Which NFL records are real?
Before Week 8: close wins have carried the Browns and Broncos to 5-2, and the 4-3 Seahawks are a little better than their record. From here, all three project about the same.
Cleveland and Denver have records their scoring doesn't support
All 32 teams as a table
| Team | Record | Point diff | One-score | Adj. per-play rank |
|---|---|---|---|---|
 Vikings | 6-0 | +51 | 4-0 | 13th |
 49ers | 6-1 | +58 | 3-0 | 2nd |
 Chiefs | 5-1 | +40 | 2-0 | 4th |
 Bills | 5-1 | +26 | 4-1 | 11th |
 Bears | 5-2 | +37 | 2-1 | 1st |
 Steelers | 5-2 | +15 | 4-1 | 14th |
 Broncos | 5-2 | +3 | 4-0 | 16th |
 Browns | 5-2 | -13 | 5-1 | 19th |
 Jaguars | 4-2 | +50 | 2-2 | 8th |
 Bengals | 4-2 | +41 | 2-2 | 3rd |
 Lions | 4-2 | +19 | 3-1 | 17th |
 Packers | 5-3 | +7 | 3-1 | 5th |
 Seahawks | 4-3 | +43 | 2-3 | 9th |
 Patriots | 4-3 | +17 | 2-2 | 7th |
 Ravens | 4-3 | +2 | 3-2 | 18th |
 Giants | 4-3 | +2 | 3-2 | 28th |
 Rams | 3-4 | +11 | 1-3 | 10th |
 Cowboys | 3-4 | +9 | 1-3 | 25th |
 Falcons | 3-4 | -2 | 0-3 | 6th |
 Raiders | 3-4 | -14 | 1-2 | 21st |
 Panthers | 3-4 | -15 | 1-2 | 15th |
 Buccaneers | 2-5 | +11 | 1-5 | 12th |
 Texans | 2-5 | -5 | 1-4 | 22nd |
 Jets | 2-5 | -13 | 1-4 | 26th |
 Colts | 2-5 | -26 | 1-3 | 20th |
 Saints | 2-5 | -32 | 2-3 | 23rd |
 Eagles | 2-5 | -48 | 2-2 | 27th |
 Cardinals | 2-5 | -66 | 1-1 | 29th |
 Commanders | 1-5 | -40 | 1-3 | 24th |
 Dolphins | 1-5 | -59 | 1-2 | 31st |
 Titans | 1-6 | -48 | 1-4 | 32nd |
 Chargers | 0-6 | -61 | 0-3 | 30th |
The Browns are 5-2 and have been outscored by 13 points. The Broncos are 5-2 at +3. Together they are 9-1 in one-score games, and that is the whole gap in the chart.
The Buccaneers are the mirror image: 2-5 while outscoring opponents by 11.
Every Browns win came by 4 points or fewer
The margins explain the records. Colored bars are one-score wins; gray bars are losses.
Every game's final margin, 2026
Browns5-2 · one-score 5-1Broncos5-2 · one-score 4-0Seahawks4-3 · one-score 2-3Buccaneers2-5 · one-score 1-5- Browns: won by 4, 3, 3, 3 and 3. Lost by 24 and 5.
- Broncos: 4-0 in one-score games, plus a 14-point win. Lost by 21 and 10.
- Seahawks: three losses by a combined 8 points; two wins by 17 or more.
- Buccaneers: five losses by 7 or fewer. Their two wins were by 8 and 30.
Cleveland turns it on late, and two quarterbacks have dodged picks
Two things that rarely last have propped up the 5-2 teams: late-game play and interception luck. Deshaun Watson and Bo Nix have each made 12 turnover-worthy plays (PFF's name for throws that should have been picked) but thrown only 3 interceptions. About half usually are.
Turnover-worthy plays vs interceptions, 2026 quarterbacks
- Browns: in close fourth quarters (win probability 20% to 80%), Cleveland's offense gains +0.13 EPA per play over 98 plays, against −0.15 otherwise. EPA, expected points added, is how much a play changes a team's expected score.
- Watson's legs: 10 runs in close fourth quarters for 10.4 yards a carry, and no Cleveland turnovers in those spots.
- Buccaneers, the reverse: they pass 71% of the time in close fourth quarters, with 2 turnovers and 4 sacks. A small sample (53 plays), but it fits their 1-5 one-score record.
- Seattle got the bad end: Sam Darnold has 8 interceptions on 11 turnover-worthy plays, about 2 more than the league rate.
- Neither carries over: early close-fourth-quarter EPA had a 0.11 correlation with the rest of the season (158 team seasons since 2019); a quarterback's interception rate barely predicts his next season (0.07).
Team by team: the players and injuries behind it
- Running back Quinshon Judkins averages 3.5 yards a carry, so the late run game is mostly Watson.
- Guard Teven Jenkins is on injured reserve (back).
20.2%Nik Bonitto's pass-rush win rate, 2nd in the NFL (PFF)- Adjusted for opponents, the offense ranks 17th and the defense 12th.
- Running back RJ Harvey is on injured reserve (knee).
- Two losses turned on fourth-quarter Darnold interceptions, one a pick-six against Washington.
- The caveat: the edge is all defense, which carries over less than offense (a 0.32 correlation with the rest of the season, against 0.53).
- Left tackle Charles Cross (ankle) and receiver Rashid Shaheed (knee) are out for Monday's game against Chicago.
- Baker Mayfield has been about average (−0.01 EPA per dropback); Daniels, who started Weeks 4 to 6, sits at −0.13.
- Bucky Irving leads a run game worth +0.21 EPA per carry.
- Receiver Emeka Egbuka is out today (foot).
Close-game records mostly even out
Since 2016, teams that won 80%+ of their early one-score games won just 57% of them afterward. Teams that lost most of theirs bounced back to 52%.
One-score game win rate, first 7 games vs the rest of the season
Per play, Seattle ranks 9th and Cleveland 19th
Strip out garbage time, adjust for opponents and weight every play, and the NFL sorts itself out.
2026 NFL teams ranked by adjusted net EPA per play
Bears49ersBengalsChiefsPackersFalconsPatriotsJaguarsSeahawksRamsBillsBuccaneersVikingsSteelersPanthersBroncosLionsRavensBrownsColtsRaidersTexansSaintsCommandersCowboysJetsEaglesGiantsCardinalsChargersDolphinsTitans- Seahawks, 9th: the NFL's best defense, held back by the 26th-ranked offense.
- Buccaneers, 12th: the 3rd-ranked defense, the 25th-ranked offense.
- Vikings, 13th: 6-0 on the No. 2 defense and the 27th-ranked offense.
- Broncos, 16th: average, against one of the NFL's toughest schedules.
- Browns, 19th: the 23rd-ranked offense. Unadjusted they rank 11th; an easy schedule flatters them.
Most analysts see the Browns. Fewer see Tampa Bay.
- RG.org (Oct. 26) calls Cleveland "nearly two wins better than its point differential supports" and gives Tampa Bay the league's biggest shortfall.
- ESPN's Daniel Oyefusi (Oct. 26): the Browns' minus-13 differential is "the second lowest by a team to start 5-2 or better over the past decade."
- NFL.com's Week 8 power rankings have the Seahawks 4th, Broncos 8th, Browns 12th and Buccaneers 19th.
What the data adds: Cleveland's late wins run through Watson's legs and Tampa Bay's late losses through its passing game. Per play the Bucs rank 12th, well above NFL.com's 19th; Seattle (9th) sits below its 4th.
Seattle and the 5-2 teams project about the same
Both 5-2 teams should still finish around 10 or 11 wins. Seattle plays better than its record, but a 4-3 start counts too: the model has the Seahawks winning 54% of the rest, level with Denver (55%) and Cleveland (56%).
Projected win % for the rest of the season
Vikings6-0 nowBrowns5-2 nowSteelers5-2 nowBroncos5-2 nowSeahawks4-3 nowBuccaneers2-5 nowThe model doesn't adjust for schedule. After adjusting for opponents, Denver (16th) rates a little above Cleveland (19th), so of the two 5-2 teams, lean Denver.
Questions people ask
Are the Browns good in 2026?
Their record is ahead of their play. Cleveland is 5-2 but has been outscored by 13, is 5-1 in one-score games and ranks 19th per play after adjusting for opponents and garbage time.
Are the Broncos for real?
Partly. Denver is 4-0 in one-score games and +5 in turnover margin, and ranks 16th per play after one of the NFL's toughest schedules.
Which NFL teams are better than their record in 2026?
The Buccaneers most clearly: 2-5, 12th per play and 1-5 in one-score games. The Seahawks (4-3, 9th per play) are a little better than their record but project about the same as the 5-2 teams.
Is winning close games a skill in the NFL?
Mostly not. Since 2016, teams that won 80%+ of their early one-score games won 57% of the rest.
How we did this: Snapp's NFL play-by-play warehouse, ESPN win probability, PFF and injury reports, regular-season games through Oct. 29, 2026, before Week 8. Garbage time is removed by keeping plays with pre-snap win probability between 10% and 90% (the previous play's win probability). Per-play ratings weight every play, adjust for each opponent's average, and count interception and fumble return touchdowns against the offense that gave them up (the source data had scored them in that offense's favor). Close fourth quarters: pre-snap win probability between 20% and 80%. The projection regresses rest-of-season win rate on record and unadjusted net EPA per play (same garbage-time filter) over each team's first seven games, fit on 320 team seasons from 2016 to 2025; its typical miss is ±20 percentage points of win rate. One-score game = decided by 8 or fewer.