First, the bar should be 10, not 9. In any event, Kelly averages 9 wins now only because of 2012. A majority of the time (3 out of 5 years), he's gone 8-5, and he needed an upset win yesterday to pull it off this year. That obviously isn't good enough.
And while 10-3 should be the average, for Kelly to reach that by 2019 he's going to need to average 11-2 the next 5 years. At this point I probably could be talked into settling for a 10-3 average the next 5 years, which would bring Kelly to 95-35 (.731).
But even to average 10-3, he'd need some seasons where he does better than that -- just as he badly needed that 12-1 season just to get to a 9-4 average to this point. 2015, in particular, strikes me as a season where we should win no fewer than 11 games, including the bowl(s). A 10-3 mark next year, with all the advantages we figure to have in personnel and scheduling, would be a disappointment and a bad sign for the future -- although I'm sure ND would be delighted with it and start printing "10 WINS!" t-shirts.