If you're talking about the personnel "delta," guys that will be with the team this coming year that didn't play at all last year, you definitely have to include all the returned missionaries, regardless of whether or not they played before their missions.

Bill Connelly's preseason S&P+ rankings, for example, include a recruiting element that would absolutely need to count all of these additions, plus hypothetically any players who missed last year due to injury and will return in 2019, to properly predict the year-over-year changes in personnel.

Imagine an unrealistic scenario: you have ten 5-star players in the Class of 2016. All ten played right away and helped BYU win a national championship in 2016. Then, they all leave on missions and all ten return for the 2019 season. The "true" recruiting rankings would need to include them for both 2016 and 2019, since they were players that were added to the roster and didn't contribute to the team's ranking the previous year.

That's the point of the true signing class. It's all the players that are being added to the roster that will help replace the departing players and affect a team's ranking from the previous year for better or worse.