Uh… is there any?

I've been fretting about Michigan's future APR scores for a while now without actually looking at the numbers."Wah wah wah," I wah, "APR mumble bits mumble." There's a possibility I'm mildly concerning all of you for no reason, so there's no time like the present to put some numbers behind the concerns.

Varsity Blue has helpfully listed the full dossier of transfers and departures since 2008, when the APR scores leave off. (Actually, the Spartan Tailgate "Rodriguezed" thread is more complete. In retrospect, this is obvious. Warning: useful content goes to zero after post #1.) With this information we can divide and multiply our way to knowledge, like they did in the olden days.

Assumptions

I'm still a little unclear on how this thing is calculated. The NCAA's explanation:

The APR is calculated by allocating points for eligibility and retention -- the two factors that research identifies as the best indicators of graduation. Each player on a given roster earns a maximum of two points per term, one for being academically eligible and one for staying with the institution. A team's APR is the total points of a team's roster at a given time divided by the total points possible.

This seems insufficiently detailed. Do walk-ons put on scholarship count? Do walk-ons count in general? What happens when a player like Mallett transfers halfway through the year? What about early NFL draft departures? I've searched the NCAA's web site and haven't found explanations. (I do have my second attempt at an email in; the first went unanswered.)

Let's make some common sense assumptions in the meantime:

• Walk-ons count if they're on scholarship. It would be hard for the NCAA to distinguish between a recruited player and a walk-on who earned his way onto the team.
• They do not count if they are not on scholarship. Allowing any walk-on to count would allow teams to pack their rosters with 5'7" chemical engineers.
• A transferred player only hits you once.
• There are two semesters with 170 points available in each, for a yearly total of 340.

A sanity check on that last point: the NCAA has stated a 925 APR corresponds to about a 50- or 60-percent graduation rate and .925 to the eighth power is about 53%. It goes by terms.

Should You Be Mildly Concerned?

Under those assumptions, Michigan's yearly APR maximums for Rodriguez years one and two:

2008

Three transfers (VB missed Quintin Patilla), two medical scholarships—which don't count against M since both players remained in school— and one Marques Slocum 0-for-2 leave Michigan down a minimum of five points. Since the decisions of Ciulla, Mitchell, DeBenedictis, and Gallimore—AKA the entire 2004 offensive line recruiting class—to leave early were not accompanied by transfers I don't think they'd count against Michigan unless any of those guys left school without a degree.

Then there are the departures of Adrian Arrington and Mario Manningham for the NFL. I'm not sure what the NCAA does in the case of early entries. A review of the Greg Oden stuff at Ohio State is inconclusive:

Oden was classified an "0 for 2" in APR jargon, meaning he left school without completing the term and was ineligible for the following season when he left. An "0 for 2," combined with a program's overall APR score of less than 925 (about a 60 percent graduation rate), triggers such a penalty. …

The early departures of Mike Conley Jr. and Daequan Cook are not believed to have negatively affected the score because they completed spring quarter in 2007 before leaving.

Wait, what? Oden gets 0 for 2 for leaving for the NBA but the other two guys don't affect the score even a little? That's odd. I'd guess at least one of the receivers left ineligible; if both did that would cost Michigan another four points.

In the worst case, 331/340 = 973 APR. Even the improbable double-worst case where every player who left Michigan did so with a 0.0 GPA, the attrition only gets Michigan down to 958.

Those numbers immediately makes me think this calculation is goofy even if you factor in some unpublicized dings from players who exhaust their eligibility but don't graduate. But the sanity check is the sanity check. If they were grading this out of 170 points per year the graduation rate they are shooting for would be 75%.

2009

Nine transfers—Hill, Clemons, Babb, Chambers, Horn, McGuffie, Threet, Wermers, and O'Neill—have shoved off. Wermers left ineligible, so Michigan is down at least ten points. Others could have left ineligible, too; we wouldn't know because they'd have to sit out as they got their grades up due to the transfer anyway. I'm willing to bet many dollars that Carson Butler left ineligible, which would make the number 12.

Andre Criswell left the team but remains at Michigan as a grad assistant, so he got a degree and won't count against them. Jason Kates left the team but may not have left school, in which case he wouldn't count. Taylor Hill is also an interesting case since he left after about two weeks and immediately transferred to Kent State. If he got out by the drop/add deadline there's a possibility he doesn't count either.

And then there's the strange case of Marcus Witherspoon, who apparently enrolled enough to invoke transfer rules when he moved to Rutgers, but managed to do so without being eligible, but managed to enroll and redshirt at Rutgers despite supposedly not being eligible, which I guess he wouldn't have to be except then he'd… well. I don't know.

There's a lot of gray area here. The bare minimum is 12 , which would be a 964 APR. The maximum reasonable loss—not everyone left ineligible—would be around 20, which would be a 941 APR.

Limitations

These numbers appear too optimistic. Michigan's rate of attrition under Carr seemed considerably lower than it does in the first couple years of the Rodriguez era, and those teams checked in with APRs near the worst-case scenario of Michigan's 2009 Transfer Spectacular. Either I'm calculating these wrong or there's a big unknown minus from players who run out of eligibility but leave without a degree. I lean towards the latter pending someone from the NCAA actually responding to an email.

Well, Should You?

If we can squint at the grim transfer parade of last year, make the maximum reasonable negative assumption, and then tack on another six points for non-graduating seniors—which would be over 50% of the class, well outside the Michigan norm—and still get a 923 APR, Michigan is going to be fine.

Even in this unlikely worst-case scenario, that one-year number is barely below the line and should be surrounded by years much higher than that. Rodriguez's attrition should drop considerably as the guys who didn't sign up for this Barwis nut leave the program via means natural and un-.

While Michigan's APR will continue to dip over the next couple years, it probably won't even approach the Mendoza line, let alone dip below it.