Bill Connealy of SBnation argues that he has a better way of measuring returning talent. Rather than just looking at returning starters, he has correlated various stats with the following season production. Some returning starters have a high correlation, particularly receivers on offense, while some have a low correlation, such as offensive line returning.

https://www.sbnation.com/college-football/2018/1/31/16950...

By his measure, returning talent for MAC teams are:
Ball State 84%
Kent 77%
BG 76%
Miami-Ohio 73%
Buffalo 73%
NIU 66%
EMU 64%
Ohio 62%
Akron 61%
Toledo 58%
WMU 57%
CMU 38%

Note that the three worst teams from last year have the most returning. That bodes well for the bottom not being quite so far behind the top as last year. Last year the combination of Ball St, Kent, and BG were 2-20 against MAC teams outside that group. Interestingly, the two wins outside that group were wins by BG and Kent over Miami, solid evidence yet again that Miami sucks.