The fact they've now done it two years in a row does at least suggest it isn't ALL luck. A significant component? Of course. But they've found a style of play that works for them -- much like we would constantly let our opponents try and drive for game-winning drives during Grobe's heyday, and they would either screw up or our D would make a big play.
That style only had a three-year shelf life of working due to our personnel on D and the lack of talent on other squads during that time. We'll see if Duke can keep theirs up longer
They certainly have a good game plan. They've also beaten up on bad teams and play in a relatively weak division in the Coastal.
Over the past season and a half they've played 22 games and are 17-5 which is good for anybody (obviously). Here's how those wins break down and this uses Sagarin as the rating method since human polls are generally awful:
Record against teams ranked 1-25: 0-2 (FSU and TAMU last year)
Record against teams ranked 26-50: 4-2 (last year beat UNC, Miami, and VPI, this year beat GT. Lost to GT last year and Miami this year)
Record against teams ranked 51-100: 6-1
Record against teams ranked 101+: 7-0
So Duke is 13-1 against teams ranked worse than 51 and 4-4 against teams in the top 50. I can't imagine there's another Power 5 team who has played seven games against teams ranked worse than 100, much less won seven games.
Again, I'm not taking anything away from Duke but they're just beating up on bad teams. The best part of their overall resume is that they're 6-1 against teams between 51-100 which indicates that they've solidified themselves as a pretty solid top 50 team. The funniest part is they've only played 8 games against teams ranked in the top 50 and 2 games against top 25 teams. 2 out of 22 is seemingly an impossibly low rate for playing top 25 teams.
ETA: While the SEC West is an anomaly because of how difficult it is, if you want to see what Duke would do with a difficult schedule just look at Arkansas this year. Duke and Arkansas are one spot apart in total Sagarin and Arkansas is ten spots better in predictor - so they're generally equivalent and Arkansas may even be a little bit better (would be favored by around four on a neutral field). Duke is 7-1 with an inside track to play in the ACCCG and Arkansas is 4-5 and has to win two of three against LSU, Ole Miss, and Missouri just to make a bowl.