the deck gets stacked against the coach/team whenever you start measuring records vs teams that finished up ranked in the top 10 or top 25.
For example, this year USC and Stanford had identical 7-4 records in games other than ND. But in this analysis, ND gets punished for the USC loss but does not get credit for the win vs Stanford. Even though Stanford would almost certainly would have re-entered the polls at 8-4 had they beaten us, just as USC is ranked now. Similarly, MSU 2011 and Oklahoma 2012 would have been losses vs Top 10 teams, but instead are credited as wins vs 11-25 teams. I'm not saying this to defend Kelly's record (which I think is adequately judged already based on 5,5,1,4,5), I'm merely pointing out an inherent flaw in the analysis.
There's nothing wrong with evaluating this type of stat, just as long as the evaluator understands that the baseline isn't .500 as it would normally be in a game where there is a winner and a loser.
I don't really know what a good baseline win % is for this stat, unless we compare it to other coaches. I think it's a waste of time to compare it to the win % that guys like Holtz/Saban/Meyer have vs teams that finished ranked in the top 10/25, because we already know that Kelly isn't in their league. I would be more interested in comparing this to other coaches who would be presumable future candidates for the ND job whom you believe would be capable of winning a national championship at ND.
FWIW, your list omitted the 2013 win vs USC.
Regardless, thank you for taking the time to compile this.