LAS VEGAS — Dan Mullen was asked Thursday night to grade Jackson Arnold through three games and immediately explained why he dislikes grading the quarterback without grading everything happening around him.
Mullen can stop the film at the point the football should leave Arnold’s hand and see a free rusher already forcing him out of the pocket. Another play might show Arnold holding the ball after the protection gave him enough time. A running back can miss a block. A tackle can lose his matchup. The quarterback can be late moving from one read to the next.
All of them can produce the same statistic.
A sack.
UNLV has allowed 11 of them through three games, one of the clearest reasons an offense averaging 375 yards has produced only 16 points per game. Mullen does not believe the problem can be isolated to Arnold. He also does not believe Arnold is finished developing inside the offense.
His assessment Thursday was more specific than anything he had offered previously.
“I don’t see him make a ton of bad reads,” Mullen said. “I think he’s a little conservative at times with the ball and needs to be a little bit more aggressive. I think he needs to get through reads a little bit faster.”
That is the next part of UNLV’s offensive self-scout.
Arnold has completed 51 of 82 passes for 601 yards, two touchdowns and no interceptions. He threw for 291 yards and both scores against Memphis, attempted only 18 passes while UNLV rushed for 248 yards at Hawai‘i and completed 18 of 27 for 194 yards in the 44-6 loss at North Texas.
The box scores describe three different games. The recurring problem appears when the offense falls behind schedule.
UNLV’s success rate on passing downs is 18 percent. Memphis sacked Arnold six times, all after halftime. Hawai‘i produced only one sack during a game controlled by the run. North Texas added four after holding UNLV to 98 rushing yards and forcing the Rebels into more obvious passing situations.
Mullen spent part of the bye working through why Arnold sometimes gets to an answer quickly and why another play takes too long.
His explanation began with the quarterback’s feet.
On a rhythm concept, Mullen wants the progression tied to a three-step sequence. The quarterback’s eyes have a job on the first step, another on the second and another as the third step completes. If the first throw is there, the ball leaves. If it is covered, the quarterback hitches and moves to the next answer. If that is covered, he resets again.
“One, two, three, hitch, throw,” Mullen said.
The description sounds simple until the entire process is compressed into the time available during a college football play. Mullen estimated some of those decisions have to happen inside roughly 1.6 or 1.7 seconds.
A foot slightly out of position can change the throw. Eyes arriving late can delay the progression. A defender rotating after the snap can change which route is available. A pressure can force the quarterback to move before the original read has finished developing.
Mullen believes those decisions become easier with repetition because experienced quarterbacks begin recognizing the answer before every part of it unfolds.
He used Dak Prescott as the obvious example from his own coaching career.
The comparison was not that Arnold should already play like Prescott. Mullen’s point was the opposite. Prescott spent several seasons in the same offense and accumulated thousands of college snaps before reaching the level where much of the picture could be anticipated before the ball was snapped. Arnold is still early in his first season running Mullen’s offense.
Mullen described showing Arnold a play in which the quarterback initially worked one side of the field. The primary target slipped. Mullen wanted Arnold to already know what was developing on the opposite side based on the coverage picture before the play had fully unfolded.
That is the processing speed he wants accelerated.
The same principle was part of Mullen’s work with quarterbacks long before UNLV. He recalled former Utah quarterback Alex Smith reaching a point where coverage structure and defensive movement could be identified quickly enough that unusual defensive behavior required specific “awareness” rules rather than ordinary progression teaching.
Arnold is not there yet.
Mullen also made clear that the timeline is not an excuse.
“My expectations are really high,” Mullen said.
UNLV cannot wait three seasons for its quarterback to understand the offense because the games being played now still count. The coaching job is to shrink the learning curve, remove concepts Arnold does not trust and get him to the comfortable answers more quickly.
That is why Mullen’s earlier bye-week comments remain relevant.
“I might like a play, but if he’s not comfortable running it, it really doesn’t matter what I like,” Mullen said earlier this week. “I’m standing on the sidelines. No one’s coming to hit me and I don’t have to throw a ball.”
Three games have given Mullen and offensive coordinator Corey Dennis real evidence rather than training-camp projections. If Arnold repeatedly hesitates on one concept, the staff has to decide whether the answer is more repetition or fewer calls. If another concept consistently gets the ball out on time, UNLV should have more reason to return to it.
Memphis provided the clearest evidence of what happens when Arnold sees the answer quickly.
Against the Tigers’ blitzes, Arnold completed six of seven passes for 102 yards and both of his touchdowns. His average time to throw on those snaps was about 2.14 seconds.
The additional rusher did not automatically make the play harder. It often defined the coverage.
Arnold knew where the replacement throw was.
The ball left.
The more difficult plays have come when the defense keeps enough players in coverage, the initial picture changes and Arnold has to move deeper into the progression while the protection deteriorates around him.
That becomes more important when his career sack history is considered.
Arnold was sacked 35 times on 246 pass attempts at Oklahoma in 2024. He was sacked 30 times on 215 attempts at Auburn last season. He has been sacked 11 times on 82 attempts through three games at UNLV.
Using pass attempts plus sacks as a simple denominator, while recognizing that it is not a true dropback rate because scrambles are excluded, the percentage has remained remarkably similar across all three stops.
The offensive lines have changed.
The systems have changed.
The sack problem has followed him.
That does not make every sack an Arnold statistic. Memphis produced 21 total pressures in the opener. Mullen has identified a North Texas sack caused by a missed running-back block. UNLV’s tackles have lost individual matchups.
It does suggest the offense cannot solve the problem solely by asking five linemen to become perfect.
Some improvement has to occur before the rusher arrives.
Mullen believes Arnold has already shown enough to make that realistic. He is not turning the football over through the air. Mullen does not see a quarterback repeatedly choosing the wrong receiver. His criticism is narrower: trust the throw when it is available, move through the progression faster and connect the feet to the decision.
Akron provides a particularly useful test because the Zips have recorded only two defensive sacks through three games.
There should be opportunities for UNLV to keep the pocket from becoming the defining feature of the offense.
Austin Boyd will remain unavailable at center, leaving Griffin Scroggs and BJ Tolo to handle the position again. But North Texas showed why Boyd’s injury does not explain every pressure. Scroggs allowed one hurry in the pass-blocking sample from Denton, while much of the pressure arrived elsewhere.
Saturday therefore gives UNLV a cleaner evaluation.
If Arnold has time, does he use it efficiently?
If the first throw is available, is the ball gone?
If the defense removes it, does the progression move before the pocket forces him to?
Mullen believes the quarterback is seeing more of the field correctly than the sack total suggests.
The next step is seeing it sooner.
See something that needs correcting? Report a correction.