Tracking Efficiencies and Lineup Combinations Suggests That Nicholas Baer Was Really Good

By houksyndrome on April 30, 2019 at 9:00 am
go baer go
© Kevin Jairaj-USA TODAY Sports

I recently discovered that, for some reason, Iowa’s starting five was extremely ineffective during the last third of the season.  I wrote an article about it previously so I won’t go into too much detail here but will use it as a jumping off point for this article.  In short, from the Minnesota game through the end of the regular season, Iowa’s starting five was outscored by about 0.17 points per possession (which is terrible), while their bench was only outscored by about 0.03 points per possession.  So the play of our non-starting five lineup combinations was the only reason why that awful stretch to end the season wasn’t even worse.  The disparity between starters and non-starting five lineup combinations was entirely on the offensive end.  The starters were actually slightly more effective defensively than the non-starting five; however, the non-starting five lineup combinations were about 0.18 points per possession better than the starters on offense.

In one of the comments to that article, therealCatnuts suggested that the difference in effectiveness between these lineup combinations was largely, if not entirely, due to Nicholas Baer.  Another hypothesis is that the starting five was generally going up against the other team’s best players, when they were fresh, while our bench was going up against the other team’s weaker players / tired starters.  These hypotheses are not mutually exclusive and it could be a combination of both of them.  It is impossible to know exactly how much of the bench vs. starters difference is due to the other team’s lineup and how much is due to us, because I didn’t record the other teams’ lineups.  However, I did find some pretty striking evidence that Baer was critical to our success, which I will share with you in the following paragraphs.  Based on this evidence, I think the success of our non-starting lineup combinations was largely, perhaps entirely, due to Baer.  Congratulations, therealCatnuts, you win a fabulous prize!*

*due to budget cuts, GoIowaAwesome is no longer awarding prizes.

First I will briefly review my methods.  I recorded Iowa’s player combinations, defense type (we won’t go into that in this article) and number of points scored or allowed for pretty much every possession, excluding some garbage time, from the Minnesota game until the end of the year.  I included the Illinois BTT game and the NCAA tournament games against Cincy and Tennessee.  I didn’t want to watch the BTT game against Michigan again, so I didn’t track possessions from it.  So now I have 14 of the 35 games that the team played this year (i.e. 40%) with a total of about 950 possessions.

Let’s start with some simple metrics, points per possession scored (O PPP) and points per possession allowed (D PPP) when each of our players were on the floor (I am omitting the non-scholarship guys because of small sample sizes).  These values are shown in table 1 along with +/-, which is equal to O PPP minus D PPP.

Offensive and Defensive Efficiencies for Each Player
Wieskamp 1.09 1.10 -0.00
Dailey 1.05 1.12 -0.06
Baer 1.13 1.08 0.05
Cook 1.07 1.09 -0.03
Moss 1.06 1.12 -0.06
Garza 1.12 1.12 0.00
McCaffery 1.11 1.12 -0.01
Bohannon 1.08 1.09 -0.02
Kriener 1.05 1.05 -0.00
Starting 5 1.02 1.08 -0.07
Non-Starting 5 1.11 1.10 0.01
Total 1.09 1.10 -0.01

Another way of looking at this is by measuring how much better our offensive (or defensive) efficiency gets when a player is on the court.  I have shown these data in table 2.  O diff is simply OPPP when a player is on the floor subtracted by OPPP when that player is on the bench.  The same is true for D diff.  A positive O diff and a negative D diff suggests that a player is better than whoever is replacing him on offense and defense, respectively.

How did our effectiveness change when each player is on the court?
Player O Diff D Diff Net Diff
Wieskamp 0.03 0.00 0.03
Dailey -0.05 0.02 -0.07
Baer 0.08 -0.03 0.12
Cook -0.08 -0.02 -0.05
Moss -0.07 0.06 -0.12
Garza 0.08 0.05 0.03
McCaffery 0.05 0.04 0.01
Bohannon -0.04 -0.01 -0.03
Kriener -0.05 -0.07 0.02
Starting 5 -0.10 -0.02 -0.07
Non-Starting 5 0.10 0.02 0.07

One thing jumps out right away:  Baer is one of only two Hawkeyes with a positive +/-, with the other being Garza, whose +/- was only slightly greater than zero.  Overall, we were 0.12 PPP better when Baer was on the court, about 0.08 PPP better on offense and 0.03 PPP better on defense.  At face value, these data could be interpreted to mean that Baer was actually our best player last season.

It is possible that some (or most) of Baer’s effectiveness, as evidenced by his +/-, was due to him playing against weaker or tired opposition.  I don’t think that was the case, however, because if you look specifically at all of our “non-starting five” lineup combinations and compare the ones with Baer vs. the ones without Baer, you still see a dramatic difference in effectiveness when Baer is in the game.  The data are in table 3.  I would intuitively expect that our “with Baer, non-starting five” and “without Baer, non-starting five” lineup combinations would face about the same quality of opposition, in aggregate.  The fact that our “non-starting five” lineup combinations are 0.05 PPP better on offense and 0.07 PPP better on defense when Baer is in the game suggests to me that Baer’s +/- is largely, if not entirely, due to Baer and not some hidden variable like opponent lineup composition.  When our starting five were on the floor together, they were outscored by 0.065 PPP.  Our “non-starting five” lineup combinations that did not include Baer were outscored by 0.067 PPP.

Our non-starting lineups were ineffective when Baer was not in the game
  With Baer With Baer Without Baer Without Baer      
  OPPP DPPP OPPP DPPP O Diff D Diff Net Diff
Non-Starting 5 1.13 1.08 1.08 1.15 0.05 -0.07 0.12

Another possible hidden variable would be if Baer is only on the floor with specific teammates and one of those teammates, or one particular lineup combination, is skewing the data.  That doesn’t appear to be the case, though.  I identified 47 different lineup combinations that contained Nicholas Baer across his 426 possessions.  Of these, the Bohannon, Moss, Wieskamp, Baer and Cook combination was by far the most prevalent (49 possessions total) and that lineup combination put up 1.12 OPPP and surrendered 1.08 DPPP, which is nearly identical to his overall averages.  His remaining 377 possessions were spread out over 46 different lineup combinations so the idea that there was one particular lineup combination that was driving his +/- does not seem to hold water.  Furthermore, with the exception of Moss, every player on the team’s +/- improved when they were on the court with Baer, sometimes by as much as 0.24 PPP.  We can rule out the possibility that Moss was secretly responsible for Baer’s on court success because when Baer and Moss were on the court together, we actually got outscored by 0.12 PPP.  In summary, Baer was making his teammates better, not the other way around.

We all know that Baer has a great story and seems like a great guy.  I think we all recognize that he’s a pretty good player.  Some of his advanced statistics, such as BPM (an estimate of +/- based on box score stats), suggest that he was actually our best player this season.  Frankly, I was very skeptical of Baer’s BPM when Catnuts cited it in his previous comment because I am not sure how precise of an estimate it is.  However, in this case, it appears to have been right on the nose.  Nicholas Baer is a great player and he should be recognized for the extremely high quality of his play, not just his effort level and determination to make himself into a B10 player as a former walk-on.  Until I did this analysis, I didn’t realize just how much of an impact he made for us.  At face value, these data suggest that he will be missed even more than Cook, who I would have automatically assumed was our best player.  We are really going to miss him next year.

So if you’ve been debating about whether or not to get that Nicholas Baer face tattoo, go ahead and get it.  If you are having a baby soon and don’t know what to name it, name it “NicholasBaer”.  One word.  You could name it “Nicholas Baer”.  But that’s a cop out.  People usually don’t write down their middle names and the Baer needs to be recognized.  Otherwise people might think you named your kid after some other Nicholas, like Tsar Nicholas II.  Do you want your kid to be mistaken for some Russian dude who got chopped up by the Bolsheviks?  Hell no.  It has to be NicholasBaer or you aren’t a true fan.  Now, what if it’s a girl?  I thought about that.  You should name her NicholasBaer**.  You could call her by a nickname, like “Nicole”, if you wanted, or you could call her “NicholasBaer”.  People might give you (and her) crap if you did that but those people would be losers who don’t understand the game of basketball.  Name your kid NicholasBaer.  Do it right now!


**Actually, just name her MeganGustafson.  It’ll be way less trouble.

View 17 Comments