[Cracked Sidewalks] Is Defense Marquette's Achilles' Heel?

[CrackedSidewalksSays]

Is Defense Marquette's Achilles' Heel?

Written by: noreply@blogger.com (Rob Lowe)

After last year's senior day game (which Marquette lost.  Again.  Of course) I had a conversation with someone at a Milwaukee watering hole.  This person knows a lot more about basketball than I do, having served as both a team manager and a high school coach.  Me... I mostly work with numbers and watch hoops. I was complaining that Marquette's defense was not good.  His response was that Marquette was playing defense a lot better than they were at the beginning of the year.

We were both right.  Back to this in a minute.  Let's consider how important defense is to any program and what the gaps are for the Warriors.

Point #1 - Marquette will never be elite until their defense is elite.

The best summary of this fact comes from CNNSI's Luke Winn, whose holiday tradition of Early Warnings shows which teams are pretenders or contenders around New Year's Eve.  What I particularly like is that he shows the offensive and defensive efficiency of NCAA Elite teams.  Last year's list had the Elite Eight teams going back to 2003-2004.  Winn's 2010 Early Warnings List (ps - he bags on UConn and ND) shows other teams that might be contenders or pretenders this year.  Now, the point of Winn's list is to demonstrate which teams, ranked or unranked, might be real Elite Eight/Final Four/National Championship teams and which ones might not.  However, it's all built on the notion that there is a profile for each Elite Eight team.

The average

Elite Eight team has a national Defensive Efficiency Ranking of 16.  The average Final Four team has a rank of 9. For giggles, National Champs average 5.3 nationally in defensive efficiency.   But once your team's defensive efficiency gets worse than 25, you're really talking about a crazy run to make the Elite Eight (think Pittsnogle's WVU team or Stephen Curry's Davidson team).  Honestly, even when there, Elite Eight is your ceiling.

But what does this mean for Marquette?

Here is the same chart with Marquette's defensive efficiency numbers from the last two and a half years -- the Buzz Williams Era.



There's the issue.  Marquette has managed to develop a nearly elite offensive capability in the last three years.   Buzz not only modifies the offense to the strengths of his team, but does it so effectively that the team runs a top 25 offense.

However, no matter how good MU is offensively it's the defense that will continue to hold the program back and is the reason that, despite an elite offense,  Buzz's teams play so many close games.  Ultimately, MU's inability to defend at a high level decreases the program's margin for error night after night.

Let's look at Buzz's progression as the MU head coach. Yes, I know that two years ago, Marquette was short.  And last year MU was also short with an even shorter bench.  This year Marquette is taller but now more inexperienced.   Still, rationalizing poor defensive efficiency eventually just becomes a pile of excuses, especially after 2.5 years.  After all, MU successfully makes the adjustments offensively year to year, and Buzz inherited a program that was consistently a good defensive team (#31 defensively in 2007 and #10 defensively the year before he became head coach).

Point #2 - Not all aspects of defense are the same

Essentially, defense can be broken up into the four factors.  Those are not letting your opponent get good shots, forcing turnovers, defensive rebounding, and not letting your opponent get to the line.    However, their relative weights are not the same.  This is very important.  Very.  Important.



If one breaks down what contributes to defense, 27% of defense just comes from facing a D1 opponent for 40 min.  However, 41% of defense is effective Field Goal percentage (eFG%). Defensive  eFG is twice as important as forcing turnovers.  Effective field goal percentage is three times as important as preventing offensive rebounds.  And almost fourteen times more important than not letting your opponent get to the free throw line.

What this tells us is that the primary goal of a defense should be preventing open looks.  Then it's a tossup between forcing turnovers and limiting offensive rebounds (although forcing turnovers is better).  Finally, and way down the list, is preventing your opponent from getting to the line.

But what does this mean for Marquette?



Here's how Marquette ranked/ranks on the four factors.  This is the big issue with Marquette's defense and the ceiling the Warriors' defensive performance places on the program.  The aspect that is the most important is where MU ranks the worst.  Consistently.  The area that is the least important for defensive performance is where Marquette ranks the best.  Consistently.  What the numbers say is that Marquette is emphasizing the wrong type of defensive priorities, year after year.

Vandy, West Virginia, and NCAA games

The last two games against Vanderbilt and West Virginia show how Marquette remains competitive.  Our offense is great.  Realize that an average offense scores 1.00 points per possession (ppp) and Marquette scored 1.10 ppp against Vanderbilt (who has the #21 defense and only allows 0.88 ppp).  On the road!   MU had an offensive eFG% of 60% against one of the nation's better defensive teams.  However, as that final basket showed, MU just couldn't make the stops.  Against West Virginia, Marquette had an extraordinary 1.22 points per possession but gave up a remarkable 1.15 ppp while allowing WVU to shoot an eFG% of 56%.  With offensive efficiency of this magnitude, MU should be winning games like these by double digits.

What MU is doing is winning games with their offense, but creating the lack of margin because of lousy defense.  Last year Marquette had an Offensive Efficiency  of 1.16 points per possession against Washington.  Two years ago MU  scored 1.11 points per possession against Missouri. Those efficiencies should have been good enough to win an NCAA game.

With this recent history, does anyone really believe Marquette can play lock-down defense when it matters?

Which brings us back to the story from the beginning of this post.  I have no doubt that Buzz can coach defense.  Well... better said, I don't know basketball well enough to question if the coaching staff is properly teaching defense.  So I trust when someone that knows basketball better than I do says the team is playing better defense.  He's right, MU is demonstrating modest improvement. Yet, the numbers are the numbers.  Not only has Marquette consistently not played good enough defense the last two and a half years, but it's been neglecting the most important aspect of defense and emphasizing the least important.  We're both right.

http://www.crackedsidewalks.com/2011/01/is-defense-marquettes-achilles-heel.html

[brewcity77]

Excellent article. Does a good job of conveying its point and I guess the positive that comes out of it is that our eFG% seems to be continuously improving, Granted, 150 isn't good, but it's better than 252. Hopefully that keeps going in the right direction, and the maturation of some big bodies up front can help with those rebounding percentages.

Just checked on kenpom.com and he has us at #25 in offense and #73 in defense. I would be curious to see how much 3-pointers influence a team's success in March. So many small teams seem to make deep runs with the three-ball, but I suppose that can be an Achilles' heel too when it goes cold.

[bilsu]

Two other factors:
It appears to me that MU is very good at playing defense without fouling, which is a plus. MU so far has not had a single player foul out this year. There cannot be too many other teams that can say that.

Defense ends when the defensive team secures the rebound. Generally MU losses can be pointed to MU giving up to many offensive rebounds, which is  a hugh negative.

Two observations:
Crean had big centers. JohnMueller(?), Scott Merritt, Robert Jackson, Marcus Jackson, Ousman Barro. Also had Claussen and Grimm. Until Buzz has similar bigmen his defense will suffer in comparison to Crean's.

I think you will see MU's overall defense improve as Buzz uses less players. The weaker defenders will not be playing much in Big East games.

[dw3dw3dw3]

Very interesting, thanks for that.... I believe both Huggy and Stallings noted there were some easy looks because of the pressure early on in games, but as the game progressed MU backed off in its overall defensive pressure to negate that.  Your stats suggest the pressuring defensive philosophy probably isn't worth it to get a few more steals since MU would give up open looks in return (against good teams).  I wonder if Buzz is adapting to this since I would think he knows  where his defense ranks.

I thought we played pretty good defense against WV from a perspective of having a hand in the face on every shot. They seemed to be hitting a high percentage of contested shots. The back door seemed to be shutdown. Usually our pressure and rotations lead to wide open 3s.

[MarquetteDano]

Very interesting analysis.  One thing I will say about not fouling is that it keeps your best players on the floor longer.  That cannot be quantified in the statistics presumably can it?  Crean's teams fouled a lot more and I can remember games where our best players only played 15 minutes due to foul trouble.

It does beg the question though that would it be wise near the end of games to become more aggressive in not giving up open looks and potentially fouling more (not a purpose of course)?  I don't know if you can realistically teach kids to play defense one way for 35 minutes and then tell then another way for the last 5 however.

[tower912]

Is there a statistical analysis for points created off of steals?   IIRC, we generated a lot of points off of our steals against WVU.   When we stopped getting steals off of poor passes, WVU got back into the game. 

[Henry Sugar]

Two other factors:
It appears to me that MU is very good at playing defense without fouling, which is a plus. MU so far has not had a single player foul out this year. There cannot be too many other teams that can say that.

Defense ends when the defensive team secures the rebound. Generally MU losses can be pointed to MU giving up to many offensive rebounds, which is  a hugh negative.

Your first point is defensive free throw rate (FTR).  That is how often an opponent shoots free throws per shot attempt.  That's the part that is 3% of defense.  Your second factor is defensive OR% (or just defensive rebounding).  That's the part that is 12% of defense.

Two observations:
Crean had big centers. JohnMueller(?), Scott Merritt, Robert Jackson, Marcus Jackson, Ousman Barro. Also had Claussen and Grimm. Until Buzz has similar bigmen his defense will suffer in comparison to Crean's.

I think you will see MU's overall defense improve as Buzz uses less players. The weaker defenders will not be playing much in Big East games.

In Crean's final year, our effective height ranked 142.  The last two years our effective height was 305 and 308, respectively.  This year we are back up to 195, so our height has improved overall but the defense has not.  I guess I'll just say you might be right, but let's see how the numbers get better (especially eFG%).

Also, did you just give Crean credit for big men?  what?

[GGGG]

I also hope that this article dispells the notion that we don't attack the zone effectively.  Vandy and WVU played almost entirely zone, and we scored 1.10 and 1.15 points per possession respectively.

Crean had big centers. JohnMueller(?), Scott Merritt, Robert Jackson, Marcus Jackson, Ousman Barro. Also had Claussen and Grimm. Until Buzz has similar bigmen his defense will suffer in comparison to Crean's.

But isn't that Otule's role?  I mean, he's a lengthy 6'11" guy with very limited offense.  

Honestly, I think part of the problem is that Butler and DJO are *at best* average defensively.  DB and Blue are much better of course.

[Canadian Dimes]

I agree Butler and DJO suck on defense....2 or 3 times on Staurday Otule on the weak side block helped stop the ball on the ball side block leaving his man on the weakside block.  In all occassions Jimmy butler who was on the weakside elbow stood and watched.  Each time the shot went up and caromed to the WVU player (Jones) who scored or got fouled as Butler rotated late.  DJO is lost.  Oyule is improving but is still not htere yet.  croweder has been here a few months.

I agree our defense is our achilles heal when you are scoring 75+ points it should not be a one possesion game.  I think it will get better.  As otule and Crowder play more.  Our best defenders are Buycks and Blue

[NersEllenson]

Some good analysis.  That said, the FT rate seems to be undervalued.  Also this doesn't take into consideration quality of competition/SOS.  The Big East has been the best conference the last 3 years, arguably..better offenses to defend..

[Henry Sugar]

Just checked on kenpom.com and he has us at #25 in offense and #73 in defense. I would be curious to see how much 3-pointers influence a team's success in March. So many small teams seem to make deep runs with the three-ball, but I suppose that can be an Achilles' heel too when it goes cold.

Three pointers help increase your odds of an upset.  There are studies out there, and I've talked about it before.  If you are an underdog, your team benefits by playing higher variance strategies (more threes, pressing) from a risk/reward perspective.  The three is a shot that's worth more but goes in less.  I think an underdog can increase their odds of winning by something like 5%-15% depending on how many threes they chuck up.

Very interesting analysis.  One thing I will say about not fouling is that it keeps your best players on the floor longer.  That cannot be quantified in the statistics presumably can it?  Crean's teams fouled a lot more and I can remember games where our best players only played 15 minutes due to foul trouble.

This was my starting point with Buzz's first year.  We fouled less but were so good offensively (#9) because our best players were always on the floor.  Plus, I think part of the final defensive numbers are skewed because James was out.  

Is there a statistical analysis for points created off of steals?   IIRC, we generated a lot of points off of our steals against WVU.   When we stopped getting steals off of poor passes, WVU got back into the game.  

One could be generated, but the data would have to be gathered manually.  No offense, but I'm not doing that.

Some good analysis.  That said, the FT rate seems to be undervalued.  Also this doesn't take into consideration quality of competition/SOS.  The Big East has been the best conference the last 3 years, arguably..better offenses to defend..

No, it isn't.  FTR is not that important.  You may believe it is, but regression analysis disagrees with you.

Also, here are some defensive rankings for other BE teams
2008-2009 = UL (#2), UConn (#3), Nova (#15), Syracuse (#29), WVU (#14), GU (#22)
2009-2010 = Cuse (#18), WVU (22), Pitt (26), GU (47), Uconn (33), SJU (51)
2010-2011 = Cuse (19), Cincy (21), Pitt (35), Nova (20), SHU (39), UL (17).  There's more.

[GGGG]

Three pointers help increase your odds of an upset.  

Which is probably why WVU and Davidson both exceeded tournament expectations despite their poor defense.

[Henry Sugar]

Here is my question, for someone that knows basketball.

With the current roster, how would one adjust defensive strategies to focus on effective field goal percentage defense? 

[babytownfrolics]

Here is my question, for someone that knows basketball.

With the current roster, how would one adjust defensive strategies to focus on effective field goal percentage defense? 

I don't know basketball, but could the philosophy against fouling too much be adjusted to more in your face, physical defense that risks fouling more when Buzz begins to develop the type of depth throughout the roster that he wants to have?  In other words, we couldn't foul as much the last 2 years because we had almost no one capable on the bench of contributing.  We have more capable contributors this year but maybe not to the level that Buzz trusts completely?

[Dr. Blackheart]

Excellent, excellent article Hank!

Last season, we were small and had limited depth which affected Buzz's strategy. He limited the tempo (# of possessions), tried to limit the fouls, and defended the perimeter (#1 in the BE in arc defense). But we got pulverized inside (2FG% and rebounding).

This season, I get the feeling he is still deciding. In the "Revealed" preseason episodes, he was stressing rotations in defending the paint.  So far, our 2FG% defense is 45.5% which is 104.  Last season it was 50.0% or 263rd. So, we are defending the paint better and not fouling (our FT rate defense is one of the tops in the country--23.9% FTR or 4th nationally). We are 2nd in the BE in not giving up points  (Pts./Game 67.2 or 2nd). And that isn't all bunny driven as the biggies were also held in relative check.  Rebounding (as painful as it seems) is also improved. So, Buzz is trying to limit exposure through positioning.

However, in defending the paint, we are 250th in defending the trey. In all our major games and many bunnies, we were being scotched on the perimeter.  Buzz then made in-game adjustments, and then we get beat on the inside (as seen in offensive rebounding, Plumlee, Huggy/Stallings going to backpicking, etc.).  Our defense becomes worse when Buzz has gone small in these match-ups (all bigger teams in our losses). Otule and Gardner take up space and should always be in the game (space eaters) as it allows us to extend our defense to cover the arc--limitations noted.  

We are not good at team defense--rotations, switches, looks, schemes. Look how the Illini played the Badgers yesterday. They went man and 2-3 match-up...always bracketing Taylor as he is the engine and the only one who can create a shot. U of I is big and quick granted, but UW could not get inside and create--so they became chuckers.  Marquette let Taylor have his way with us and let him penetrate which created open looks for the others. The end of the Vandy game is an example where we can be broken down.  

[MarquetteDano]

Excellent, excellent article Hank!
 Our defense becomes worse when Buzz has gone small in these match-ups (all bigger teams in our losses). Otule and Gardner take up space and should always be in the game (space eaters) as it allows us to extend our defense to cover the arc--limitations noted.  

I basically agree with this with one minor change.  I think Otule, Gardner, OR FULCE should be in the game at all times.  Otule is obvious.  When he is not in I think either Gardner or Fulce should be in.  I lean towards Fulce right now especially if there is an athletic front line.  And especially if the five spends a lot of time outside the paint (e.g. Leuer type).

[Jay Bee]

No, it isn't.  FTR is not that important.  You may believe it is, but regression analysis disagrees with you.

I'm struggling with this one too.  Can you help out with how you get to a 3% figure?  I could deal with it being 15%-20%, but 3%?  

[GGGG]

I don't know basketball, but could the philosophy against fouling too much be adjusted to more in your face, physical defense that risks fouling more when Buzz begins to develop the type of depth throughout the roster that he wants to have?  In other words, we couldn't foul as much the last 2 years because we had almost no one capable on the bench of contributing.  We have more capable contributors this year but maybe not to the level that Buzz trusts completely?

We are already in your face enough.  In fact, too much IMO. 

If I were Buzz I would pack it in a little more and even consider some sort of zone.  You can hide defensive issues in a zone occasionally, but I think one of the reasons he doesn't do this is because much of our offense is generated by turnovers, and you don't get turnovers out of a zone that often.

[dwaderoy2004]

I'm struggling with this one too.  Can you help out with how you get to a 3% figure?  I could deal with it being 15%-20%, but 3%? 

agreed.  I suppose when we only allow 7 free throws, it is a low percentage.  But if we gave up 20 free throws/game, how is that not more like 20%.

[dwaderoy2004]

We are already in your face enough.  In fact, too much IMO. 

If I were Buzz I would pack it in a little more and even consider some sort of zone.  You can hide defensive issues in a zone occasionally, but I think one of the reasons he doesn't do this is because much of our offense is generated by turnovers, and you don't get turnovers out of a zone that often.

We already have trouble rebounding the ball.  A zone would only make it worse.

[Henry Sugar]

I'm struggling with this one too.  Can you help out with how you get to a 3% figure?  I could deal with it being 15%-20%, but 3%?  

Short answer = when you control for eFG%, Turnovers, and Offensive Rebounds, opponent's free throws are not as important.

Long answer = I regressed Marquette's defensive efficiency on def eFG%, def TO%, def OR%, def FTR, and pace.  Adjusted R^2 was at 95%.  All the t-stats/p-values were significant at a 99% value.  Pace was not significant.

Coefficients for each Factor based on the regression analysis
Intercept 41.9
eFG% 1.28
TO%    -1.25
OR%    0.59
FTR   0.14

Average Value of each Factor
eFG% 50
TO%    21.8
OR%    31.3
FTR   29.3

Sum of contribution
Intercept 41.9
eFG% 1.28 * 50 = 64.4
TO%    -1.25 * 21.8 = -27.3
OR%    0.59 * 31.3 = 18.4
FTR   0.14 * 29.3 = 4.2
Total = 101.7 (compared to actual of 98.2 - pretty close)

I took the absolute value of each factor for a total of 156 and then divided each contribution above by 156.  4.2/156 = 3%

[Henry Sugar]

We already have trouble rebounding the ball.  A zone would only make it worse.

defensive eFG% is three times as important as defensive rebounding.

[dwaderoy2004]

fair enough.  you should honestly consider writing a letter to Buzz.  He would probably read it, and with as numbers-centric as he is, probably consider it.

[Jay Bee]

Short answer = when you control for eFG%, Turnovers, and Offensive Rebounds, opponent's free throws are not as important.

Long answer = I regressed Marquette's defensive efficiency on def eFG%, def TO%, def OR%, def FTR, and pace.  Adjusted R^2 was at 95%.  All the t-stats/p-values were significant at a 99% value.  Pace was not significant....

Ummm, I don't recall Probst teaching me this.  Thanks for the answer, I'll try to understand this a little better.. but, the 3% just doesn't pass the reasonableness test for me...

If I try to keep it simple, I think of putting another team on the line as a 1.38 ppp situation for them (ignoring a slight adjustment for front ends)... 2 shots, 69% avg FT shooting... maybe it's the fact you're controlling for the other factors..  

[Henry Sugar]

How about this for reasonableness?

We held WVU to a FTR of 11.7%.  We held Vandy to a FTR of 25.4.  UW@Madison had a FTR of 25%.  Those are all awesome for FTR.  By contrast, we are #4 nationally at def FTR at 24%.

And yet those opponents still had 1.15, 1.12, and 1.19 points per possession, respectively.  That is not awesome. 

How could our FTR be so good, and yet our defense so bad unless FTR were practically insignificant?