Georgia Tech begins the final third of its season with no clear indication whether we will see Good Tech or Bad Tech. That’s what coach Paul Johnson calls the dichotomy that’s resulted in a 3-1 record over the second third of the season after a 1-3 mark in the first.

Good Tech pummeled ACC opponents in its last two road games. Bad Tech lost to South Florida of the AAC (not to be confused with the ACC) and farcically lost fumbles on three consecutive plays in the homecoming loss to Duke.

Bad Tech can lose at North Carolina on Saturday for the third defeat in four games as a road favorite. Good Tech can win that game with no sweat, then follow it with a victory over Miami for just the second time in nine years.

The Yellow Jackets are 4-4 overall, 3-4 against FBS foes and 2-3 in the ACC. Tech has been slightly-below average overall, not with a steady stream of OK play, but with wild swings from great to garbage.

Just one Tech game ended with a one-score margin, the 24-19 loss at Pitt. The Jackets blew out two ACC opponents on the road, one not bad (Virginia Tech) and the other not good (Louisville). Tech was blown out at home by two ACC opponents, one of them (Clemson) elite and the other (Duke) just OK, like the Jackets.

You get the picture, and it’s muddled. Even after that dominant victory at Virginia Tech last week, Johnson wasn’t ready to say Good Tech is the norm.

“We will see,” Johnson said Wednesday. “The last four games are going to be tough games against good opponents. But I’ve said all along, outside of Clemson, everybody in our league is pretty close.”

Mathematically, the Jackets still can finish atop the ACC Coastal. Realistically, they won’t. Tech must win two of its final four to be eligible for a bowl and that’s more feasible.

The Yellow Jackets are favored by 4½ points at North Carolina. They figure to be slight underdogs at home against Miami. The Virginia game looks like a toss-up. Some seemingly over-matched Tech teams ambushed Georgia, but this one won’t.

Will the Yellow Jackets win two more games and make a bowl game? I think so because, for three of the last four games, the option offense has operated at near-peak efficiency. But, honestly, you probably should take a skeptical view of my opinion about the Jackets because I haven’t been able to figure them out.

I picked them to win at South Florida and Pittsburgh and they lost both games as favorites. I did call Tech’s loss to Duke, and correctly picked Clemson to cover the point spread. But I also thought the Jackets would lose handily at Virginia Tech, and instead they won going away.

Before the season I predicted Tech would win seven games. My optimism about Tech largely was based on the expectation that Johnson would get his offense back to humming with more experienced players.

At the time, I noted it was a good sign that Johnson said he liked both of his quarterbacks. It’s worked out that way. Tobias Oliver was great against Virginia Tech as the injury replacement for TaQuon Marshall.

Still, the offense has been erratic. Johnson weirdly seemed to lack confidence in his option during crucial moments at South Florida and Pitt, opting to pass with predictably bad results. The option couldn’t be stopped at Louisville and Virginia Tech but, in between those games, it misfired against Duke.

In the preseason, I figured Tech’s defense, with new coordinator Nate Woody, would create more turnover chances by being disruptive. That’s gone according to plan. The Jackets have 17 takeaways in eight games after they had 10 all last season.

But the Jackets couldn’t stop South Florida and had their defensive shortcomings covered up at Louisville and Virginia Tech. Then again, Tech’s defense gave it a chance against Pitt and Duke, only to see the offense stumble.

“At times they are better than others, just like on offense,” Johnson said. “I think overall what we are doing (defensively) is sound. I think the more they do it, they better they will get at it.”

As for luck, that’s also swung in both directions for Tech.

Promising B-Back KirVonte Benson suffered a season-ending injury after two games, and his position is crucial to Johnson’s option. Injuries along the offensive line hurt against Duke. Allowing consecutive kickoff returns of nearly 100 yards each to South Florida wasn’t bad luck, exactly, but how many times do you see that?

The Yellow Jackets have had good luck recovering fumbles, which tends to be a random thing. They got back just three of their 11 fumbles but recovered all seven they’ve forced and one that they didn’t. The Jackets lost those three fumbles in a row against Duke but recovered two others they had in that game.

To get a better hold on Tech, I looked to the advanced statistics. The model created by Bill Connelly at SB Nation gives the Jackets a 36 percent chance of finishing with six victories. It gives them a slightly higher chance (38 percent) to finish with five.

Even the numbers seem unable to decide if Good Tech or Bad Tech is the Real Tech. The last four games should provide clarity.