Georgia Tech football nosedived in 2015. Was it a blip or the beginning of the end?

Check out the advanced-stats glossary here. Below, a unique review of last year’s team, a unit-by-unit breakdown of this year’s roster, the full 2016 schedule with win projections for each game, and more.

In 2014, we saw what happens when it all goes right. Georgia Tech had a quarterback come into his own, with a set of senior skill guys, behind a line that featured an All-American. The defense stayed mostly healthy, and Paul Johnson’s Yellow Jackets won 11 games for the first time since 2009 and finished in the AP top 10 for the first time since 1998.

In 2015, we saw what happens when it all goes wrong. Georgia Tech headed into the year without its top two fullbacks, top five slotbacks, top two receivers, and said All-American up front. The Yellow Jackets then proceeded to get wrecked by injuries at fullback, slotback, offensive line, and defensive line. The defense didn’t regress much, but it didn’t matter. The quarterback fell into a bit of a funk, the big plays dried up, and Tech went 1-6 in games decided by one possession.

The result: a season that featured a likely step backwards instead featured about seven. Tech fell from 11-3 to 3-9, taking Johnson from king status to mildly hot seat.

Now what? We can see whatever we want to see from Johnson and Tech heading into 2016. If you want to say that this is the beginning of the end for Johnson, that the talent departed from the 2014 squad was of a higher caliber than its 2015 replacements, then you haven’t been proven wrong yet. Justin Thomas is a wonderful athlete and option quarterback, but your success is still dictated by the talent around you. If you choose to believe the defense is never going to improve enough to offset whatever talent drain has taken place on the other side of the ball, you might be right.

After four straight years in the Off. S&P+ top 40 and a peak of third in 2014, Tech fell to 62nd in 2015. Again, injuries could be attributed to some of that struggle, but we indeed do not immediately know that the new pieces will ever reach the heights of the old ones.

But if you want to point out that this downfall has happened before, you are also right to do so. Following the 11-win breakthrough of 2009, in which Tech ranked 16th in Off. S&P+, the Yellow Jackets fell to 74th in 2010. The new skill guys weren’t at the level of the old ones, and only an easier schedule kept the Ramblin’ Wreck bowl eligible. They then responded with said run of top-40 offenses.

I lean to the latter explanation, both because I’m an optimist by nature and because there’s a precedent. And because Tech was so banged up last year. Not a single fullback played in all 12 games. Leading returning slotback Broderick Snoddy missed four games. Slotback Qua Searcy was carrying a big early load, then missed the last nine. Fullback C.J. Leggett, a presumed future star, missed the entire season. Quarterback TaQuon Marshall had to fill in at slotback for a while. On top of that, only two linemen started in all 12 games.

The lineup was both limited and ever-changing. And after recording at least 40 rushes of 20-plus yards in four of the last five years, Tech managed just 27.

Meanwhile, only one defensive lineman could stay on the field for 12 games, which prevented the Yellow Jacket defense from improving enough to make up any of the difference.

Last year I thought Tech could remain at a top-10 level or something close to it despite the turnover in the backfield. That was clearly an opinion destined to be incorrect, but it took quite a bit for the Yellow Jackets to fall as far as they did. I would be surprised if they didn’t rise back into the top 40 this fall. How quickly we lower the bar...

Two weeks into the season, I was feeling as confident as ever in my Tech-as-top-team prediction. Against two thoroughly overmatched opponents, the Yellow Jackets looked like a well-oiled machine. Tech averaged 9.7 yards per play against Alcorn State and limited the Braves to 4.1 per play; against Tulane, those averages were 8.7 and 3.8. There is a predictive quality to blowouts that stark.

By this point, Tech had already lost linebacker Beau Hankins, defensive end Kenderius Whitehead, and backup quarterback Tim Byerly for the season. They were about to lose Searcy. The injury cavalcade had begun and would not stop. And the team we saw over the last 10 weeks ... well ... it wasn’t very good.

If you’re looking for reasons to be optimistic (or at least, not as pessimistic), you could certainly note that some of Tech’s better games during this 10-week slide happened against the better teams on the schedule. The Jackets knocked off Florida State with a stunning blocked field goal return and stayed within a touchdown of Notre Dame, North Carolina, and Georgia, all F/+ top 30 teams.

At the very least, this helps with the idea that pure talent wasn’t Tech’s problem. If talent were the issue, you’d assume Tech would lose tight games to mediocre teams and get blown out by good ones.

The Tech offense regressed across the board in 2015, but the regression was strongest in the play-maker/explosiveness categories. The Jackets fell from third to 27th in Success Rate+ and from first to 26th in Standard Downs S&P+; from an efficiency and chains-management perspective, they went from elite to merely good. But they plummeted from third to 73rd in IsoPPP (which measures the magnitude of the successful plays) and from fourth to 105th in Passing Downs S&P+. Twenty-yard gains turned into 10-yard gains, and once behind schedule, Tech couldn’t catch up.

Justin Thomas was asked to pass a bit more frequently, in part because Tech was dealing with more deficits; Tech’s run rates fell from 84 percent to 82 on standard downs and from 57 percent to 49 percent on passing downs. This was still clearly an option team, but Thomas went from averaging 8.4 yards per pass attempt, with a 51 percent completion rate and a 3.2 percent INT rate, to averaging 6.5 per attempt with 42 percent and 4.4 percent, respectively. All the averages went in the wrong direction.

Johnson’s spread option is brilliant at creating an offense that overachieves compared to its recruiting rankings. It is unique and requires a level and type of defensive discipline that opponents don’t have to deal with that much. but in 2015, we saw what can happen if a little bit of sand gets in the machine. Tech creates big plays, in part, because of the ruthless efficiency it generates. When the efficiency gets trimmed a little, the big plays get trimmed a lot. There isn’t a ton of margin for error here.

Note: players in bold below are 2016 returnees. Players in italics are questionable with injury/suspension.

On paper, the Tech attack looks fine. Marcus Marshall could become a unique big-play threat from the fullback position, Leggett is healthy, and quite a bit is expected of redshirt freshman Quaide Weimerskirch if he can also stay healthy. And in the slotback spots, Georgia transfer J.J. Green joins Clinton Lynch, Searcy, senior Lynn Griffin, etc. Plus, freshman Xavier Gantt was one of the more well-regarded signees in the 2016 class.

Meanwhile, at receiver, the trio of Jeune, Micheal Summers, and Brad Stewart (combined: 8.2 yards per target) was unable to replicate the 2014 success of DeAndre Smelter and Darren Waller (10.6). Jeune and Stewart are back, however, and big youngsters like Harland Howell (sophomore) and Christian Philpott (redshirt freshman) could threaten for playing time. I expect more out of this unit considering the higher level of experience and, in theory, the higher level of efficiency (which would force fewer passing downs).

All-American guard Shaq Mason was gone, off starting for the New England Patriots. Presumptive starting tackle Chris Griffin missed the entire season with an ACL injury. Center Freddie Burden and guard Trey Braun were the only two linemen who could stay on the field without limping off.

It made sense that Tech’s offensive line would struggle in 2015, and it did. The Yellow Jackets fell from first to 81st in stuff rate and from 34th to 78th in Adj. Sack Rate. Opponents spent a lot more time than usual in the Tech backfield, and nothing in the world can disrupt an option offense more than sudden losses.

The line isn’t guaranteed to improve. Braun and two others are gone, and at the moment, Griffin’s status is uncertain and Shamire Devine’s size is an issue. If they both end up on the field, Tech will have six players with starting experience at their disposal. Without them, that number drops to four, and the pool of potential replacements isn’t particularly deep. The line doesn’t have to be great for Tech’s offense to thrive, but it still needs to be decent.

To pull off a bend-don’t-break defense, you still have to prove you can make plays. The idea is that you wait for an opponent mistake, and then you pounce. That means you need to be good on passing downs -- once they have fallen behind schedule, they can’t be allowed to catch up -- and it means you must be good at stiffening as you get closer to the end zone. It probably also goes without saying that you can’t give up big plays.

Georgia Tech had the “bend” part down pat last year. The Yellow Jackets ranked 78th in Success Rate+, a major improvement over 2014 (111th) but still not very good. They also ranked 29th in Passing Downs Success Rate+, which was awfully impressive considering the Tech pass rush was more theoretical than literal.

Still, there were issues. The successful plays Tech allowed on passing downs were quite successful -- opponents completed just 47 percent of their passes on third-and-4 or more but averaged 16 yards per completion -- and the Jackets ranked just 93rd in Redzone S&P+.

Tech also stunk at defending the run, but when you look at the injuries up front, that one makes a little bit of sense.

The Jackets needed more from their defense than they got in 2015, which has pretty much been the story of the Johnson era. And now Ted Roof’s defense reinforces its front seven just in time to start over on the secondary.

The ceiling of Tech’s defense isn’t particularly impressive, but we can at least understand why the Jackets couldn’t defend the run or rush the passer very well in 2015.

The defensive line was hit early and often by injury, which left the one healthy guy, end KeShun Freeman, tired and outmanned. His production fell from 9.5 tackles for loss to 4, but of course it did. Tech had 10 linemen who averaged at least 0.5 tackles per game played -- five tackles, five ends -- and Freeman was the only constant. The five tackles combined to miss 19 games, and there was no such thing as continuity.

Taking this into account, the fact that Tech improved from 94th in Rushing S&P+ to 83rd was semi-impressive. The Jackets should NEVER rank that low, but the bar was awfully low after 2014.

So they get a half-mulligan. Great. It expires this fall. Of last year’s 10 “regulars,” seven are back, including Freeman, tackle Francis Kallon (a one-time four-star recruit who has one last chance to live up to that rating), and sophomore Anree Saint-Amour (a nearly four-star recruit who made a couple of plays). Plus, JUCO transfer Desmond Branch and four-star freshman Jordan Woods could squeeze into the rotation. There are a lot of candidates, at least, which helps when you’re looking for four particularly decent linemen. But there are no guarantees.

The front four suddenly has quite a few options, but the back seven has been thinned out quite a bit. Linebackers P.J. Davis and Brant Mitchell are decent, but they’re they only reasonably proven pieces there, and now four of the top five defensive backs are gone.

The safety position appears to be in decent shape from an experience standpoint. Juniors Lawrence Austin and Corey Griffin and sophomore A.J. Gray all saw quite a bit of action last year, though they combined for just 1.5 tackles for loss and four passes defensed. With the top three cornerbacks gone, however, that position is a mystery. Juniors Step Durham and Lance Austin and USC transfer Lamont Simmons all seem to have decent upside, but they barely saw the field last year.

This isn’t a complete reset. Tech probably isn’t going to be starting three freshmen in the secondary by any means -- it will be mostly sophomores and juniors. Still, the turnover here is obvious, and the pass rush will need to improve quite a bit to help the secondary out. That isn’t the safest bet in the world.

Assuming the offense is solid again, having a good kickoffs weapon in Harrison Butker will help more than it did last year. Butker is also a strong place-kicker; he hit four of seven longer kicks last year and missed no PATs.

Butker checks two boxes, and there’s enough speed in the skill corps to figure that decent return options emerge to replace Jamal Golden. But punting is a question mark; Ryan Rodwell averaged barely 39 yards per punt, and Tech allowed a mediocre 7.8 yards per punt return. That’s a bad combo.

The 2010 parallels should be comforting if you’re a Georgia Tech fan. Tech double dipped last year, losing a ton of the previous year’s contributors on offense, then losing a ton more to injury. The Yellow Jackets were too young to pull off any level of offensive consistency. That likely won’t be the case in 2016.

With the offense likely heading back toward the Off. S&P+ top 30, that will keep Tech in games and likely punch a bowl ticket (S&P+ disagrees a bit with that, but S&P+ doesn’t take injuries into account very well).

The defense, however, gets none of the faith I feel I can put in the offense. Tech has only once ranked better than 50th in Def. S&P+ since 2008, and with a rebuilt secondary I can’t figure this is the year that changes.

Rinse, repeat. Georgia Tech should have a good offense and a defense that can’t make enough plays. We’ve seen this episode before. But hey, that’s still better than last year’s episode.