Broncos 26, Panthers 21 Nov 8, 20263rd quarter, 1:28Carolina 2nd and goal at the 1
In the air at the snap.
Sean Payton's red flag left his hand 0.57 seconds before Carolina snapped the ball. It hit the turf 0.49 seconds after. When the ball moved, the flag was 5.9 feet up and 12 feet short of landing.
The flight from the side, drawn to scale. One dot for each video frame: found in the video by colour analysis, rebuilt from its flight and its shadow. Each ±, and each “give or take”, is one standard deviation.
What happened
First and goal: Bryce Young's pass is ruled incomplete, though Devon Key has caught it. Carolina hurries to the line. Evan Rothstein, Denver's director of game management, screams “Challenge it!” into Payton's headset. Payton pulls the flag from his back pocket and throws. Carolina snaps, and Chuba Hubbard is stopped for no gain. Only then do the officials stop the game. New York confirms the flag was out in time, the pass becomes an interception, and Denver goes on to win.
By eye, the throw seems to come right at the snap. Measured, a red patch that moves like the flag is in flight from 0.10 s before the snap. That comes from colour analysis; no reader could confirm it by eye that early. Its shadow and gravity put it near the top of its arc at that moment. No public footage shows it leave his hand: that moment is rebuilt from physics.
The throw, the snap and the landing, rebuilt in 3D and in step with the fans' own video. Press play or drag the time bar. Drag the 3D picture to look around. After a click on the replay, space plays and the arrow keys step one video frame.
This browser could not start 3D graphics, so the replay below runs on the real footage alone.
View
The real footage, same instant
Time from the snap
0.000 s
Flag
Ball
02
Who was right
The league
“…the flag was out prior to the ball being snapped.”
True. It was out of his hand about half a second or more before the ball moved.
The fans
“Red Flag hits the ground after the snap.”
Also true. At the snap it was still 5.9 feet up. The same post says the throw “starts at the snap”. That is how it looks, and it is not what happened: see why.
The rule
A coach challenges “by throwing a red flag onto the field of play before the next legal snap or kick.”
It does not say which moment counts. By the throw, Payton beat the snap easily. By the landing, he missed it. The league counted the throw. Rule 15, Section 1, Article 1(a), 2026 rulebook.
03
Why it looked so close
By eye, nobody could see it leave his hand
On the fan video the flag seems to come from nowhere, right at the snap. Three readers who had not seen these results were asked to find the throw by eye. Two of them had been told that the dark red beside the coach was not the flag, which probably made them late: they saw nothing in the air before frames 166 and 169. The third, given neutral questions and the colour maps, saw the flag free of any hand by frame 162, three or four frames before it saw the line move, and read it, tentatively, as coming from a man standing beside Payton. None of them could see a release, and none of them, alone, supports the 0.57 seconds.
What its shadow shows
The flag's own shadow shows why the eye is wrong. In frame 166, a few frames after the flag seems to leave that second man, it is 25 feet from him (give or take 2) and moving over the ground at about 14 mph. To have come from his hand when it seems to, frame 160 at the earliest, it would have needed about 85 mph (at least 72). It had been in the air for a while. It just could not be seen.
Why it may have left no trace
The reconstruction has Payton tossing it from hip height about 14 frames before any trace of it can be found. Why it leaves no trace is inference, not observation. It was thrown almost straight at the camera (17° off the line of sight), and two ounces of sand make a knot about 3 pixels across at that distance. When the red patch does show, in frame 153, it is about 16 pixels across. It shrinks to about 3 by 6 in frame 156 and grows to 27 by frame 161. That fits a cloth tumbling and opening out behind the knot, but in its first frames the patch may also take in the second man's hand or face. From that fan's seat this part of the arc lines up with his head. He is about 16 feet behind it (give or take 5).
See it , then .
14frames
in the air before any trace of it can be found (give or take 1.2)
25ft
from the second man in frame 166, a few frames after it seemed to leave his hand (give or take 2)
85mph
what it would have needed to get there from him (at least 72). Over the ground it was doing 14 mph, give or take 6.
Five moments from clip A. Above, the picture. Below, only what turned redder in that frame; a moving hand lights up as well as a red flag. 138: Payton's right hand comes away from his body (he is at the right, in the white cap). 141: his arm is out. 147: nothing to see. 153: a red patch shows, in line with the second man's head. 159: it is falling, as the line starts to move.04
How we know
The sun was a second camera
One camera cannot tell how far away a small object is. Its shadow can. From frame 164 a dark streak slides across the turf and meets the flag where it lands. It is faint in frames 164 and 165 and strong from 166. The line of sight to the flag and the sun's ray through that streak pass within 3 inches of each other in all 6 frames measured, as they must if the streak is the flag's shadow. Where they cross is where the flag is: in frame 166, 10 feet inside the field (give or take 1) and 2.9 feet up. That is 25 feet (give or take 2) from the man it seems to come from. Over the next four frames it covers the ground at 20 feet a second, give or take 8. To have left his hand when it seems to (frame 160 to 162) it needed 85 to 128 mph; even from the first frame red shows near him (153) it needed 39 mph. None of this assumes who threw it.
The flag and its shadow closing on each other. Red dots: the flag, frame by frame. Blue ellipses: the dark streak on the turf in the same frames. Each grey line is the sun's ray. Dashed: the fitted flight and the shadow it predicts. They meet where the flag lands.
Gravity is the stopwatch
Between frames 166 and 170 the flag is falling at 18 feet a second. Gravity alone says it stopped rising about 0.5 seconds before the middle of that stretch: frame 152, about 6.3 feet up (give or take 0.5). So when the red patch first shows, the flag is at the top of an arc, not leaving anyone's hand. Traced back further, the flight lines up in the picture with Payton's hand in frame 139, give or take 1.5 (typically 5 pixels away; the 95% range runs up to 15), when it would have been 3.5 feet up, give or take 1.3: hip to chest height. Left free, gravity comes out at 10.0 ± 0.4 m/s² against the true 9.80, so the clip plays at true speed to within a few percent.
Every frame at once. Each pixel that turned red in a frame, and was not red there four frames earlier or later, is painted in: yellow early, red late. Nothing was picked by hand. The band is the flag from the top of its arc down to the turf. The blob at right is Payton's hand and the lettering on his chest.
One clip holds almost every event
Fan clip A shows Payton, the flag and the line of scrimmage in the same frames, so the throw, the line's movement and the landing need no syncing between cameras. His right hand moves out from his body to chest height in frames 138 to 141. Colour analysis picks the flag up from frame 153. By eye, one reader, helped by the colour maps, saw it clearly from frame 162; the other two from 166 and 169. It reaches the turf between frames 170 and 171. The quarterback is in steady movement from frame 155.5 and the line follows at about 159.5; to the readers the line is clearly moving at 164 to 166.
The snap, seen from behind
The ball is hidden in the fan clip, so the snap is timed from the men around it. A CBS shot from behind the offense shows the quarterback under center and the lineman in front of him, the center. Their helmets start to move within 0.05 s of each other. The quarterback's arms jump 0.14 s after his helmet starts. That is taken as the ball reaching his hands: the track is a noisy one, but the rest of the line starts in the same gap between two pictures, 0.16 s after his helmet. The fan clip shows the same order, with the line 0.13 s behind the quarterback. On the fan clip's clock the center first stirs at about frame 153.5 and the quarterback's arms jump at about 159. The snap is taken as the middle of that window: frame 156, give or take 2 frames (0.07 s).
Two cameras, one field
Each fan's camera was located from the painted lines alone, to within about 2 pixels. Working separately, they put the referee within 1.2 feet of the same spot and the resting flag within 4.5 feet.
The stadium clock
Clip B has the play clock in shot. Its ticks come 30 or 60 frames apart (one tick was missed), so that clip is true speed too. It shows the same order: the line moves, then the flag reaches the turf.
The checks that were held back
The two faintest frames of the shadow, 164 and 165, were kept out of the fit; it predicts them to about 4 pixels. The referee's own shadow, read off the same clip, puts the sun 24.1° up and 7° round from where the almanac has it (22.4° up at 3:09 pm). Using that sun instead moves the flag by 1.4 feet.
05
The throw
The conditions on the day
06
Why the play was not stopped
The officials were set for a goal-line snap. The flag came down 23 yards upfield, 12 feet inside the Broncos sideline. For the referee that is well behind his shoulder, 121° off his line to the ball. The one official measured as roughly facing that way was about 170 feet off, where the flag is a red patch a few inches across. Nobody stopped the play. Payton afterward: “Was anyone able to see that I threw it before the play was snapped? And fortunately, New York can.” See , or stand in the .
Every official, and how far off his line to the ball the flag fell
Official
Flag, off his line to the ball
Away
Position
07
What is measured, and what is not
Measured from footage
The flag in 17 frames, by colour analysis (by eye, one reader saw it clearly from frame 162, the other two from 166 and 169), and its shadow in 7, of which 5 go into the fit. Payton's hand, his spot and his walk. The first movement of the quarterback and of the line. Where the flag came to rest. The referee, umpire, back judge and both wing officials. Ten of the eleven defenders, the backs and the width of the offensive line. The stadium's shadow.
Rebuilt with physics
The first 14 frames of the flight, the moment of release, the speed and angle, the drag, the instant of touchdown to a fraction of a frame.
Inferred
Why the flag leaves no trace early (a small knot coming at the camera, opening later). That the ball moved some time between the center first stirring and the quarterback's arms jumping. That the officials were watching the snap.
Standard, not seen
The two deep officials at the end line. How deep the linemen sat (the defense hides their feet).
Drawn in
How the pile moves after the snap, and the people on both benches. The footage cannot resolve them and nothing here depends on them.
08
Limits
Nobody can see the release.
The first 14 frames of the flight are rebuilt, not watched, and a search of those frames along the rebuilt path found nothing above the noise. The release is where the flight meets his hand: frame 139 ± 1.2. Its 95% range, frames 137 to 141.5, is 0.48 to 0.63 s before the snap; that range covers the release only and takes the snap frame as given. Across the picture the meeting with his hand is tight. Along the camera's line of sight it is loose, because the shadow is only measured in the last frames: there the flight could pass anywhere from 1 to 22 feet from where he stands (a 95% range). What rules out a later throw is gravity: the flag is at the top of its arc, six feet up, in frame 152, while his hand is at chest height.
The eye does not show the throw, and three readers confirm it.
Readers 1 and 2 got plain enlarged frames and a note that wrongly told them the dark red near the second man was his clothing, which probably made them late. Reader 1 first saw the flag in the air at frame 166 and guessed the release at 162 to 165, about level with the snap as that reader saw it (164 to 165). Reader 2 first saw it in the air at frame 169 and reported a dark-red patch at Payton's hand in frame 168, after the snap as it saw it (164). One frame later the flag is 24 feet from him and about a foot above the grass, so that patch was not the flag. Reader 3 got neutral questions and the colour maps made for this analysis: no distinct red object at Payton's hand in frames 134 to 150, though it could not rule out something small, a reddish blotch around the second man from 153 that could have been skin, the flag clearly free of any hand at 162, and the line clearly moving at 165 or 166. So reader 3 saw the flag in the air three or four frames before it saw the line move, and read it, tentatively, as thrown by the second man. None of the three could see a release, and none supports the 0.57 s on its own.
The shadow is faint.
It is a streak 5 to 17 pixels wide and 2 to 7 high. Five frames go into the fit; the two faintest, 164 and 165, were held back and fit. It moves steadily and meets the flag at the landing. If it were not the flag's shadow, the case that the second man did not throw it would rest on his hand being on his hip until frame 152, which reader 3 and the frames agree on, and on every account naming Payton.
The ball is never seen leaving the ground.
In the fan clip it is hidden behind the line, and in the CBS shot behind the center. So the snap is a window, not an instant: after the center first stirs (frame 153.5, carried over from the CBS shot) and before the quarterback's arms jump (frame 159). Its middle, frame 156, is used, give or take 2 frames: 154 to 158. Where a movement starts depends on how it is read. The quarterback's first steady movement in the fan clip is frame 155.5 (give or take 1) and the center's, from the CBS shot, is 156; a fit that assumes each man started from rest puts both 2 to 3 frames earlier. The CBS file repeats frames, so it holds about 29 different pictures a second. It is lined up with the fan clip through the linemen, fitted the same way in both: 3 of them in the CBS shot, 6 in the fan clip. Lining it up through the quarterback instead moves it by 0.7 of a frame, and the alignment is good to about 1 frame. Rerunning both trackers with other settings (135 combinations) moves the middle of the window between frames 155.6 and 157.2; its start moves between 152.7 and 154.5 and its end between 158.5 and 159.9. The rest of the line starts later: frame 159.5 is the median of 6 linemen, who start between 157.5 and 163.3 (a seventh track was too noisy to use). If the ball moved only then, the flag left his hand 0.68 s before the snap and landed 0.38 s after.
By eye the surge is not plain until about frame 165.
Take that as the snap and the flag left his hand 0.87 s before it, could be detected in flight 0.40 s before it, was clear to one reader 0.10 s before it (to the other two only after it), and landed 0.19 s after. The order of throw, snap and landing does not change.
Two readers put the landing a frame or two early.
Readers 1 and 2 put the flag on the grass in frame 169. In that frame it is still a foot up; its shadow has not reached it. Contact is between frames 170 and 171.
The posted clips are re-encoded at 30 frames a second.
One frame is 0.033 s; the two margins are about 17 and 15 frames.
The league's own angles have not been published.
So this cannot check exactly what the replay center saw.
09
Ways to make this better
Ask both fans for the original files. A phone's own file is sharper, often 60 frames a second, and carries its capture time. That alone should roughly halve the timing errors and might show the flag in the frames where it now leaves no trace.
Follow the shadow all the way back. With a sharper file the shadow should be traceable across the white border, where it stands out most. That would pin the whole arc from the sun alone.
Find a fan behind the Broncos bench. A view of Payton's hand from behind would show the release itself, the one thing rebuilt here and not seen.
Use the league's tracking chips. The ball and every player carry one, read ten times a second. That gives the snap instant and all 22 positions without video.
Get the league's angles. The replay center used views nobody else has seen. One sideline shot of Payton would settle the release frame outright.
Throw a real flag. Two ounces of sand in a red cloth with three rubber bands, filmed at high speed in the same light, would show how long it stays balled up and pin down its drag.
Track every body. Pose-tracking models on the broadcast and both clips could replace the drawn-in movement after the snap with the real thing.
Fix the rule's wording. One word, “released” or “landed”, ends this argument. A sideline button that time-stamps the challenge ends it for good.
10
Sources
2026 NFL rulebook, Rule 15: Section 1 (how a coach challenges, page 67) and Section 9 (replay may assist the officials, page 71). Contents page
