The book is empty. Every crew rate, every team tendency, every claim we make later comes out of games that have not happened yet, and the honest thing to do before any of it starts is say what we will count, what we will refuse to count, and how you will know if we are any good at this.
So this is the first entry. Three claims about the 2026 season, each with a number attached that makes it wrong, published before a single regular season snap.
What preseason cannot tell you, and why we are saying so first
The temptation right now is to write about crews. Three weeks of preseason football just happened, flags were thrown, and the data exists.
We are not going to use it, and it matters that we say that in August rather than in December when a number stops being convenient.
Three reasons. Preseason crews are not regular season crews; the league uses August to evaluate and rotate officials, so the seven names working a game in Week 2 of preseason may never work together again. The people committing the fouls mostly will not be employed on Sunday, because rosters cut from 90 to 53 on August 30 at 6 PM eastern. And officiating in preseason is calibration, not enforcement. New rules and points of emphasis get called tight in August on purpose, to send a message to players who are still learning them.
That last one is not a theory. It is the most reliable pattern in officiating, and it repeats every year: a point of emphasis produces a flag spike in preseason that regresses once the games count.
So the rule for this book: preseason data never enters a crew or team profile. It is useful for exactly one thing, which is watching what the league is emphasizing. That is what these claims are about.
What is actually knowable right now
Five rule changes passed for 2026. Four of them are worth a fan’s attention, and one of them changes officiating more than anything the league has done in years.
| 2026 change | What it does | Flag impact |
|---|---|---|
| Replay can add a flag | Replay may now eject a player for a flagrant football or non-football act, and may drop a flag when the crew on the field did not throw one | New category of foul entirely |
| Kickoff alignment | Receiving teams put five players on the restraining line instead of six, with up to four floaters and limits on overloading one side | New formation foul to learn |
| Onside kick any time | A team may declare an onside kick at any point in the game, not only in the fourth quarter | More kick plays, more kick fouls |
| Touchback at the 20 | When a penalty moves the kickoff to the 50, touchbacks and out of bounds come out to the 20 rather than the 25 | Raises the cost of a kickoff foul |
The first one deserves a paragraph on its own. Until this year, replay in the NFL could take a flag off the field but never put one on it. That asymmetry was deliberate and it was the answer to every “why didn’t New York fix that” argument for a decade. It is now gone for one narrow category of foul.
Nobody knows how often that power gets used, because it has never existed. That is the best kind of thing to make a claim about.
The three claims
Claim 1: Replay adds a flag five times or fewer all season
New powers get written broadly and used narrowly, especially when the league is nervous about being accused of officiating games from an office in New York. We expect this to be reserved for the obvious: a flagrant hit the crew missed, an ejection-level act away from the ball.
Resolves as a miss if replay adds a flag six or more times in the regular season. Resolution date: end of Week 18.
Basis: no historical rate exists, which is the point. This is a genuine forecast and it is the one most likely to be wrong.
Claim 2: Kickoff formation fouls spike this weekend, then fall by more than half
The new receiving team alignment is exactly the kind of rule that produces August flags. It is a formation requirement, it is new, and preseason is where officials teach it.
Resolves as a miss if the kickoff formation foul rate in regular season Weeks 1 through 4 is more than half the rate we count across preseason Week 3. Resolution date: end of Week 4.
Basis: the preseason spike and regression pattern, which has repeated on every significant point of emphasis we can find, in some years by a factor of three or more on the emphasized foul.
Claim 3: A crew’s preseason numbers will not predict their regular season numbers
This is the claim that justifies the rule we set at the top, and it is the one we would most like to be wrong about, because being wrong would mean three extra weeks of usable data every year.
Resolves as a miss if the correlation between preseason flags per game and regular season flags per game, for referees who worked both, comes out above 0.3 at the end of the season. Resolution date: end of Week 18.
Basis: preseason crews are mixed and rotated for evaluation, so the unit being measured in August is not the unit being measured in October. If that is true, the correlation should be close to nothing.
How the scorecard works
Every claim published here carries the number that makes it a miss, written before the games. That is the whole mechanism, and it exists so that grading day is not a negotiation.
- A miss is a miss. There is no partial credit, no “directionally correct,” no quiet edit after the fact. A claim that gets softened once makes every other claim on the board meaningless.
- Claims are published with their sample size. Under eight games of history behind a crew and we will not make a claim at all, we will say the book is thin. That will be a common sentence in September.
- The running record is public. Hits, misses, and the worst miss of the season, updated every Monday.
- Nothing here predicts a game result. These are claims about officiating behavior. No scores, no outcomes, no lines.
One of the three claims above resolves at the end of Week 4. The other two run to the end of Week 18, and we will carry them on the board all season, visible, whether or not it is going well.
What happens next
The final preseason games run Thursday through Saturday. Rosters cut to 53 on Sunday evening, and the first regular season crew assignments follow after that, which is when the book actually starts filling.
The first real entry publishes the Thursday before Week 1, with claims about crews rather than about rules. From there it is the same three beats every week: what we expect on Thursday, what happened on Monday, and what changed for your team on Wednesday.
Claim one is already live in the rulebook. We took the new replay power apart in what the replay center can and cannot do now.
The app is where the calls get rated while they are still happening. This page is where we write down what we thought beforehand and admit how it went.