I don't want this to feel like I'm piling on or beating a dead horse, but I wanted to fact check Ron Everhart's claim that the turnover we saw at Duquesne was a college basketball problem not a Duquesne problem. My immediate thought was "this guy is sooooo full of crap." I decided to put him to the test ala Politifact.com, a favorite web site of mine, and rate his comment on their scale.
In an article in the Post Gazette on Wednesday addressing the reaction to TJ McConnell, Mike Talley and Danny Herrera electing to leave the school, Everhart said the following:
"We haven't won (at Duquesne) in 40 years and have had five winning seasons. I'm very proud of that. It's not where I'd like to be, but quite frankly, we're not any different than any school in our league who go through cycles," he said. "Everybody wants to make a big deal about guys leaving the program. And we have a couple things that work against us, but it's also an epidemic in college basketball.
"It's not a Duquesne University men's basketball issue. It's a college basketball issue."
Ron's claim should be broken down into two parts. The first is that turnover is a problem in college basketball and second is that Duquesne is just another victim of some national trend. To evaluate the claim, I researched roster changes for 24 teams to see how many players did not complete their senior season from 2007-2008 to the beginning of last year. I have not included any transfers that have happened since the most recent season ended. 10 of the 24 teams came from the Atlantic 10 and 14 teams from two randomly chosen conferences, the Big 10 and the Colonial. I included both walk ons and scholarship players as part of my analysis. I did not take into account the reasons why players left the team such as grades, illness, injury or entering the NBA draft.
We'll begin by looking at the claim that turnover is a universal problem in college basketball. The schools I looked at averaged 8.92 players leaving early over the full four years or 2.23 players per year. This is not an insignificant number. Assuming a roster with 13 scholarship players and 2 walk ons, the average team loses 15% of its roster early every year for whatever reason. Results varied fairly widely from school to school. One, Wisconsin, saw a single player leave early. Two saw 16 fail to finish their careers, which leads me into my next question.
In terms of raw players lost, Duquesne tied with Fordham for the most turnover in the survey. It's interesting to note because Fordham saw a coaching change during the analyzed period while Duquesne did not. Really, having the most players in the sample is meaningless since I only looked at 24 teams. How far outside of the norm is Duquesne when the sample is generalized to all of the Division I basketball? The standard deviation for the test schools was 4.32 transfers over the four offseasons placing Duquesne 1.6 standard deviations above the average. Assuming a normal distribution for the entire population or a neat bell curve, there are some schools that would have higher turnover than the Dukes but not a lot of them. The Dukes would be in the 94th percentile for turnover in the country. To put this into more familiar terms, if Ron Everhart performed as well on the SAT's as he did in turnover, his score would be a 2080 (1400 old composite). Not exceptional, but good enough to get you into Cornell or a nice fat scholarship at Duquesne.
My Rating: Ron Everhart made two claims. The first was that turnover is a problem every where. On average, the total number of players transferring per team was higher than I would have expected. There were exceptions that could keep their rosters stable but Wisconsin was an outlier even further to the left on the distribution than Duquesne was on the right. Evidence suggests that the first part of Everhart's statement is true. However, the problem is far worse at Duquesne than the rest of the country and if the sample is representative, it means only 6% of all schools have higher turnover rates. It seems Everhart was correct with one assertion but not with the other. Therefore, I would rate his overall claim as half true.
.
