Defensive Report Card: Individual stats, transition problems

Michigan 80, Penn State 67 - #21Michigan’s defensive struggles have been obvious over the last month, and especially in the last two games. Heading into a stretch of schedule that will pit them against some of the best offensives in the country, the Wolverines need to figure out a way to make improvements on that side of the ball – quickly. (Photo: Scott Mapes)

Evaluating defense can be difficult because there are only so many metrics available. Team defense can generally be analyzed by the same statistics as offense, but tracking defense at an individual level is extremely difficult. The only standard statistics kept are blocks, steals and rebounds, which obviously aren’t enough to make a complete judgment of a players’ defensive effort.

There are some methods available – including defensive score sheets and Synergy Sports statistics – which can get us closer. After the jump we look at two different sets of individual defensive stats and take a closer look at one team-level issue hurting the Wolverine defense.

 

Synergy Spots Statistics

I posted individual defensive numbers from Synergy Sports in mid-December and have updated numbers to pass along. Synergy Sports logs and categorizes every play and assigns the blame or credit to a given player. The resulting numbers tell us how many possessions a player has defended and how many points have been allowed, providing a points per possession defended figure. (Disclaimer: These numbers are from before the Penn State game.)

image

Robinson grades out as Michigan’s best defender by a wide margin according to these stats. Stauskas, Irvin and Albrecht were all better on defense in the last month. Of that group, Irvin’s performance looks most promising. LeVert’s team-leading usage rate is a sign that opponents continue to attack him, and are still being fairly successful.

Defensive Score Sheet: Penn State at Michigan

The concept of a defensive box score has been around for a while, a movement spearheaded by Sports Illustrated’s Luke Winn and David Hess. It’s not much different than the Synergy model with the exception being that the numbers are tracked internally (by me) and some calculations go into the final numbers.

I took a crack at tracking each defensive possession against Penn State and came up with the following figures.

image

A quick run-down on some unfamiliar stats above. The following stats are tallied manually throughout the game.

  • FM - Forced field goal miss (includes blocks)
  • FTO - Forced Turnover (steals, charges taken)
  • FFTA - Forced missed Free Throw Attempt
  • DFGM – Allowed Defensive Field Goal Made
  • DFTM - Allowed Free Throw Made

Defensive Rating (DRtg) is calculated based on the stops and scoring possessions assigned to the player, it’s an estimated measure of points per 100 possessions.

In this first pass I assigned just one player to every Penn State shot attempt. Other charters have split shot attempts for shared defensive efforts and this is something I plan to start doing in the future.

There are things these stats (both defensive score sheets and Synergy numbers) don’t account for (most importantly: who guards the best players?), but they could provide some value going forward – at least providing a benchmark for discussion regarding the defensive effort after each game.

Observations

  • Jon Horford played a great game on both ends of the floor. His rebounding was great and he was solid defensively. Penn State’s big scoring runs generally came when he was on the bench.
  • The numbers are ugly for Glenn Robinson III and Nik Stauskas. Stauskas’s issues were documented in Five Key Plays, while Robinson had an up and down game. He was praised by John Beilein for his effort against DJ Newbill late in the game but he gave up the majority of his baskets to Donovan Jack and Ross Travis.
  • Caris LeVert wasn’t great, but he was better than his wing counterparts because he was able to force a handful of turnovers.
  • Spike Albrecht’s defensive numbers look pretty good, but he was mostly tucked away guarding Graham Woodward and John Johnson. Neither player tried to do much other than fire up threes. He did cause a couple turnovers including drawing an offensive foul on a push off in the first half, further bolstering his stats.

Transition defense

While the individual numbers are interesting, it’s impossible to discuss Michigan’s defense without looking at its problems in transition.

The Wolverines are allowing a 66.1% eFG% in transition situations, the second worst in the country. Michigan’s half-court eFG% allowed is a more impressive 43.7%, but that leaves a 22.5% gap between Michigan’s halfcourt and transition defenses. That transition defense drop-off is the largest in the country.

Michigan’s saving grace continues to be its ability to slow down games and limit transition opportunities. The Wolverines are ranked 58th nationally in percentage of initial field goal attempts allowed in transition. If they were allowing as many transition opportunities as an average Division I team, their defensive numbers would be much uglier.

That manifests itself in other statistics as well. The Wolverines have essentially stopped crashing the offensive glass – failing to post an offensive rebounding rate over 31% since December 7th – and are playing the slowest basketball in the Big Ten at 60 possessions per game. Those adjustments in strategy are almost certainly caused, at least to a degree, in protecting an underperforming transition defense.

Bottom Line

Michigan’s defense is a problem. There’s no overnight fix and there isn’t an easy scapegoat.

The Wolverine offense appears to be very good, good enough to win a few games that an average to good offense probably wouldn’t. But, sooner or later, the defense is going to start costing Michigan games.

  • Picknroll21

    Good article as defense is definitely key! Everyone has to step up big time on D and of course rebounding. Morgan and Horford have to be tough inside. Robinson, would like to see him block more shots as well. Has to be a total team effort and am looking forward to see how the team comes together during this stretch. I feel like they are ready! Go Blue!

  • mikey_mac

    The next step, which Dylan alluded to, is to compare against the competition’s average ORtg, to see a +/- versus that, to neutralize for individual matchups. For example, GRIII giving up that 119 on possessions for which his opponents would typically average a 110 isn’t nearly as bad as Albrecht giving up a 106 against guys who average 90. (The ORtg averages are just notional in my example.)

    • http://www.umhoops.com/ Dylan Burkhardt

      The problem with that is that with switching and everything, that becomes too much to track. Synergy could probably come up with something like that because they literally log every play.

      • mikey_mac

        Yeah, I wasn’t trying to suggest you spend your next 12 hours going back and logging matchups! That would be painful. I would just love to see that next step turned into a metric.

  • mikey_mac

    I think FFTA is supposed to indicate fouls on shooters (Forced Free Throw Attempts). I’m not aware of any way to force a missed free throw, other than waving your arms/praying/voodoo dolls.

    • http://www.umhoops.com/ Dylan Burkhardt

      Yes.. Fouls on shooters (that were missed) vs. fouls on shooters (that were made).

      FFTA vs DFTM

      • Giebz

        I still don’t get it. Can you give a little more explanation on what a forced free throw miss or an allowed free throw made are?

        • http://www.umhoops.com/ Dylan Burkhardt

          Player A fouls Opponent A. Opponent A makes 1 of 2 free throws. That’s one forced free throw miss and one allowed free throw made. It’s just accounting for all the FTs…

          • mikey_mac

            OK. I guess my confusion is, why does the result of a free throw attempt factor into a defensive metric?

          • http://www.umhoops.com/ Dylan Burkhardt

            A full rundown of the calculations can be found here: http://audacityofhoops.blogspot.com/2011/01/project-defensive-score-sheet.html

          • mikey_mac

            Ah, thanks! That webpage refers to Appendix 3 of Basketball on Paper, which describes the formulas in great detail. (It can be found on Google Books.)

          • JGiebz

            I see, you need to know the result of the free throw to determine the points/100 possessions. My first reaction would have been to say that you can’t control the free throw so why not count every free throw attempt as 0.75 pts or whatever the opposing team average is, but I guess you could argue that a defender could intentionally foul a poor free throw shooter as part of a defensive strategy. I still think the calculation would be more accurate if you took out the actual result of the free throws and instead multiplied the number of free throw attempts a player was awarded by that individual player’s season ftm average. That way if you intentionally foul a 45% shooter and he happens to get lucky your defensive stats aren’t brought down as much.

  • Mattski

    I wonder, apropos of this, what people made of this analysis at mgoblog, in which the poster asserts that M is far more effective with Horford in. Obviously his analysis takes in defense less than offense, but maybe Horford’s shot-blocking ability is enough to cinch this argument in his behalf? To quote, in part:

    “Horford is currently 2nd among active players in PER and ws/40, trailing only Stauskas. His rebounding % are better than even McGary’s and his block rate is double that of the next best. Morgan is producing well enough when hes in, but on top of the discrepancies in individual numbers Horford has a DRtg 10 points lower. Basically he is producing at almost the same rate McGary was defensively, exceeding that on the glass, and is far and away the best shot blocker.”

    http://mgoblog.com/mgoboard/horford-vs-morgan

    I don’t feel like I have a dog in this fight, want to see both of them thrive, but maybe this is evidence that Horford is really coming on. I know I’ve had the belief–not quite fulfilled until recently–that he could be serious-good. But I think there’s also been the sense that Morgan was the guy you wanted in when you needed stops. Obviously, both of their offensive games have been both more reliable and developing some, too. . .

    Hope like heck we can become a little more reliable on D, and don’t just get completely exposed down the stretch here some time when the O goes cold on the road.

    • guestavo

      PER is an outdated stat.

      • Mattski

        Well, I know it emphasizes offense, and tried to anticipate this in my comment. Beyond that, I bow to–and solicit–the greater knowledge/wisdom of my fellow boarders. :)

        • guestavo

          Not even a good offensive measure, since it favors volume scoring more than it does say efficiency or the difference between different kinds of shots.

          • Mattski

            You clearly possess a lot of knowledge, guestavo, but you’re also just a total buzzkill. Your responses never seem designed to elicit further back and forth, just to demonstrate your superior knowledge. If you’re not here for the give and take, why bother to post? Your ego can’t be THAT weak.

          • guestavo

            You take it as a “demonstration” when I was attempting to “inform.” I don’t claim to know everything but if I do know something and see someone who misses some of the nuance, why not correct it? I guess I come off as cold on here and should do a better job of communicating on here since some think I’m going out of my way to be a d*ck.

    • MGoTweeter

      Horford does provide more rebounding and help side blocking than Morgan but he is far worse in other areas. His on ball defense in the post is not nearly as good as Morgan. He is also much worse in pick n roll situations mainly because he plays far too vertical but he is also slower than Morgan in recovering. The vertical issue also hurts when he helps on drives in the lane since it is very easy to make passes around him.

      That said I thought he was much better against PSU than he has been all year. He did a better job of staying in a stance and seemed to be moving better in recovery. Both guys have their strengths and weaknesses and it is probably going to remain a hot hand approach for the rest of the season with a slight lean towards Morgan because he is a bit more consistent on both ends of the floor.

  • guestavo

    Wouldn’t Synergy’s % time just indicate that Robinson and Levert usually guard the most potent offensive players?

    • guestavo

      In addition to steals, deflections would be cool to keep track off but probably much harder to do?

    • mikey_mac

      Yeah, my initial takeaway was that LeVert is doing better than I thought, since I feel like he ends up with more difficult assignments. And that Walton and Albrecht are doing even worse.

      • guestavo

        Penn St game shows improvement for Walton so hopefully that trends. Spike is guy who has to create more offense than his opponent which he usually does in his 10 minutes or so. I have to think that guys like Robinson and LeVert are being kind of handicapped by not using their length to defend because JB doesn’t want them to pick up fouls i.e. Charlotte game.

  • MGoTweeter

    Despite what the numbers say, I think we have seen a lot of positive signs over the last couple weeks that the defense is improving, albeit very slowly. Especially in the PSU game where the half court defense was pretty solid except for a stretch at the start of the second half. I don’t think Michigan has the pieces without McGary to be good defensive team, but they can get to average which might be enough to help them make a run.

    I still believe that with the new rules, Michigan needs to get more aggressive on the perimeter. Beilein’s man defenses have always been more containment without fouling oriented with reliance on allowing the ball to certain areas and taking charges when needed. In past years even if you were sitting back on your man, you could redirect a ball handler coming straight at you to a certain extent and if that was not feasible you could rely on a helpside player stepping in to take a charge. Now once that ball handler gets going in a straight line at the basket, it is virtually impossible to stop them without fouling. This is especially true for teams that try to avoid fouls like Michigan, since the occurrences of contact on drives is less common than other more physical teams, making it more likely that the refs will call a foul.

    I would like to see Michigan try and pressure the ball a little more to force drivers away from straight line drives. It probably means a few more silly fouls on the perimeter but hopefully less fouls at the rim on drives and less no contest layups. I would also like to see the 1-3-1 worked in a little more randomly, even if it is not that effective, just to change the rhythm of the opposing offenses.

    • mikey_mac

      I think it’s fair to say JB took a wait-and-see approach to the rules changes. Now that we’ve seen, I agree it’s time to adjust. You need to prevent clean lanes to the bucket without trying to resort to drawing charges. I think fouling before the ball handler gets to the rim is an option, but I think packing the paint on the ball side might be just as effective, without getting opponents into the bonus early.

    • guestavo

      He needs to seriously reconsider the 2 foul rule

Previous post:

Next post: