilled on the low block — his .952 points per possession rated in the 81st percentile — and I’d expect Michigan to get him more touches on the low block, especially if he can continue to improve as a passer.

Hand Off

% of Poss. Lost: 71%
% of Offense: 3%
Top Candidate: Charles Matthews

The dribble hand off in Michigan’s offense over the last few seasons has primarily been a way to get open mid-range jumpers for Zak Irvin in the middle of the floor. His inconsistency with this shot was maddening to Michigan fans at times, but the action is still a staple of Beilein’s offense.

When I think about how this action might look in Michigan’s offense in 2017-18, I envision much more Manny Harris than Zak Irvin. That’s because I can see Charles Matthews playing this same role, but using it as an opportunity to slash to the cup.

Here’s a look back at Manny Harris using an almost identical set to slash to the basket against Oklahoma in the 2009 NCAA Tournament.

  • jblair52

    Love the breakdown! I like that you dug deep and found the Manny-Charles comparison for potential hand offs. Impressive!

    When you break it down like this, it makes me less anxious about replacing the production. We look to be fine if guys contribute as expected.

  • Champswest

    I think this is likely one of those UM teams that isn’t lead offensively by one or two stars. We have multiple guys that are capable of averaging double digits or going for 17-24 points in any one game.

  • bobohle

    I think Austin Davis will have no problem receiving the ball in the pick and roll if his hands are is good as JB says they are. Whether he gets the rock thru the hoop consistently is the question. Also do not under estimate what X will mean to this team on both ends of the floor.

  • MrLG

    I assume that you would put alley-oops in the Cut category, but I think it should be broken out. Bottom line: we were pretty terrible considering we had DJ and Mo to finish, but DW just wasn’t that good at this. Look for a solid improvement with Simmons.

    • The Synergy play types are representative of the action that creates the shot more than the actual shot. So an alley-oop could come any number of ways (transition, cuts, p&r roll man, etc.), but yeah usually probably a cut of some kind or in transition.

      Are you assuming that Matthews will be a better alley-oop finisher than DJ? The easiest way to throw an alley-oop is if you have a guy who can go get anything and Matthews certainly has some bounce, but DJ was pretty long.

      • MrLG

        We didn’t have many alley-oops last year and I believe the problem was the PG. With a change in PG, we should do better. Matthews could potentially be an awesome alley-oop finisher, we’ll see. I’m hoping that Davis will be good at this as well. Maybe even Duncan for a few as he cuts well. And there’s Mo, of course.

        • I think the number of alley-oops is a lot more dependent on the number of athletes on the roster. I wouldn’t put Duncan Robinson or Austin Davis high on the list of guys to throw alley-oops to. Duncan’s backdoor cuts are usually that little alley-oop layup which is more just to help him get off the ground quicker.

  • GTFOmycourt

    I am surprised livers was not mentioned. He seems to be the most obvious replacement for DJ. Maybe he will factor in as a replacement within the cutter category?

    • I ended up cutting Livers out of the cut portion just because the article was getting long. If he can give rotation minutes at the 4 that is huge, but his game is more mid-range jump shooting IMO than a pure cutter. Of course that could change this year in a new role from his HS, etc. If he’s Michigan’s best scorer in that are and does well, that’s a couple extra wins IMO.

    • Leslie Hoerwinkle

      Guessing that Livers will find that the college game is a big adjustment from high school.

  • rlcBlue

    This is a great analysis, Dylan, but it might lead people to the wrong conclusion. Beilein’s offense is not designed to get a fixed percentage of certain types of shots – it’s supposed to force the defense to give up a good shot; what type of shot that is doesn’t matter so much as long as the offensive players recognize the shot and take it.

    What qualifies as a good shot depends on the personnel on the floor – when Wagner, Livers, Robinson, Rahk and Poole are on the floor, any open 3 is a good shot; if it’s Simpson, Matthews, Watson, Davis and Teske, not so much. The proportion of types of action that lead to scoring will shift depending on who is on the floor. Who is on the floor will depend on who can understand the offense and who can perform on defense.

    Luckily, the ball screen offense can be made very simple, and Simmons has plenty of experience executing it, so that will be available from Day 1. Hopefully Matthews and Davis have enough practice experience under their belts to run the more intricate parts of the offense. I don’t hold out much hope for the true freshmen beyond running the break, spotting up for jumpers and making the occasional cut to the basket.

    What will our strongest defensive unit be? Who knows? Probably Simpson, Rahk and Matthews on the perimeter, but the 4 and 5 are complete mysteries to me. Wagner will be on the court as much as possible, but will he be an asset or a liability on defense? Can Teske use his size for anything more than changing light bulbs? Is Livers’ athleticism enough to make him a better defender than Robinson, or does Duncan’s experience give him the edge? The answers to these questions will determine the minutes distribution, which will dictate where the scoring will come from.

    • Leslie Hoerwinkle

      I know you’re probably being illustrative, but in what world would Teske, Davis and Watson all be on the floor at the same time?

      • rlcBlue

        A very, very unhappy one for Wolverine fans.

    • I get what you are saying, and several categories talked about potentially seeing more or less possessions of a certain play type. On the other hand, these #s don’t change too much year to year (I could run the numbers and check for sure), and the one trend we’ve really seen is the increase in PNR offense.

  • Melatoto

    I just want to say that Duncan’s open 3 was a beautiful play.

  • Barth Applefeld

    Great article as always, and I appreciate the way you broke the plays down into categories. I’m sure that you did it before, but could write something up on why Michigan’s put back rate is so atrocious. There are some obvious reasons that are not bad necessarily: M makes a lot of shots so there is less opportunity for put backs. M has had offensively limited big men so that they can’t consistently rebound (Wagner so far) or can’t put back (JMorgan). M’s offense is not the type, in general, to produce a lot of put-backs because of ??? Thanks

    • Michigan doesn’t crash the glass very often as a team, opting instead to get back on defense. For example, some teams will crash 2,3,4 players to the glass whereas Michigan will usually send 1. Now that has to do with offensive rebounding rate, not necessarily putbacks. But I assume you are talking about the # of putback possessions that Michigan had last year, which is obviously very dependent on that.

      • Barth Applefeld

        Yes, thank you, but that goes to the type of players they recruit and what they teach them to do in the time that they have them, doesn’t it? Tarpley, Vaughn,Relford and crew could rebound the heck both offensively if I remember correctly and still get back on transition enough times.
        UM not ranking above 275th except three times in all these years is so statistically significant for a program that has enjoyed the success that Michigan has that it cries out for a fuller explanation than they chose to get back on defense. Why must they sacrifice offensive rebounding so much compared to all other teams. How can they at least become mediocre and still defend in transition. If a team has to worry about getting a defensive rebound, then they can’t transition as quickly or with as many players. My thought: I think it has at least something to do with athleticism and height, but even UM’s more athletic and tall players have consistent problems, but they too often can’t play defense when they get back anyway.
        That’s why I wish Donovan had stayed, I would think UM could at least be mediocre on offensive rebounding and average or above on defense. It reminds me of RichRod sometimes: great offensive mind but no mind at all for defensive aspects of the game. Just saying. Love Beilein and you’re the best!

  • nswan

    Love the use of GIFs. So much easier to see the play than linking to a youtube video and loading each. Great article