Franalysis: Player Tracking Results From 11 High Major Games

By houksyndrome on January 28, 2021 at 10:13 am
go hawks go
© Joseph Cress/Iowa City Press-Citizen via Imagn Content Services, LLC
8 Comments

This article will have a somewhat different format from how I usually write them.  Instead of a long article with an in-depth focus on one topic, I'm going to bounce around between a few different topics.  Think of it like The Hybrid but with math instead of humor.  Sounds fun, right?  Away we go.

The Indiana game

Yeah, that one was no fun but there was one interesting thing that I noticed.  In my last article I noted that Iowa's zone defense has been more effective than the man-to-man defense and advocated that Iowa should run more zone.  However, in the same analysis I noticed that the zone becomes far less effective after it has been heavily used in a game.  In particular, prior to the Indiana game, Iowa's zone was allowing 0.83 PPP on the first 30 possessions in which it was used and 1.17 PPP after that.  But this was based on a very limited sample size since Iowa has only used zone on more than 30 possessions a few times this season.  I figured it was probably worth running more zone to see if it could be effective for more than 30 possessions. 

Apparently, Fran felt the same way.  Against Indiana, Iowa ran 39 possessions of zone defense -- and allowed 0.9 PPP on the first 30 zone possessions and 1.8 PPP on last nine zone possessions.  The calamitous drop-off in the effectiveness of the zone defense was a major factor in that nightmarish second half against the Hoosiers.

 

CJ Fredrick the Great, Emperor of Plussia

I record the lineup combination for every possession (excluding some garbage time) in every Iowa men's basketball game.  I use that data to compute how many points Iowa scores or allows on a per possession basis (O PPP and D PPP, respectively) when each player is in the game.  O Diff is how much Iowa's O PPP changes when that player is in the game.  Total Diff is computed as O Diff minus D Diff.  In essence, it's how much better (or worse) Iowa does in terms of overall scoring margin when that player is in the game versus on the bench.  These are a type of "plus/minus" data and since Iowa has now played 11 games against high major opponents, we have a decent sample size. Here's how Iowa has performed when each of the nine major rotation players has been on the court.

Player efficiency data for first 11 games against high major opponents
           
Player D PPP O PPP D Diff O Diff Total Diff
Total 1.06 1.21 NA NA NA
Nunge 1.01 1.13 -0.07 -0.14 -0.06
Garza 1.07 1.22 0.05 0.07 0.02
Bohannon 1.10 1.24 0.16 0.13 -0.02
Fredrick 1.05 1.26 -0.03 0.16 0.19
Wieskamp 1.10 1.19 0.16 -0.08 -0.24
C. McCaffery 1.08 1.25 0.05 0.09 0.04
P. McCaffery 1.00 1.10 -0.08 -0.15 -0.07
Keegan Murray 1.00 1.16 -0.09 -0.07 0.02
Toussaint 0.93 1.16 -0.17 -0.07 0.10

There are a number of things that really pop out here but the most glaring one is that Iowa is a staggering 0.19 points per possession better when Fredrick is in the game.  Over a 71 possession game, that comes out to a 13.5 point difference.  It's almost entirely due to the offense being WAAAAAAYYYYYY better when CJ is on the floor.  The offense is also dramatically better when Bohannon is in the game; however, the defense gets dramatically worse which effectively negates the offensive improvement. 

As we all know, Fredrick missed the second half of the Indiana game, which was also the worst half of basketball that Iowa has played all year.  So maybe CJ's +/- is being inflated by him missing out on that massacre (although maybe Iowa struggled because CJ wasn't out there).  In any event, if we omit the Indiana game altogether, Fredrick still has very good +/- numbers (second best Total Diff on the team) as shown here:

Player efficiency data for first 10 games against high major opponents (No Indiana)
Player D PPP O PPP D Diff O Diff Total Diff
Total 1.05 1.22 NA NA NA
Nunge 1.01 1.16 -0.07 -0.12 -0.04
Garza 1.06 1.24 0.06 0.09 0.03
Bohannon 1.09 1.26 0.15 0.13 -0.02
Fredrick 1.05 1.25 0.00 0.08 0.08
Wieskamp 1.09 1.20 0.16 -0.10 -0.26
C. McCaffery 1.07 1.26 0.04 0.08 0.03
P. McCaffery 0.99 1.13 -0.09 -0.13 -0.04
Keegan Murray 0.98 1.19 -0.11 -0.04 0.07
Joe Toussaint 0.93 1.18 -0.17 -0.07 0.10

Fredrick has tended to run hot/cold in terms of his production.  Some games he's been blisteringly hot and he tends shoots a lot in those games.  In the other games, he doesn't shoot much.  For example, over the last five games he's only averaging about 5 PPG.  However, the offense has still been about 0.06 PPP better when he is on the court.  This makes sense, as the other team has to fear his shooting ability and that, combined with his constant movement, forces the opposing defense to pay a lot of attention to him which opens up easier scoring opportunities for his teammates.  

The player with the second best +/- data after Fredrick is actually Joe Toussaint.  Our defense is 0.17 points per possession better when Toussaint is in the game.  The offense is 0.07 points per possession worse but that defensive improvement more than makes up for it.  I hope Fran finds a way to get Toussaint on the floor more going forward (more on that in a bit).

It has not escaped my notice that Wieskamp's +/- numbers are by far the worst on the team.  Frankly, I don't know what to make of that.  He hasn't improved as much from his freshman year as I hoped he would, but he doesn't look anything close to terrible to my eye.  

Lineup combinations and the battle of the bigs

It was pretty surprising to me that Garza's presence on the floor does not have a major impact on our offense: an improvement of only 0.07 PPP, whereas Fredrick and Bohannon have 0.16 PPP and 0.13 PPP, respectively.  I noticed this earlier in the season and attributed it to offensive struggles when Nunge and Garza are on the floor together.  The double big man lineups are still struggling offensively (about 0.09 PPP worse than our other lineups), which is a major reason why Garza's O Diff isn't much larger.  If we specifically focus on the possessions where Garza is at center and someone other than Nunge is at power forward, our offense scores 1.26 PPP - an improvement of 0.12 PPP over our other lineups.  If we apply the same analysis to lineups where Nunge is at center, we find that those lineups are 0.11 PPP worse on offense but 0.08 PPP better defensively.  So the defense appears to be considerably more effective when Nunge is at center but the offensive drop off between him and Garza (~0.12 PPP) more than negates this.  It is also worth noting that Joe Toussaint has been at PG for 59% of Nunge's center possessions but only 16% of the possessions where Garza has been at center without Nunge.  

The efficiency of some lineup combinations
Lineup D PPP O PPP D Diff O Diff Total Diff
Nunge and Garza 1.03 1.13 -0.03 -0.09 -0.06
Garza w/o Nunge 1.08 1.26 0.06 0.12 0.06
Nunge w/o Garza 0.99 1.12 -0.08 -0.11 -0.03
Toussaint and Bohannon 1.03 1.49 -0.03 0.3 0.33

Speaking of Toussaint, Iowa is scoring a preposterous 1.49 PPP when Toussaint and Bohannon are in the backcourt together.  The catch is that those two have only played together for about 38 possessions, total, throughout all 11 games against high major opponents.  So the sample size is minuscule.  It is still pretty striking, though, and I'm surprised Fran hasn't utilized that player combination more frequently.  The effectiveness of Fredrick is likely a major reason why we don't see more of Bohannon at SG.  However, if Fredrick has to miss time due to injury, the Toussaint/Bohannon backcourt combination could help soften the blow of his absence.

8 Comments
View 8 Comments