Three Outlets Say Tua Is Pulling Away. The Scheme Hasn't Voted Yet.

In Kevin Stefanski's final season in Cleveland, the Browns ran 85% of their offensive snaps out of two personnel groupings: 11 personnel and 12 personnel. That number -- 44% from 11, 41% from 12 -- tells you everything about what this offense is and what it asks of its quarterback. It is a system built on condensed formations, play-action deception, and intermediate accuracy off the run fake. It does not ask its quarterback to improvise. It asks him to process.

Three outlets in the last 48 hours have declared that Tua Tagovailoa is "pulling away" in the Falcons' quarterback competition. Blogging Dirty used that exact phrase. ESPN's Jeremy Fowler, via Heavy.com, reported that Tua has "the backing of Matt Ryan and Stefanski" and cited his accuracy as the decisive trait. Atlanta News First framed the competition timeline as accelerating in Tua's favor.

I have written about this competition three times in the last two weeks. I have established that it is structurally asymmetric -- Tua is generating 11-on-11 data that Penix physically cannot generate yet. I have established that Penix's cognitive processing is intact in 7-on-7. I have established that Stefanski's patience is a structural advantage, not a weakness. None of that has changed.

What none of those pieces answered -- and what none of the three outlets declaring a winner have addressed -- is the schematic question that actually determines who should start this offense. Not who looks sharper in OTAs. Not who has institutional backing. Who fits the scheme.

What Stefanski's System Demands

The Kubiak coaching tree -- which runs through Gary Kubiak to Kyle Shanahan to Stefanski himself -- has been producing a specific quarterback profile for two decades. The system is built on wide-zone rushing that sets up play-action passing. The play-action creates intermediate windows. The quarterback's job is to identify those windows before the defense recovers from the run fake and deliver accurately at 10 to 18 yards.

This is not a system that rewards arm talent. It is a system that rewards processing speed and intermediate accuracy executed within structure.

The evidence from Cleveland is specific. When Stefanski had Joe Flacco for the 2023 late-season run, the play-action rate hit 29.2% of dropbacks -- which would have ranked fifth in the NFL had Flacco qualified. On those play-action snaps, Flacco averaged 12.2 yards per attempt with a 125.8 passer rating. Those are extraordinary numbers, and they came from a 38-year-old quarterback with a below-average arm. Flacco's gift was the same gift the system has always demanded: he could read the second and third levels of the defense off the fake and deliver the ball on time at the intermediate level.

That is the quarterback profile. Quick processing off play-action reads. Intermediate accuracy between the numbers. Comfort in heavy formations -- 12 personnel with two tight ends, sometimes 13 with three -- where the run fake is most convincing because the defense has to honor the extra blockers. Willingness to operate within the structure rather than extend outside it.

The Case for Each

Map both quarterbacks against those demands, and the picture is more complicated than "Tua is pulling away" suggests.

Tua's case is straightforward and strong. His 2022 season in Miami -- 69.3% completion rate, 25 touchdowns, 8 interceptions, 105.5 passer rating -- is the cleanest evidence that he can execute a Kubiak-tree offense at a high level. The Dolphins' system under Mike McDaniel was a direct descendant of the same coaching tree. Tua's time-to-throw was among the fastest in the league. His intermediate accuracy was consistently above the 75th percentile. He processed quickly and delivered on time. That is the version of Tua that Stefanski is betting on -- not the 2025 version that threw a career-high 15 interceptions in Miami's collapsing offense and was benched for the final three games.

The question with Tua was never whether he could run this scheme. The question is whether the 2022 version or the 2025 version is closer to what walks into Mercedes-Benz Stadium in September.

Penix's case is different in kind but not necessarily weaker. His 2025 sample -- nine starts, 88.5 passer rating, identical to Tua's 2025 mark -- was defined by conservative decision-making. That 1.1% interception rate was the lowest in the NFL among qualified starters. In a play-action system, that profile has a specific and underappreciated value: it keeps the offense in favorable down-and-distance situations where play-action is most effective. You cannot run play-action on third-and-nine. You run it on first-and-ten and second-and-six. A quarterback who rarely turns the ball over keeps the offense in those situations more consistently than a quarterback with a higher ceiling but a higher floor, too.

The gap in their profiles is not talent versus no talent. It is quick-trigger intermediate accuracy (Tua's strength) versus ball security that preserves play-action-friendly game states (Penix's strength). Stefanski's scheme benefits from both. The question is which one matters more -- and that is a coaching evaluation that cannot be made while one quarterback is limited to 7-on-7.

What 7-on-7 Cannot Show

This is where the "pulling away" narrative fails its own logic.

Stefanski's offense depends on a quarterback's ability to sell the run fake, feel the pocket, and deliver from a boot-action platform that requires lateral movement and a planted front foot. None of that is visible in 7-on-7 work. In 7-on-7, there is no offensive line. There is no rush. There is no play-action fake because there is no running back. The quarterback stands in open space, reads a defense, and throws. It is a cognitive exercise, and it is a valuable one -- but it cannot evaluate scheme fit for a system that derives its identity from the run-pass interaction.

Tua is generating 11-on-11 data -- boot-action execution, pocket presence behind the line, timing with Bijan Robinson on play-action fakes -- that Penix is structurally barred from generating. Declaring a winner based on that asymmetry is a category error. It is like comparing a pitcher's bullpen session to a starter's five innings and concluding the bullpen arm is sharper.

Stefanski knows this. His comment about being "very intentional" with rep distribution was not throwaway coachspeak. It was an acknowledgment that the evaluation framework is incomplete. He is gathering data from one quarterback and patience from the other. When reporters describe this as Tua "pulling away," they are describing the expected output of a structurally unequal process and treating it as competitive separation.

Minicamp Is the Scheme's First Vote

Mandatory minicamp runs June 16-18. If Penix is cleared for 11-on-11 work by then -- and his ahead-of-schedule recovery from the November ACL surgery suggests it is plausible -- the evaluation framework changes fundamentally. For the first time, both quarterbacks would be operating within the same schematic conditions: play-action fakes, boot-action reads, pocket movement, a live rush (or at least the approximation of one in a non-padded setting).

That is when Stefanski can actually evaluate scheme fit. Not accuracy in a vacuum. Not cognitive processing in an empty backfield. Scheme fit -- the ability to execute the specific concepts this offense runs, under the specific conditions it creates.

The receiver room adds a layer to this evaluation. Drake London, Zachariah Branch -- the third-round rookie from Georgia who profiles as a YAC-after-catch weapon at 5-10, 180 -- Jahan Dotson, and Olamide Zaccheaus all present different target profiles. Branch in particular is an interesting variable: his twitchiness and run-after-catch ability could amplify a quick-release passer like Tua or complement a conservative decision-maker like Penix who favors high-percentage throws. How each quarterback distributes targets in 11-on-11 settings -- if both get those reps -- will reveal connection preferences that the current evaluation simply cannot capture.

Kyle Pitts, on the franchise tag, is the other schematic variable. Stefanski's system in Cleveland featured the tight end as a seam-threat weapon in 12 personnel looks. Pitts at 6-6 with a 4.44 forty is the version of that role David Njoku never became. But Pitts's seam routes off play-action are only meaningful if the quarterback trusts the fake and delivers on time. That is an 11-on-11 evaluation, not a 7-on-7 one.

The Verdict That Hasn't Arrived

The media has reached a consensus. Stefanski has not.

I understand the appeal of declaring a leader. The offseason is long, and certainty is a better headline than patience. Tua has looked accurate and confident in 11-on-11 work. That is real, and it matters. But the question this competition is supposed to answer is not "who looks better right now?" It is "who gives this specific offense its best chance to execute at the level the September schedule demands?"

That question requires both quarterbacks operating in the same schematic conditions. It requires play-action evaluation, not pocket-passing evaluation. It requires boot-action reads under (simulated) pressure, not open-field processing without a rush.

Until minicamp provides that framework -- or until Penix's knee forecloses it -- the scheme has not voted. Three outlets have. Stefanski has not. I will wait for the one that actually calls the plays.