This week’s power rankings include the major update of the methodology from a couple weeks back, plus a few tweaks each subsequent week. You’ll see that many of the visuals have changed, and I’m also directly comparing the implied Week 8 win probabilities and point spreads from my power rankings to what prediction markets and sportsbooks are projecting.

The major additions to the methodology are as follows:

• Better adjustments for non-quarterback injury clusters

• Better decay weighting on older results to improve predictive power

• Adjusting dropback efficiencies for PFF grading inconsistencies

• Using the Bayesian Quarterback Rankings to estimate quarterback change effects

• Daily updating of the paid subscriber Google sheet for injuries and other relevant news items

For more info, you can find the archive of previous Power Rankings posts here, which detail the previous methodology.

TEAM OFFENSIVE AND DEFENSIVE STRENGTH

This plot shows the SOS-adjusted performance for every team in their 2026 games. At the risk of confusing things further, I added arrows next to the team logos to show how they have moved week-over-week. The diagonals represent a 55/45 mix for offense/defense strength, so teams along the same diagonal are roughly equal in total strength.

Surprisingly to the average fan, the San Francisco 49ers didn’t move much after their loss to the Atlanta Falcons, maintaining the league’s best offensive performance in 2026, with a slightly above-average defense. The reason is that the adjusted scores model for the game had them as the better team, buried underneath a slew of turnovers that, hopefully, won’t become the norm going forward.

Other movements of top teams include the Chicago Bears falling somewhat in defensive performance after giving up 75th percentile EPA efficiency to the New England Patriots’ offense. However, they maintain positioning as the second-best team so far this year, with their power ranking (see below) a bit lower due to preseason priors and more uncertainty that Caleb Williams has sustainably “made the leap” in this breakout season. Williams is near the top of my adjusted efficiency numbers, but his Bayesian projection is a bit weaker on a lack of sample size.

The Bears are joined in a tight second-tier of teams with the Los Angeles Rams (better defense, worse offense) and have nearly the exact SOS-adjusted efficiencies this year as the Cincinnati Bengals and Jacksonville Jaguars. The Bengals got a big boost in offensive efficiency with their second-half throttling of the Baltimore Ravens, scoring five touchdowns and posting an elite 62% offensive success rate.

The biggest hit was taken by the Carolina Panthers, who fell more on offense than defense, which might seem strange after giving up 43 points to the Tampa Bay Buccaneers. But their game was out of hand so quickly that the Bucs offense didn’t overwhelmingly prove its superiority, getting scoring boosts from a fumble-6, a muffed punt recovery and an extra-point block and 2-point return.

The Kansas City Chiefs also dropped out of the second tier, although not as far as you might think after a pathetic-looking performance on Sunday Night Football. Playing against the Seattle Seahawks defense on the road is always challenging, so the effect on their offense was a bit muted.

In the lower half of the NFL, the Denver Broncos, New York Giants, Cleveland Browns, and New York Jets all moved up, mostly due to offensive success.

POWER RANKINGS VERSUS THE BETTING MARKETS

Let’s start with the power rankings (exact numbers updated daily for injury news on the subscriber Google sheet) before heading to the markets.

*Power rankings and “Model to Market” analysis updated as of 10/27, 10:00AM ET.

Read more

The adjusted scores quantify team play quality, with emphasis on stable metrics (success rate) and downplaying higher variance events (turnovers, special team, penalties, fumble luck, etc). Adjusted expected points added (EPA), in conjunction with opportunity-based metrics like total plays and drives, projects adjusted points. Adjusted scores have been tested against actual scores and offer slightly better predictive ability, though their primary benefit is explanatory.

All 2022-2026 and historical Adjusted Scores and other site metrics are available in a downloadable format to paid subscribers via Google Sheet.

Find previous advanced reviews here

Unexpected Points is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

** Adjusted Scores table:

• “Pass” - Pass rate over expectation (based on context of each play and historical averages

• “Success” - Success rate on offense, a key metric in adjusted score vs actual

• “H & A” - Home or away team

Eagles vs Cowboys (-4.5, 66%)

Strong win for the Dallas Cowboys on the road against what was supposed to be a good Philadelphia Eagles defense, putting up 416 yards (198 on designed runs) and 38 points. The adjusted scores had the Cowboys as the clearly better team, with a 17-point actual differential and 8-point adjusted differential, as the Cowboys offense was consistently better at making positive plays (54.5% success rate to 42.6%), plus winning the turnover and late-down advantages.

The game was fairly tight through the mid-third quarter, after the Eagles took the second-half kickoff and drove down the field over the course of 9 minutes and 16 took a one-point lead (21-20). The estimated win probability at that point was as low as 59.8% for the Cowboys, before they scored 18 points on their four second-half drives and pulled away.

Both teams performed well on late downs, especially so for the Cowboys, converting 8 of 13 third downs (+12.7 EPA) and their only fourth try (+2.0). The combination of third and fourth down the Eagles converted in normal play or via penalty totalled +13.3 EPA, helping keep them in the game, despite weaker efficiency on a per-play basis.

Read more

The adjusted scores quantify team play quality, with emphasis on stable metrics (success rate) and downplaying higher variance events (turnovers, special team, penalties, fumble luck, etc). Adjusted expected points added (EPA), in conjunction with opportunity-based metrics like total plays and drives, projects adjusted points. Adjusted scores have been tested against actual scores and offer slightly better predictive ability, though their primary benefit is explanatory.

All 2026-2022 and historical Adjusted Scores and other site metrics are available in a downloadable format to paid subscribers via Google Sheet.

Find previous advanced reviews here

Unexpected Points is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

** Adjusted Scores table:

• “Pass” - Pass rate over expectation (based on context of each play and historical averages

• “Success” - Success rate on offense, a key metric in adjusted score vs actual

• “H & A” - Home or away team

Seahawks (-3, 58%) vs Chiefs

The Seattle Seahawks got a dominant defensive performance in the win over the Kansas City Chiefs, with the help of incredible third-down conversion success and a couple of Chiefs’ missed field goals. Seattle never trailed, and its lowest WP was 52.8%, not far off from pre-game 58%.

The Seahawks defense held the Chiefs offense to 37% offensive success rate (excluding garbage time, < 1% WP), 4.3 yards per play, and 23rd percentile EPA efficiency.

The Seahawks went 11 of 16 on third down for an insane +18.9 EPA, while the Chiefs went 3 of 11 for -2.7, or net +22.6 advantage for the winners. Weirdly the EPA model gave the Seahawks -1.6 in value on their final, game-sealing 4th and 1 conversion, so that can be ignored.

Read more

The adjusted scores quantify team play quality, with emphasis on stable metrics (success rate) and downplaying higher variance events (turnovers, special team, penalties, fumble luck, etc). Adjusted expected points added (EPA), in conjunction with opportunity-based metrics like total plays and drives, projects adjusted points. Adjusted scores have been tested against actual scores and offer slightly better predictive ability, though their primary benefit is explanatory.

All 2025-2022 and historical Adjusted Scores and other site metrics are available in a downloadable format to paid subscribers via Google Sheet.

Find previous advanced reviews here

** Adjusted Scores table:

• “Pass” - Pass rate over expectation (based on context of each play and historical averages

• “Success” - Success rate on offense, a key metric in adjusted score vs actual

• “H & A” - Home or away team

Cardinals vs Broncos (-4.5, 67%)

The Denver Broncos won convincingly against the Arizona Cardinals, getting their offense going by EPA efficiency (+0.17 per play, 78th percentile), but only posting a 39% success rate (in non-garbage time, < 99% win probability). They were even better defensively, holding the Cardinals to a 33% success rate and only 4.7 yards per play. The Broncos never trailed with their lowest in-game WP at 61.1% (Q2 6:19), against 67.6% pregame.

Neither team was great on late downs, which contributed to the Broncos underperforming in EPA relative to their outstanding 7.0 yards per play (91st percentile).

Read more

The adjusted scores quantify team play quality, with emphasis on stable metrics (success rate) and downplaying higher variance events (turnovers, special team, penalties, fumble luck, etc). Adjusted expected points added (EPA), in conjunction with opportunity-based metrics like total plays and drives, projects adjusted points. Adjusted scores have been tested against actual scores and offer slightly better predictive ability, though their primary benefit is explanatory.

All 2025-2022 and historical Adjusted Scores and other site metrics are available in a downloadable format to paid subscribers via Google Sheet.

Find previous advanced reviews here

** Adjusted Scores table:

• “Pass” - Pass rate over expectation (based on context of each play and historical averages

• “Success” - Success rate on offense, a key metric in adjusted score vs actual

• “H & A” - Home or away team

Falcons vs 49ers (-2.5, 57%)

Big swing in the adjusted scores versus actual, with a nine-point differential in favor of the Atlanta Falcons moving to the San Francisco 49ers being three points better by the fundamentals. The 49ers had a materially better 55.2% offensive success rate (48.5% for the Falcons), but got crushed by four turnovers totaling 17.8 expected points lost (zero for the Falcons).

The 49ers’ win probability peaked at around 84.2% early in the second quarter, up 14-3. The 49ers turned the ball over on five of their next six possessions, if we include a turnover on downs, going for it on 4th-and-10, down 12 points on the ATL 23. Fourth down models had the win probability slightly leaning towards kicking it in that situation, which was ironic considering the fourth down discourse around Kyle Shanahan all week.

It was a pretty good offensive showing for the short-handed 49ers, who totaled substantially more yards at a more efficient clip, but had their dropback EPA efficiency crushed by turnovers.

The Falcons’ run game has been weak the last couple games, only averaging 3.3 YPC and losing 0.22 expected points per rush. The 49ers added +0.11 EPA per rush, despite losing 5.6 expected points to a Kaelon Black fumble, the second costliest play of the game.

The four biggest EPA and WPA swings in the game were mistakes for the 49ers, three turnovers and the 4th-and-10 failure.

Brock Purdy’s two interceptions were the game’s biggest win probability swings: -15.8% (Q3, Jessie Bates at the ATL 10) and -15.2% (Q4 2:42, Kendal Daniels, -5.7 EPA).

Drake London finished with seven catches for 66 yds and 1 TD on 10 targets, totaling +7.0 EPA.

Bijan Robinson was inefficient on the ground (63 rushing on 20 carries, -5.3 EPA), but caught all five of his targets for five catches and 45 yds (+2.5).

Christian McCaffrey left the game and came back later, limiting his work on the ground to seven carries for 39 and a score, adding six catches for 61 yds on seven targets. McCaffrey totaled +7.9 EPA, the best of any non-QB in the game. McCaffrey’s loss was acutely felt in light of the 49ers’ thin receiving group.

Kaelon Black ran fairly well, but his fumble more than offset the positives from 12 carries for 63 and a touchdown.

Brock Purdy, in a way, did well to finish with positive EPA after losing 15 expected points to turnovers and sacks. He had +17.2 pass EPA on non-INT pass attempts, with most of the value coming via air yards.

Michael Penix wasn’t special, but completely avoiding turnovers and sacks lifted his EPA efficiency to elite territory (+0.34 per play).

As of 5:13PM ET, the 49ers’ chance to win the NFC West in prediction markets was at 73%, down from as high as 79.6% leading into the game.

Titans vs Browns (-1.5, 54%)

Read more

The adjusted scores quantify team play quality, with emphasis on stable metrics (success rate) and downplaying higher variance events (turnovers, special team, penalties, fumble luck, etc). Adjusted expected points added (EPA), in conjunction with opportunity-based metrics like total plays and drives, projects adjusted points. Adjusted scores have been tested against actual scores and offer slightly better predictive ability, though their primary benefit is explanatory.

All 2022-2026 and historical Adjusted Scores and other site metrics are available in a downloadable format to paid subscribers via Google Sheet.

Find previous advanced reviews here

Unexpected Points is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

** Adjusted Scores table:

• “Pass” - Pass rate over expectation (based on context of each play and historical averages

• “Success” - Success rate on offense, a key metric in adjusted score vs actual

• “H & A” - Home or away team

Saints vs Steelers (-3, 59%)

The adjusted scores agreed with this outcome, but by a bit more narrow margin than the final score (3 vs 6 point differential). The Pittsburgh Steelers’ offense was more successful on a play-by-play basis, with a positive EPA rate at 45% to the New Orleans Saints at roughly 41%. The two big factors swinging the game in the Steelers’ favor were turnover differential (+1, +6.2 EPA) and a 6.8 expected points advantage on penalties and special teams, which includes a Saints missed field goal.

Pittsburgh never trailed, and its lowest win probability in prediction markets was 52.2%, with 8:22 left in the third quarter.

Late downs went both ways, netting a small advantage for the SAints. The Steelers converted 10 of 16 third downs, while the Saints went 6 of 12 on third down and 2 for 2 on fourth, including Shough’s 1-yard TD to Devaughn Vele (+2.5 EPA).

The Saints had 18 total rushing yards, only 12 when restricted to designed runs and excluding scrambles. This was their fewest in any single game over the last 10 seasons (163 games). The Steelers had a more balanced offense and better per-play efficiency.

Read more

THE CLAIM

Drake Maye had a turnover-plagued start to the season with six interceptions and a fumble lost in his first three games, but had seemed to calm things down, with only two turnovers in the next three games. But the ugly turnover bug came back last night, costing the New England Patriots an excellent opportunity in the Chicago Bears’ red zone.

The Ringer’s Sheil Kapadia points out that turnovers are the Achilles’ heel for an otherwise successful Patriots offense.

How worried should the Patriots be about Maye’s turnover woes, and it is simply part of his game to take elevated risks with the ball?

THE VERDICT: AN UNLUCKY GUNSLINGER

It’s true that turnovers have detracted from otherwise excellent play more for Drake Maye than any other quarterback in the NFL, and the Patriots have suffered as a consequence.

But Maye’s play in 2025 doesn’t show a reckless quarterback, with the notable exception of his struggles in the playoffs.

If anything, Maye has been unlucky to have as many turnovers as he’s had this year, meaning the natural regression in his play, combined with better luck, should elevate his impact back to one of the top quarterbacks in the NFL, benefiting the Patriots greatly.

THE EVIDENCE

Drake Maye is on an island if you plot the landscape of NFL quarterbacks in terms of non-turnover efficiency and turnover impact: He’s excellent in the former, horrible in the latter.

Maye ranks fourth in EPA per dropback (excluding garbage time) on non-turnover plays, trailing only Lamar Jackson, Brock Purdy and Josh Allen. But while nearly all of the other top quarterbacks by that measure also avoid turnovers, Maye has been the fourth-worst in turnover EPA per play. He’s all by himself in the plot above.

But that wasn’t always the case, as Maye’s 43.8 expected points lost to turnovers in 2026 is already approaching his 17-game total in 2025 (48.1).

If we look back at Maye’s 2025 regular season, his turnover rate per play was significantly lower (1.8% to 3.9%), averaging 1.3 interceptions per 100 dropbacks, which has jumped to 3.2 this year. While Maye struggled with turnovers in the playoffs last season, it was as much due to taking sacks and fumbles lost as to interceptions. Maye’s much larger history as a quarterback in 2025 shows that his elevated interceptions this year will likely regress to a “true” number closer to 2%.

Fumbles lost are the other part of the turnover equation, and Maye appears to be on a poor luck streak there also. Maye’s sack rate has improved from 7.8% (22nd best) in 2025 to 5.2% (12th) this year, but he has only one fewer fumble lost in seven games than all of last year, having taken 34 fewer sacks. Maye had an elevated fumble rate in the 2025 playoffs, but that corresponded with a leap in sack rate that we haven’t seen in 2026.

Read more

The adjusted scores quantify team play quality, with emphasis on stable metrics (success rate) and downplaying higher variance events (turnovers, special team, penalties, fumble luck, etc). Adjusted expected points added (EPA), in conjunction with opportunity-based metrics like total plays and drives, projects adjusted points. Adjusted scores have been tested against actual scores and offer slightly better predictive ability, though their primary benefit is explanatory.

All 2022-2026 and historical Adjusted Scores and other site metrics are available in a downloadable format to paid subscribers via Google Sheet.

Find previous advanced reviews here

Unexpected Points is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

** Adjusted Scores table:

• “Pass” - Pass rate over expectation (based on context of each play and historical averages

• “Success” - Success rate on offense, a key metric in adjusted score vs actual

• “H & A” - Home or away team

Bears (-3, 59%) vs Patriots

The Chicago Bears and New England Patriots finished with similar offensive success rates and EPA efficiencies, but the Bears enjoyed advantages in the three highlighted categories of high-variance play: +5.3 EPA in net turnovers (+1 differential), +9.2 EPA in net late-down value (3rd & 4th) and +2.5 EPA in special teams and penalties.

But adding the context of how the game unfolded, it appeared to be more of a dominant victory for the Bears, who never trailed and held a 20-7 lead with three seconds left in the first half (86.3% estimated win probability) before Rhamondre Stevenson took a short pass 48 yards for a shocking touchdown. That play kept the Patriots in the game (+14% win probability), but they never got above a 35% win probability in the second half.

The Bears’ two fourth down conversions were massively positive (+4.6 EPA), and sometimes these can go under the radar in the minds of observers, though they came on drives that only resulted in three total points (one missed field goal). For the Patriots, their initial drive ended on 4th-and-5 (-2.9 EPA), a conversion that surely would have put them in position to score points.

Read more

You can find all the previous editions of the Bayesian Quarterback Rankings here, and the backlog for Adjusted Quarterback Efficiency is here.

GO-FORWARD PROJECTED ADJUSTED EFFICIENCY

This update includes a pretty big re-work of the methodology, which projects more useful EPA per game numbers, along with details on exactly how those projections have moved from the prior week and through the season.

Additions to this update include:

• Removing the YAC and sack adjustments from the adjusted EPA data going into the model, with testing confirming those are less “luck” based than contextual information that enhances projection accuracy.

• Giving additional weight to “breakout” seasons (those occurring in years 2-4 +0.1 adjusted EPA higher than any previous season) to mitigate poor prior performance that’s less predictive.

• Establishing a “starter” designation that raises their draft prior to match franchise quarterbacks taken in that range, not backups filling in. The starter designation is phased in over the first 1,000 career snaps.

• “Stretching” (or shrinking) the projections to match the current EPA efficiency environment. We’re currently in a phase of enhanced efficiency since roughly the quarter-point of the 2025 season.

• Capping older season results at +/- 1.5 standard deviations to smooth outlier effects that are less meaningful to today’s projections.

• Lowering the weight of prior-team results to roughly 50% of current team.

These changes were all made through the iterative process of extensively testing ideas and measures at different levels and choosing the parameters that ended up with the most predictive power.

These rankings are for the go-forward projections of adjusted quarterback efficiency as of this week. They do not account for opponent strength or injuries to supporting cast, both of which go into the weekly Power Rankings.

Speaking of power rankings, this update to the Bayesian rankings and projections has been highly useful in determining quarterback adjustments when a team’s starter changes due to injury or benching.

In the tables below, I also included the adjusted EPA per play rankings for each quarterback over 2026 (min 100 play involvements) and the prior nine seasons (minimum 300 plays) so you can see the evidence going into the projections.

Older data is decayed over time, so the 2026 and 2025 EPA per play data matters more than those from prior years. That said, older data can’t be fully discounted, or else you miss bounce-back performers of great quarterbacks returning to form, like Aaron Rodgers returning to form after a few poor statistical seasons and winning his third and fourth MVPs in 2020 and 2021.

TIER 1

Well, we finally did it. The last update of the Bayesian rankings a couple weeks ago had only the tightest gap between Josh Allen and Patrick Mahomes, so it didn’t require a major rework to get Allen into the No. 1 spot. While some of the methodology changes did help Allen relative to Mahomes, he still would have been No. 2 to start the season either way, as you can see that his year-to-date projections lead is more than his current gap with Mahomes.

There’s probably less to contemplate with what could be off with Allen’s top ranking than Mahomes’ second-place position. Mahomes is only 14th in adjusted EPA this year, following multiple seasons that he didn’t crack elite status (top-8).

He’s still held up by the insanity of his earlier-career run, which combined having the top rankings with off-the-chart levels of efficiency versus the rest of the NFL. Has he been figured out? I’m not willing to say that yet, and the projections still see a material gap from Mahomes to the rest of the NFL.

Read more

In what’s turning into a lost season for the Indianapolis Colts, firing their GM Chris Ballard after falling to 2-4 in a loss to the lowly Tennessee Titans, the stats vultures have come out, pointing to the soul-crushing nature of the losses they’ve had:

Things are a bit more calm for the Baltimore Ravens, but they join the Colts with the dubious distinction of not just blowing leads, but blowing big leads. John Breech at CBS Sports highlights:

As for the Ravens, their loss to the Browns marked the 15th time since 2020 that they’ve blown a fourth-quarter lead of seven points or more. Not only is that the most blown leads in the NFL in that span, but no other team even has 10. Yikes.

What’s the likelihood that these franchises will start holding on in the fourth quarter?

THE VERDICT: PARTIALLY BUNK

The stats are all right and true that the Ravens and Colts have been awful at holding on to fourth-quarter leads, and that critique remains when adding the context of win-probability expectations, especially for the Ravens.

But the larger evidence of stickiness to teams’ abilities to play from ahead is fairly weak, so our projection for how often the Ravens and Colts will continue to suffer unlikely losses shouldn’t differ significantly from the rest of the NFL.

THE EVIDENCE

The first thing we have to do when taking any binary stat like “blown leads” is put it into context. Using win probability numbers that include betting market expectations (i.e., heavy favorites having a higher WP than underdogs at the equivalent game state: score differential, time remaining, etc.), it does confirm that the Ravens and Colts are giving up too many leads.

The plot below includes regular-season games from 2018 through 2026 where a team led at the start of any fourth-quarter play and then lost or tied. Expected blown leads is the sum of 1 minus Vegas win probability at each team’s first fourth-quarter snap with a lead. The model is well calibrated, with 565 actual blown leads league-wide against 557 expected.

Read more

This analysis is my best attempt to discount and adjust various elements of quarterback efficiency and get to the number that most accurately reflects fundamental play. There are a number of elements of quarterback efficiency - which also directly affect team performance - that are more dependent on luck (variance) than skill.

I aggregated many of those luck-based elements, with additional factors like passing scheme ability to generate yards-after-catch, in my Adjusted Quarterback Efficiency (AQE) metric.

The measures that I believe are most luck-based and part of this analysis:

• Interceptions: FTNData tracks “interception-worthy throws”, which I compare to actual interceptions on a play-by-play basis, and also adjust for expected interception return. Longer INT returns have a dramatic effect on the EPA, whether the quarterback’s fault or not.

• Drops: I calculated the expected drop rate for throws, based on location, and compared it to actual drops and determined the EPA lost/gained.

• Fumbles: Whether a fumble is recovered by the quarterback’s own team or not can turn a slightly negative play into a massive loss. I look at expectations for recovery based on different types of fumbles, and whether the quarterback himself recovers the fumble or a teammate (luckier).

• Yards After Catch: A higher portion of yards after catch should be credited to receivers or scheme than to the quarterback, at least in comparison to throws where the yards gained were mostly through the air. I adjust down EPA generated on passes with a relatively high proportion of YAC.

• Sacks caused/prevented by blocking and scheme: I introduced this element in 2024, using survival curve analysis that accounts for time-to-pressure, time-to-throw and contextual play elements from FTNData’s charting. Through this, I can divide sack credit/blame versus expectation into holding the ball too long (quarterback) and allowing pressure too quickly (offensive line). I back out the portions belonging to blocking.

• Defensive Pass Interference: There are so many underthrown balls that turn into big DPI gains that need to be recognized as partially luck. By their very nature, DPIs are not “open” receivers; the coverage defender is close enough to affect the receiver.

• Strength-of-Schedule: This is the one element that is most difficult to judge this early in the season. With a handful of games played, a great offense can have a bigger effect making the defenses they played look “bad”, and vice versa, than might be the truth. But it still matters.

• Weather: Based on expected EPA gains/losses versus average in different elements, based on wind, humidity and temperature.

If you want more details on many of the calculations, check out the Adjusted Quarterback Efficiency (AQE) primers from last season.

• AQE introduction

• Update to 2.0

• Update to 3.0

• Archive of past AQE posts, including my analysis of who deserved the 2023 MVP, the 2024 MVP and the 2025 MVP leveraging this metric

YTD 2026 ADJUSTED EFFICIENCY RESULTS

First, here’s the chart for sack adjustments, with particular note that my model estimates holding onto the ball too long (i.e., quarterback) is responsible for most of the sack rate over expected, and time-to-pressure (i.e., blocking) a lower amount. Sacks are a (mostly) quarterback stat.

The next plot below shows each quarterback who has been involved in at least 90 plays this season (dropbacks plus designed runs). There are two points for each quarterback: 1) The team-colored dot for the actual EPA per play the quarterback has this season and 2) The quarterback headshot representing the adjusted quarterback efficiency (AQE, Adjusted EPA/play). There is also a team-colored line linking the two on each row.

Josh Allen trails Lamar Jackson and Brock Purdy in raw EPA/play, but suffers lower negative adjustments than both. You can see in the table below that Allen leads the NFL in total raw and adjusted EPA by roughly 10 EPA in both categories. His usage has been a bit higher than Purdy’s, with 37 more plays through six games. But Purdy’s usage has ticked up this year, and Jackson has missed enough time to make his per-play numbers not truly comparable to other starting quarterbacks.

C.J. Stroud had the strongest week for adjusted EPA (+16.2), followed by Drake Maye (+13.7). Stroud’s seasonal number takes a bit of a hit from his weaker strength of schedule, plus his receivers have had fewer than expected drops. Maye is on a real tear the last few weeks, moving from being seen as a “disappointment” three games into the year, all the way to third in AQE following three more.

The next two best Week 6s came from Trevor Lawrence (+12.3) and Jordan Love (+11.3), with the former an AQE darling for many seasons, but never finishing in the top-5 where he currently resides. Love has been on the long march up the rankings for the last few weeks, now just outside the top-12. It’s funny how early season performance and perceptions can have Love with a meh year so far and someone like Bryce Young “breaking out” with the same efficiency. Though I should acknowledge that Young has a better PFF grade, which likely reflects how they’re viewed.

Sam Darnold has gotten a lot of grief for the Seattle Seahawks’ disappointing start. But he’s only been mediocre by AQE, excluding the effects of poor luck that have driven his production down further. Following the TNF loss last week, I went into why it would be a big mistake for the Seahawks to turn to Drew Lock.

The worst performances of the week were by Bo Nix (-10.3), Jayden Daniels (-8.7), Daniel Jones (-8.5) and Michael Penix (-8.0). Penix has tumbled from having the same AQE as the leaders after his first two starts to the middle of the NFL.

Last week, I advocated buying Josh Allen at roughly 26% win chance on prediction markets, when Brock Purdy was at 30%. Now they’re in a virtual tie around 30% after a better performance by Allen, and I’d still lean towards him being the better bet. That said, the Bills defense is not great, coming in at 15th in my Week 7 Power Rankings. But the 49ers defense ranks even worse at 17th.

Looking at longer-odds quarterbacks, the obvious +EV choice for me

Read more

This week’s power rankings include the major update of the methodology from last week, plus a few tweaks this week. You’ll see that many of the visuals have changed, and I’m also directly comparing the implied Week 7 win probabilities and point spreads from my power rankings to what prediction markets and sportsbooks are projecting.

The major additions to the methodology are as follows:

• Better adjustments for non-quarterback injury clusters

• Better decay weighting on older results to improve predictive power

• Adjusting dropback efficiencies for PFF grading inconsistencies

• Using the Bayesian Quarterback Rankings to estimate quarterback change effects

• Daily updating of the paid subscriber Google sheet for injuries and other relevant news items

For more info, you can find the archive of previous Power Rankings posts here, which detail the previous methodology.

TEAM OFFENSIVE AND DEFENSIVE STRENGTH

This plot shows the SOS-adjusted performance for every team in their 2026 games. At the risk of confusing things further, I added arrows next to the team logos to show how they have moved week-over-week. The diagonals represent a 55/45 mix for offense/defense strength, so teams along the same diagonal are roughly equal in total strength.

The week-over-week movements are much smaller here than we saw post-Week 5, especially in the top tiers of 2026 performers. Remember: this is the tiering for 2026 performance only, and the relative positioning is not equivalent to the power rankings you’ll find below.

Of the top teams, the Chicago Bears had the biggest jump, getting slightly better offensively, and receiving a more substantial upgrade on defense. The final point differential over the Atlanta Falcons was only eight points, but the adjusted scores saw the performance as much better (14 adjusted-point differential), which is also why the Falcons tumbled offensively into the bottom-4 (that includes Cooper Rush starts, which are fully adjusted out of the power rankings).

The Baltimore Ravens’ defense took a hit for not beating up more on an awful Cleveland Browns offense. And the Houston Texans defense got crushed after the shootout in London.

I guess the Dallas Cowboys should still be favored in an NFC East that some team has to win. But their SNF showing against the Green Bay Packers was poor enough offensively to drop them into the bottom-half of the NFL in 2026 performance.

POWER RANKINGS VERSUS THE BETTING MARKETS

Let’s start with the power rankings (exact numbers updated daily for injury news on the subscriber Google sheet) before heading to the markets.

*Power rankings and “Model to Market” analysis updated as of 10/24, 6:15pm. Unfortunately, my special teams adjusted wasn’t turned on for the initial Tuesday run, but it’s in there now. There are only minor changes points-wise, but it flips the top two teams.

Assumptions built into the rankings are that Baker Mayfield is in and Lamar Jackson is out.

Read more

The adjusted scores quantify team play quality, with emphasis on stable metrics (success rate) and downplaying higher variance events (turnovers, special team, penalties, fumble luck, etc). Adjusted expected points added (EPA), in conjunction with opportunity-based metrics like total plays and drives, projects adjusted points. Adjusted scores have been tested against actual scores and offer slightly better predictive ability, though their primary benefit is explanatory.

All 2026-2022 and historical Adjusted Scores and other site metrics are available in a downloadable format to paid subscribers via Google Sheet.

Find previous advanced reviews here

Unexpected Points is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

** Adjusted Scores table:

• “Pass” - Pass rate over expectation (based on context of each play and historical averages

• “Success” - Success rate on offense, a key metric in adjusted score vs actual

• “H & A” - Home or away team

49ers (-7.5, 76%) vs Commanders

The San Francisco 49ers held on to win at home against the Washington Commanders, a result that enhances the idea that they aren’t as strong as their perfect 6-0 record. The adjusted scores had them as the better team, but only by four points. The 49ers offense was moribund down the stretch, going 3-and-out on their final three possessions, opening the door for the Commanders multiple times.

While I didn’t write it on the initial Tuesday release of the Week 6 Power Rankings, the Commanders did become a play by the end of the week, with the Sunday update placing them at a 27.7% win probability in the model versus 22% in the market. Unfortunately you can’t win them all, and this was one of five 5%+ advantages that didn’t pay off, though there was some hope when the Commanders drove into the red zone with a few minutes remaining down six points.

But the reality is that the game never got that close in prediction market win probability, with the Commanders peaking at 31.2% on that drive. This is in contrast to the nflfastR win probability model that only had them as high as 27.3% (down six points with 2:26 left, 2nd & 3 at the SF 17). Modeled win probabilities tend to overvalue the leading team in these situations versus the market, especially if they’re substantial favorites going into the game.

Read more

Kenneth Walker’s 665 rushing yards through five games trail only four of the 20 best rushing seasons ever at the same point:

• O.J. Simpson 1973 (813)

• Jim Brown 1963 (787)

• Jamal Lewis 2003 (742)

• DeMarco Murray 2014 (670)

Walker is ahead of Eric Dickerson’s record 1984 season that totaled 2,105 yards. Dickerson had just 498 yards through five games.

What you notice about Dickerson’s record-breaking season is the increase in trajectory starting in his 10th game. Dickerson had rushed for only 963 yards through the first nine games of the season (107 per game, pace of 1,712 yards), including games of 38 (Week 9) and 49 yards (Week 3). He then went into overdrive the final seven games of the year, rushing for 1,142 yards (163 yards per game), hitting 200+ yards twice (Weeks 10 and 15), plus 191 yards on another occasion (Week 13).

Dickerson not only did this in a shorter 16-game season, but there weren’t bye weeks for him to recuperate. Though it appears he didn’t need the rest as much as the defenses he faced did.

The 1984 campaign was Dickerson’s second NFL season, coming on the heels of his rookie year with 1,808 rushing yards, enough to keep Dan Marino from winning Offensive Rookie of the Year.

Looking specifically at the seasons above Walker’s for rushing yards through five games, they all ended up finishing strong, except for DeMarco Murray in 2014. Murray was ahead of Dickerson’s pace through nine games, but failed to reach 100 yards in three of his last five games.

The 14-gamers Jim Brown and O.J. Simpson maintained their pace throughout, with Simpson squeezing a lot of juice (pun intended) out of the final two weeks, rushing for 200+ yards both games (419 total), breaking the 2,000-yard barrier (2,003). Brown’s pace faded a bit post-Week 10, but his per-game pace was higher through 14 games than for Dickerson’s 16.

Read more

The adjusted scores quantify team play quality, with emphasis on stable metrics (success rate) and downplaying higher variance events (turnovers, special team, penalties, fumble luck, etc). Adjusted expected points added (EPA), in conjunction with opportunity-based metrics like total plays and drives, projects adjusted points. Adjusted scores have been tested against actual scores and offer slightly better predictive ability, though their primary benefit is explanatory.

All 2026-2022 and historical Adjusted Scores and other site metrics are available in a downloadable format to paid subscribers via Google Sheet.

Find previous advanced reviews here

Unexpected Points is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

** Adjusted Scores table:

• “Pass” - Pass rate over expectation (based on context of each play and historical averages

• “Success” - Success rate on offense, a key metric in adjusted score vs actual

• “H & A” - Home or away team

Packers vs Cowboys (-1.5, 51%)

The Green Bay Packers won comfortably at home against the Dallas Cowboys on Sunday Night Football, enjoying a massive +21.8 EPA differential on late downs to get the win. Both teams posted similar, and slightly poor, offensive success rates, and averaged only around 4.5 yards per play.

The Packers went 11-for-18 on third downs (+11.4 EPA) and converted their one 4th & 3 (+3.6), while the Cowboys were 5-for-13 (-6.7) on third downs and netted basically flat EPA on their four fourth down tries.

Read more

The adjusted scores quantify team play quality, with emphasis on stable metrics (success rate) and downplaying higher variance events (turnovers, special team, penalties, fumble luck, etc). Adjusted expected points added (EPA), in conjunction with opportunity-based metrics like total plays and drives, projects adjusted points. Adjusted scores have been tested against actual scores and offer slightly better predictive ability, though their primary benefit is explanatory.

All 2025-2022 and historical Adjusted Scores and other site metrics are available in a downloadable format to paid subscribers via Google Sheet.

Find previous advanced reviews here

** Adjusted Scores table:

• “Pass” - Pass rate over expectation (based on context of each play and historical averages

• “Success” - Success rate on offense, a key metric in adjusted score vs actual

• “H & A” - Home or away team

Rams (-12.5, 87%) vs Cardinals

The Los Angeles Rams were, by far, the biggest favorites of the week, yet adjusted scores didn’t see the Arizona Cardinals as lucky winners. The Cardinals were in the game all the way, mostly playing either with the lead or in a tied contest; they had a slightly better offensive success rate, though both teams underperformed in EPA efficiency.

The Cardinals were values according to my power rankings, especially if you consider it proportionally: 19% projected win probability versus the market at 13%.

Read more

The adjusted scores quantify team play quality, with emphasis on stable metrics (success rate) and downplaying higher variance events (turnovers, special team, penalties, fumble luck, etc). Adjusted expected points added (EPA), in conjunction with opportunity-based metrics like total plays and drives, projects adjusted points. Adjusted scores have been tested against actual scores and offer slightly better predictive ability, though their primary benefit is explanatory.

All 2025-2022 and historical Adjusted Scores and other site metrics are available in a downloadable format to paid subscribers via Google Sheet.

Find previous advanced reviews here

** Adjusted Scores table:

• “Pass” - Pass rate over expectation (based on context of each play and historical averages

• “Success” - Success rate on offense, a key metric in adjusted score vs actual

• “H & A” - Home or away team

Browns vs Ravens (-3, 61%)

The adjusted scores agree that the Cleveland Browns deserved the win, though the Baltimore Ravens appeared in control for most of the game. The Ravens’ advantage on late downs (+7.7 EPA differential on third, +1.8 on fourth) was partially offset by a huge Derrick Henry fumble lost (-5.0), the only turnover of the game.

The Unexpected Points Power Rankings showed the strongest edge of the week on the Browns, with a projected 48% win probability versus prediction markets at roughly 39%. It appears my numbers were lower on Tyler Huntley, who probably played above his actual ability in an overtime win against the Atlanta Falcons last week.

The Ravens began the game with a touchdown on 4th & Goal from the 2-yard line, then a field goal from only 34 yards out. Their remaining possessions went punt, punt, punt, punt, punt, missed 49-yard field goal, fumble lost, punt and punt.

This was the Ravens’ worst offensive game of the season by EPA efficiency (29th percentile), with horrendous rushing losses (-10.8 EPA), led by the previously mentioned Henry fumble.

The team box score stats were nearly matching in yards (237 Browns, 235 Ravens), first downs (14, 13), and yards per play (4.0, 4.1).

The biggest play of the game by win probability added was a 33-yard catch by Jerry Jeudy that set up the game-winning field goal.

There wasn’t a ton of offensive production to go around, so Jeudy played well to generate +5.9 EPA on only four targets.

You’ll find more green and red on the Ravens’ table, though star players like Zay Flowers and Derrick Henry severely underperformed expectations.

Tyler Huntley had the better numbers versus Deshaun Watson, though neither quarterback put much into the box score. They were able to offset massive sack losses (eight taken combined) with some rushing success, although the EPA gains on those 13 carries netted to only marginal gains.

The markets aren’t buying the 4-2 Browns (-16 point differential), though their chance to win the division rose from around 5% pre-game to 11%. The Ravens took a tumble from 61% to 46%.

Colts (-6.5, 72%) vs Titans

Read more

The adjusted scores quantify team play quality, with emphasis on stable metrics (success rate) and downplaying higher variance events (turnovers, special team, penalties, fumble luck, etc). Adjusted expected points added (EPA), in conjunction with opportunity-based metrics like total plays and drives, projects adjusted points. Adjusted scores have been tested against actual scores and offer slightly better predictive ability, though their primary benefit is explanatory.

All 2022-2026 and historical Adjusted Scores and other site metrics are available in a downloadable format to paid subscribers via Google Sheet.

Find previous advanced reviews here

Unexpected Points is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

** Adjusted Scores table:

• “Pass” - Pass rate over expectation (based on context of each play and historical averages

• “Success” - Success rate on offense, a key metric in adjusted score vs actual

• “H & A” - Home or away team

Jaguars (-2.5, 56%) vs Texans in London

The adjusted scores see this as a bit of a lucky win for the Houston Texans over the Jacksonville Jaguars, who had material home-field advantage playing in London for the second straight week. The Unexpected Points Power Rankings showed a modest edge on the Texans as of Sunday morning, with a projected 48.1% win probability versus prediction markets at roughly 44%.

I think the adjusted scores might be taking away too much credit from the Texans. This was a game with both defenses getting carved up, and the Texans outgained the Jags 7.2 yards per play to 6.1 on the back of more explosive plays.

Both teams posted strong offensive success rates and EPA efficiencies, but the Texans did a better job maximizing their successes, gaining +5.8 EPA on late downs (5-for-11), versus the Jags at -3.1 EPA (7-for-13), with most of their loss coming on a failed fourth down try. Plus, the Texans got two conversions via penalty (+5.1 EPA), including a defensive too many men on the field on 4th & 1.

Read more