The Fulfillingness Of Fran's Finales: A Statistical Analysis Of Iowa MBB Over The Course Of A Season

By houksyndrome on February 3, 2020 at 1:08 pm
we talkin' fades?
© Jeffrey Becker-USA TODAY Sports

Let’s play a little guessing game.

Some background information first.  Bart Torvik has a metric called “Game Score” which estimates how well you played in a particular game, using efficiencies and adjusting for opponent quality.  This is logical; a one point loss to that Anthony Davis-led Kentucky team is a more impressive showing than a one point win over Chicago State (or the Chicago Bulls, for that matter) and it would receive a higher game score as a result.  Game Score is a single number that varies from 0-100 (100 being good, 0 being terrible). Iowa's win over Maryland was a 99.  Iowa's loss to Maryland was a 79. The loss to DePaul was an 11. For the purposes of my guessing game, I need a single number that reflects Iowa's quality of play in individual games.  Throughout this article when I say “Game Score” or “game score,” I mean this metric and not the actual number of points that Iowa scored or allowed in a particular game.

Okay, with the background information covered, let's play the game.  Here is a list of our monthly average Game Score over the entire Fran era.  Can you guess which one is which?

Month A = 71.8
Month B = 80.1
Month C = 74.9
Month D =  77.3
Month E = 76.3

Surely Month A is February or March, right?  I mean, we all know that Fran’s teams fade at the end of the year.  Month A is actually November.  The months are listed in order:  November, December, January, February, and March.

Perhaps we are grouping things too broadly here and, if we examined it game by game, a clear trend would emerge.  Every Iowa team under Fran has played at least 31 games (30 regular season + 1 BTT game).  In the following graph, I show the mean Game Score for each of Iowa’s first 31 games, averaged over the entire Fran era.  Error bars indicate standard error of the mean (SEM).

Average Game Scores over the course of the season in the Fran Era

Here are 5 game running averages of the same data:

Game Score, 5 game running average


Do you see a clear trend there?  An obvious drop off at the end of the season?  I don’t.  Just for reference, conference season typically begins in earnest around game 13.  What about winning percentage?  Here is Fran’s winning percentage, month by month:

November = 64%
December = 74%
January = 47%
February = 51%
March = 45%

Obviously, his winning percentage is higher in November-December than in January-March but that’s because Iowa's non-conference schedule is easier than their conference schedule.  If we want to compare apples to apples, we have to look at January-March, which are almost entirely composed of conference games.  Fran’s overall conference winning percentage, coming into this season, was 48%.  There isn’t an egregious drop-off in winning percentage in any month.  Maybe the 45% winning percentage in March could be a slight drop-off but that could just be noise, too.

I did notice some interesting trends in our offensive and defensive efficiencies, which I’ve shown in the graphs below.  Here I’m showing 5 game running averages of our average adjusted offensive and defensive over the course of the season (Note: the Y axes don't start at zero):


Iowa's adjusted offensive efficiency over the course of the season


Iowa's adjusted defensive efficiency over the course of the season during the Fran Era

One thing you should be aware of is that collegiate offenses, as a whole, get more efficient as the season progresses so that upward trend that we see in both Iowa's offensive and defensive efficiencies is common for college basketball teams.  What’s peculiar is what happens in the last few data points (representing the last 5-8 games).  Iowa's adjusted offensive efficiency falls off by close to 0.05 PPP while their defensive efficiency improves by about 0.03 PPP.  These effects more or less cancel out and Iowa's game scores stay constant over time.  I don't know what to make of those changes, frankly, but thought I'd share them anyway.  Kurt Cobain, your thoughts?

Okay.  Thanks, Kurt!

It’s hard to say whether these late season swoons are a pervasive problem for Fran’s teams.  In aggregate, there is no drop off over the last several games of the season.  However, without a doubt, there have been three seasons that have ended on very sour notes:  2013-2014, 2015-2016 and last year, 2018-2019.  That said, all teams, except for the truly great ones, have at least some periods where they struggle.  Perhaps the fact that 2013-14, 2015-16 and 2018-19 all hit rough patches at the end of the season was merely random chance in terms of timing.  I don’t know.  However, before leaving this subject, I thought it would be interesting to evaluate the struggles in those seasons to look for commonalities.

The 2013-14 team lost 7 of their last 8 games.  Prior to that time, they were 19-6.  Their average Game Score during the skid was 77, whereas before the skid it was 91.  So they were playing worse, it wasn’t simply a matter of them encountering better competition.  Pre-skid, they had adjusted offensive and defensive efficiencies of 119.5 and 91.9, respectively.  During their last 8 games, the adjusted offensive efficiency stayed good at 121.7 but the adjusted defensive efficiency collapsed, ballooning up to 108.7.  Collapsening I was pretty much entirely a defense problem.

The 2015-16 team went 3-7 in their last 10 games with average Game Scores of 68.  Prior to that time, they were 19-4 with average Game Scores of 86.  As in 2013-14, they were indeed playing worse during their last ten games.  After game #23, they had adjusted offensive and defensive efficiencies of 117.1 and 92.0, respectively.  During the last 10 games, the adjusted offensive efficiency fell to 110.5, while the defense got worse (101.54).  So it was both offense and defense that were losing steam.

Last year, 2018-19, Iowa went 2-6 in their final 8 games, after starting 21-6.  After game #27, Iowa had adjusted offensive and defensive efficiencies of 118.3 and 99.2, respectively.  During the skid, the defense didn’t change that much, increasing to 102.7, but the offense fell to 111.1.  Iowa was a pretty mediocre defensive team all season last year and, when the offense stopped being elite, they ran into problems as evidenced by Iowa's average Game Score being 60 over the last 8 games and 82 over the 27 games prior.

The bottom line is that there isn’t much commonality between the three skids.  Collapsening I was the defense suddenly falling apart.  Collapsening II was both the offense and defense struggling.  Collapsening III was the offense hitting a slump.  Given the lack of commonality, and the lack of an overall downward trend in our Game Scores over Fran’s entire tenure, I have a hard time buying the notion that Fran’s coaching somehow makes the team prone to these collapses.

View 29 Comments