For a decade the answer to “why didn’t New York fix that” was the same. Replay could take a flag off the field. It could not put one on.
That asymmetry was deliberate, and for 2026 it is gone for one narrow category of foul.
What actually changed
League personnel can now consult with on-field officials and disqualify a player for a flagrant football act or non-football act, without that act having been called on the field.
Read the boundaries carefully, because they are most of the story. It covers flagrant acts. It produces an ejection. It was approved as a one year rule, which means the league gets a season of evidence before deciding whether to keep it.
It does not hand replay the ability to add yardage. A missed holding stays missed. A missed pass interference stays missed, and pass interference remains unreviewable and unchallengeable, which has been true since the 2019 experiment was not renewed.
Which direction replay runs, before and after
| Situation | Before 2026 | 2026 |
|---|---|---|
| Flag thrown in error on a covered foul | Booth can pick it up | Unchanged |
| Flagrant hit the crew missed entirely | Nothing available | Ejection can be applied from the booth |
| Ordinary foul the crew missed | Nothing available | Still nothing |
| Pass interference, called or not | Not reviewable | Still not reviewable |
| Yardage on a missed foul | Cannot be added | Cannot be added |
One row moved. The rest of the table is the same table.
Why the league resisted this for so long
An officiating department that can add fouls from an office in New York is an officiating department that can be accused of deciding games from an office in New York. The no-adding rule was the answer to that accusation, and it was a clean one.
Giving it up for player safety is a trade the league is willing to explain. Giving it up for a holding call is not.
Which is why the scope is ejections rather than yardage. An ejection is defensible as a safety intervention. A spot foul added from the booth would be the thing the rule was built to prevent.
The claim we published, and how it gets graded
Before the season we put a number on this one. Replay adds a flag five times or fewer across the regular season. Six or more and we were wrong, and we will say so.
The reasoning: new powers get written broadly and used narrowly, especially when the league is nervous about the perception. We expect it reserved for the obvious, a flagrant hit away from the ball that everyone in the stadium saw except the seven people being paid to see it.
There is no historical rate to anchor this, because the power has never existed. That makes it a genuine forecast and the most likely of our three opening claims to be wrong. It resolves at the end of Week 18, in public, either way. The full set is in the three claims we opened the book with.
What to watch for on Sunday
The tell is a stoppage with no flag on the grass and no coach challenging anything. Officials conferring, a delay that runs longer than a spot check, then a disqualification. That sequence did not exist last season.
Watch for the second-order effect too, and it is the more interesting one. If a crew knows the booth is covering flagrant acts, does the crew on the field throw fewer of them? A safety net changes the behavior of the person on the wire. That is not something the league will report, and it is the kind of thing a season of penalty data can answer.
Three other changes passed for 2026 that touch officiating: the kickoff alignment requirement, onside kicks declarable at any point in the game, and touchbacks coming out to the 20 when a penalty moves the kickoff to the 50. The alignment rule is the one to watch for early flags, because new formation requirements are what preseason exists to teach.
When one of these lands in a game you are watching, the rule behind the call is usually the thing worth checking first.