Andy Dalton fired deep down the sideline, with Cody Core streaking downfield with Pittsburgh’s Joe Haden in coverage, step-for-step, gradually moving him closer to the sideline. Steelers safety Mike Mitchell was alone in the middle of the field but Dalton made his mind up within a few yards – he was going deep.

Core leaped, reaching over Haden for the catch – but his momentum carried him out of bounds before he could get a second foot down at about the Pittsburgh 21-yard line. Pittsburgh had seven men within five yards of the line of scrimmage and on the snap sent five. The offensive formed a perfect pocket, allowing Dalton time to throw.

It could have been a 30-plus yard completion.

“We’re inches away,” Dalton said.

Unfortunately for the Bengals, the end result of that late third quarter play in Sunday’s loss at Pittsburgh was emblematic of the struggles the team has faced in a facet of the offense that was once a dynamic strength.

Through six games the Bengals are 29th in the National Football League in scoring with 98 points, or 16.3 points per game. Since Bill Lazor took over play-calling duties four weeks ago – a quarter of the season – that has improved to 22.3.

But the deep ball has yet to return.

“I’d rather (have explosive plays),” Lazor said. “But sometimes, some of the games where we’ve won, we’ve had long drives and we’ve just been real efficient. Andy’s completion percentage is high and we put drives together. There are two ways to do it. I just think you can’t always assume (them), but it makes it easier. That’s the thing. It makes it easier. To have a 12-play drive for a touchdown is difficult. You’re assuming you don’t make many mistakes. We want to have them.”

It’s been a trend, however.

The Bengals have attempted fewer deep throws with less efficiency each of the last two seasons. Pro-football-reference.com considers a deep throw to be 15 yards, and in 2015 the Bengals went deep 109 times, completing 58 (53 percent). Dalton threw 10 touchdowns against five interceptions for a 112.8 rating.