Study · NFL · 2021 to 2024
Lose the QB,
lose 3 points.
When an NFL team plays without its starting quarterback, it does about 3 points worse. The closing spread moves about 3 points too.
226 games, 68 team seasons, 2021 to 2024 · Published 1 November 2026 · Free to quote under
CC BY 4.0 ·
Download the data (CSV)
NFL 2021-24QB out
3.1
Points the spread moved
The margin is how many points a team wins or loses by. The closing spread is the betting market's last guess at that margin before kickoff.
We compared each team with itself: the games its starter missed against its games with him in the same season.
See who is out this week on our NFL injury report.
What a missing QB
costs.
Teams went 75 wins, 150 losses and 1 tie without their starter. That is 3.3 wins in every 10 games. The same teams won 4.1 in every 10 with him, and those were not great teams to begin with.
| Per game | With starter | Without | Change | Likely range |
|---|
| Wins per 10 games | 4.1 | 3.3 | -0.8 | -1.5 to -0.1 |
| Points scored | 19.8 | 18.8 | -1.0 | -2.8 to +0.6 |
| Points allowed | 22.6 | 23.8 | +1.2 | -0.6 to +2.9 |
| Margin | -2.7 | -5.0 | -2.2 | -4.7 to +0.2 |
| Closing spread | Underdog by 1.1 | Underdog by 3.9 | -2.8 | -3.6 to -2.0 |
Each row weighs a team by how many games its starter missed. Likely range = 95 times in 100 when we resample whole team seasons. If it includes zero, we can't tell the change from luck. Figures are rounded.
Those are the plain averages. They don't allow for who a team played or where. Allow for both and the margin falls 3.1 points (likely range 0.3 to 5.9) and the spread moves 3.1 points (likely range 2.3 to 4.0). That is where our 3 comes from.
The drop in points scored alone is small: about 1 a game, or 1.6 after allowing for the opponent. Teams also allowed about 1 point more. Neither is clear of noise alone. Together they add up to the lost margin.
The spread got it
about right.
The market took off about 3 points and teams lost about 3 points. After allowing for opponent and venue, the gap between the two is 0.0 points, with a likely range of -2.9 to +2.9.
Teams without their starter beat the closing spread 100 times, missed it 114 times, and landed on the number 12 times. That is close to a coin flip.
The closing total also fell, by 2.3 points (likely range 1.5 to 3.0). Real scoring in those games did not fall by a clear amount, but the range is wide, so we would not read much into that.
This is not a betting system. The spread already knows. We stop at the 2024 season for anything about prices, because later seasons are still being used to test our own model.
Good teams
lose more.
The average hides a big split. Take teams the market usually made favourites by 2 points or more. Losing the starter moved their spread 5.6 points. They won 17 of 43 without him, or 4.0 in every 10, after winning 5.9 in every 10 with him.
Teams that were usually underdogs barely changed. The spread moved 1.4 points and they won about as rarely as before.
| Team with its starter | Games missed | Record without | Wins per 10 with | Spread change | Margin change |
|---|
| Usually favoured by 2+ | 43 | 17 - 26 | 5.9 | -5.6 | -4.9 |
| Usually within 2 | 110 | 40 - 69 - 1 | 4.3 | -2.7 | -1.4 |
| Usually underdog by 2+ | 73 | 18 - 55 | 2.8 | -1.4 | -1.9 |
Spread and margin are not adjusted for the opponent here. The spread is cleaned for home field. The favoured group is 15 team seasons, and its margin range runs from -10.4 to 0.0, so treat the size as rough.
Receivers and backs:
mostly noise.
We ran the same test on each team's top receiver and lead running back. The spread moved about 1 point for a receiver and about half a point for a back. On the field, we can't see a clear effect for either one.
| Player missing | Games | Record | Margin change | Likely range | Spread change | Likely range |
|---|
| Starting QB | 226 | 75 - 150 - 1 | -3.1 | -5.9 to -0.3 | -3.1 | -4.0 to -2.3 |
| Top receiver | 158 | 72 - 86 | -0.4 | -3.6 to +2.9 | -0.9 | -1.7 to -0.2 |
| Lead running back | 193 | 76 - 117 | -2.0 | -5.0 to +1.0 | -0.5 | -1.2 to +0.1 |
These rows allow for opponent and home field, and for other key players out in the same game.
A team that lost its top receiver won 4.6 games in 10, the same as with him (4.5). That does not mean receivers don't matter. It means one missing player is hard to see in 158 games, even when the market moves.
Some teams fell off.
Some got better.
The average also hides how far single teams swung. These are the three biggest falls and three biggest rises in points scored, among teams that missed their starter for 5 games or more.
| Team | Starter out | Games missed | Points with | Points without | Record without |
|---|
| New Orleans 2024 | Derek Carr | 6 | 24.5 | 12.3 | 0 - 6 |
| NY Giants 2021 | Daniel Jones | 5 | 18.2 | 9.8 | 0 - 5 |
| Jacksonville 2024 | Trevor Lawrence | 6 | 21.7 | 13.7 | 2 - 4 |
| Cleveland 2024 | Deshaun Watson | 5 | 15.8 | 21.8 | 2 - 3 |
| Washington 2022 | Carson Wentz | 6 | 16.0 | 21.8 | 5 - 1 |
| New England 2024 | Jacoby Brissett | 6 | 14.5 | 18.3 | 1 - 5 |
Five or six games is a small sample, and the two groups of games had different opponents. These are stories, not rates.
Why to be
careful.
We can't see why he missed. We have no injury reports for these seasons. A player who was benched looks the same as a player who was hurt. So does a suspension. A benched starter was often playing badly, which makes the backup look better than an injured starter's backup would.
Lost seasons. Late in a bad year, teams have less to play for, and they may leave a hurt starter on the bench. We left out week 18, when good teams also rest players. And we split the games by record, below.
Starters don't get hurt at random. A team's line may be struggling when its quarterback goes down. Some of the drop could come from that, not from the backup.
So we tried the test on smaller groups. The story holds in all of them, but each one on its own is less sure.
| Games where | Games | Record | Spread change | Margin change | Likely range |
|---|
| He played again that season | 119 | 43 - 75 - 1 | -2.1 | -1.6 | -4.9 to +1.8 |
| He did not play again that season | 107 | 32 - 75 | -3.6 | -3.0 | -6.2 to 0.0 |
| Team had a winning or even record going in | 73 | 29 - 43 - 1 | -2.3 | -2.2 | -6.1 to +1.8 |
| Team had a losing record going in | 153 | 46 - 107 | -3.0 | -2.2 | -5.3 to +0.5 |
Plain averages, as in the first table. The margin also jumps around by season: -6.3 in 2021, +0.3 in 2022, -1.0 in 2023 and -3.0 in 2024. Spread change: -3.6, -3.1, -2.5, -2.3. The spread is far steadier than the games.
What this
means.
Think 3 points for a starting QB. That is the typical cost, and close to what the spread does. On a good team, it can be about 5.
Don't expect much from one skill player. The market moves a little for a top receiver. The games don't show a clear change.
No promises. The closing spread already allowed for these absences, and it was about right. One game can go any way.
Who is out
this week.
Every NFL injury report, in one place. Free.
See NFL injuries
More on NFL at Bet Better: free NFL picks, NFL prop bets, NFL parlays, NFL odds, how to bet on NFL.
How we
measured.
The method
Games. NFL regular-season games from 2021 to 2024, weeks 1 to 17: 1,023 games. Week 18 and the playoffs are left out. We use no season after 2024 anywhere in this study.
Who counts as the starter. Only what was known before kickoff. He threw the most passes in his team's first game, and the most in more than half of its games so far. A team needs 3 games played first, so the earliest absence is its fourth game. Stats from the game itself play no part in picking him. A quarterback who took the job later in a season is not counted.
Who counts as missing. He has no line in the box score for that game. 238 games met this. We dropped 12 where he later played for another team that season, since that is a trade or a release. That left 226. We kept him out of both groups in the 37 games he played but was not the top passer, because he left hurt or was pulled.
Top receiver and lead back. The wide receiver with the most targets, or running back with the most carries, in the team's games so far, who played in more than half of them. A box score only shows a receiver who was thrown to, so a rare game with no targets can look like an absence. On 2024 snap counts, 1 of 53 receiver absences was a player who was on the field. None of the 61 quarterback or 41 running back absences were.
The comparison. Each team against its own games that season with that same player, 3 games at least. The adjusted figures come from a fit where every team season gets its own rating, plus home field and one number each for a missing QB, receiver and back. Ranges for those come from leaving out one team season at a time.
Closing lines. The closing spread and total in the nflverse game records.
Checks. We rebuilt the QB list in a separate database query, with no shared code. It found 228 games and a spread change of -2.9, against 226 and -2.8 here. The 2 extra games come from one tie for most passes in a Pittsburgh game. A second method fit team ratings only on games with both starters playing, then predicted the 184 games with one starter missing. It gave a margin change of -3.2 and a spread change of -3.4.
We looked at many groups, so one group clearing zero by a little could still be chance.
The data file
nfl-injuries-starters-2021-2024.csv has one row per missed game for each role (577 rows). Columns: date, season, week, team, opponent, home_away, position_role, player, points_for, points_against, closing_spread_for_team, closing_total, own_games_with, own_points_with, own_allowed_with, own_spread_with, came_back_this_season.
A negative spread means the team was favoured, as on a betting screen. The own_ columns are that team's averages with the player that season.
Cite this study
18+ (21+ in some US states). Betting is a cost, not an income. If gambling is causing you harm, call 1-800-GAMBLER
(US) or Gambling Help Online on 1800 858 858 (Australia).