Projection predicts turnaround for Georgia Tech football

Georgia Tech head coach Geoff Collins instructs during the second half of an NCAA college football game at Bobby Dodd Stadium on Saturday, November 30, 2019. Georgia won 52-7 over the Georgia Tech. (Hyosub Shin / Hyosub.Shin@ajc.com)

Credit: HYOSUB SHIN / AJC

Credit: HYOSUB SHIN / AJC

Georgia Tech head coach Geoff Collins instructs during the second half of an NCAA college football game at Bobby Dodd Stadium on Saturday, November 30, 2019. Georgia won 52-7 over the Georgia Tech. (Hyosub Shin / Hyosub.Shin@ajc.com)

There’s a standard formula for preseason college football rankings. Look at what teams did last season, count the returning starters, consider the program’s history and make an educated guess. ESPN’s Bill Connelly goes deeper with his SP+ projections.

They include a returning production formula that takes key statistics and weighs them “based on what correlates most strongly with year-to-year improvement and regression.” SP+ also considers recruiting rankings, transfers and the program’s five-year history. That approach had good results for SP+ last season, based on its record against the spread in the early weeks.

That's encouraging for Georgia Tech's second season under coach Geoff Collins. The Yellow Jackets rank No. 58 in the 2020 preseason SP+ projection after finishing No. 111 in 2019. That rosy forecast is largely explained by Tech returning the second-most production among FBS teams, according to Connelly's weighted formula.

Tech’s jump of 53 spots in the SP+ rankings is the largest by far among 130 teams. The Jackets were 3-9 in 2019. They had two victories vs. Power 5 opponents and losses to The Citadel of FCS and Temple of the American Athletic Conference. SP+ projects the Jackets improving from one of the worst teams in FBS to a slightly above average outfit.

A predicted leap that large should be met with some skepticism. But I’m buying what SP+ is selling on Tech. It’s a sound statistical analysis tool with a solid track record. More than that, SP+ might be underestimating the Jackets based on what it doesn’t measure.

Last season, Tech transitioned from Paul Johnson's triple-option offense to a conventional spread. No one knew how that would go because there was no recent precedent. SP+ ranked the Jackets No. 78 before the 2019 season. Tech ended up much worse because the new-look offense cratered.

SP+ can’t calculate how much Tech will benefit from another year of distance from the option. It projects the Jackets will improve from 117 in offensive SP+ last year to 104 this season. That represents to scoring roughly a field goal more per game. It’s reasonable to believe Tech can do significantly better than that.

Collins and coordinator Dave Patenaude spent part of 2019 trying to figure out which quarterback could throw (injuries made that task harder). They eventually settled on redshirt freshman James Graham. He showed flashes of efficiency when given time to throw. That didn’t happen consistently.

Tech’s offensive linemen had the hardest transition of all. They had to drastically change their technique to block for more passes. They had to learn how to recognize pass-rushing fronts. Things should be smoother for that group in 2020.

If the growing pains subside, Tech’s offense should be much better in 2020. There’s even more reason to believe the defense will improve. SP+ projects it will rank No. 35 next season after it was No. 71 in 2019. That would mean the Jackets allow about eight points per game less in 2020.

I could see that happening. It already started in 2019. Tech switched from a head coach who seemed to treat defense as an afterthought to one who had coordinated good units for Mississippi State and Florida. The Jackets improved their defensive SP+ ranking by 29 spots from 2018 to 2019.

And now pretty much every player who contributed significantly returns (94 percent of the production, per Connelly’s formula). The Jackets were a slightly below average defense in 2019 despite fielding a young group and suffering injuries along the defensive line. SP+ predicts Tech’s defense will be good in 2020.

One reason to be cautious about Tech’s SP+ projection is the five-year history. That gives the Jackets credit for four years in which they were 24-25 while running an efficient, specialized offense they’ve since abandoned. But Connelly said recent history accounts for less than 10 percent of the projection.

Tech's schedule is the bigger obstacle to more wins. It's not just the annual talent mismatches against Clemson and Georgia or the date against Notre Dame at Mercedes-Benz Stadium. The ACC should be better, too. SP+ rated it as by far the weakest Power 5 conference in 2019, but expects it to be significantly better next season.

Four of Tech’s Coastal opponents are projected to improve in 2020. SP+ ranks those teams significantly higher than Tech: North Carolina (No. 17), Miami (23), Virginia Tech (32) and Pittsburgh (42). The Jackets play three of them on the road. Only Miami comes to Bobby Dodd Stadium.

Tech's stirring victory at Miami was its best moment from last season. The defense made big stops and scored a touchdown. Graham delivered a 35-yard touchdown pass to force overtime. It was a hint of what the Jackets can do.

SP+ is expecting more the Jackets in 2020. I agree. It might even be underselling them.