Week 12, Thanksgiving Day: Advanced Reviews

Three games that weren't very close in the second half, though one winner we weren't expecting

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 2023 & 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

DET vs GB

The Packers pull off an unexpected win in Detroit. The Lions were 7.5-point favorites going into the game (implied win probability at 77%), yet they gave up a touchdown less than three minutes into the game and trailed the entire rest of the way often by multiple scores.

The adjusted scores have the game a bit closer than the final score, and especially closer than multiple-touchdown lead the Packers maintained until the final minute of the game.

The Lions were killed by turnovers and inability to covert late downs, including going 1-for-5 on fourth down, netting a loss of 10.8 expected points on those attempts. Dan Campbell’s fourth down decisions will get a lot of scrutiny after the game, as typical results-based analysis becomes the norm. All of the decisions were favorable according to the nfl4thR model, though only one netted more than 1% in pre-play win probability.

The turnover disadvantage for the Lions was all about Jared Goff’s inability to hold onto the ball. He lost three fumbles for a total of 15.6 expected points lost, including the most impactful play of the game, by far: a fumble-6 that cost the Lions 7.4 expected points and 11.9% win probability.

The rest of the most impactful plays of the game are littered with Lions failures on fourth down, plus a Packers 4th & 1 try and was stuffed at the line.

Where does everyone stand on Jordan Love? I’ve been leaning against the wind most of the season, with great skepticism around the sustainability of his early season results (while Packers fans were anointing him their next elite quarterback). After a series of disappointing team results, I was more positive than the crowd about his chances.

Now Love’s numbers and the Packers results have switched back to positive, winning three of the last four games. Post-Thursday night, Love sits at 11th in EPA per play, above Lamar Jackson and Trevor Lawrence. I’m still not sold on him being the answer, but he’s playing well enough to demand more chances to prove he could be it.

It’s nearly impossible for any quarterback to overcome the turnovers and lack of fourth down success Goff had in this game, not that they weren’t partially caused by his carelessness and lack of execution. Ending up at -0.14 EPA per play isn’t bad in that context, and a at least shows that the Lions offense can produce when not shooting itself in the foot. The Lions odds to win the NFC via the No. 1 seed took a severe blow with the Eagles win on Monday night and this loss. Despite being firmly in the second tier of non-49ers teams, I have them only around a 10% chance to go to the Super Bowl.

DAL vs WAS

It’s crazy how much the Cowboys pile on at the end of games to make their unadjusted point differentials so high. A record fifth pick-6 for DaRon Bland put the Cowboys up by the final score, after the Commanders hung around the first half, and were still within two scores into the fourth quarter.

Both teams derived a lot of value on third down conversions (+10.1 EPA for the Cowboys, +4.8 EPA for the Commanders), but the Commanders were 0-for-3 on fourth downs, losing 11.1 expected points.

At a 46.8% success rate, the Commanders offense was decently good on a play-by-play basis, yet the inability to convert both of their 4th & 1 chances when the game was still contested killed their chances to score more points. The largely meaningless - in term of win probability - pick-6 to end the game was, by far, the most impactful play in expected points, costing the Commanders 9.5.

At times watching the game it felt like Tony Romo and Jim Nantz were on Sam Howell’s PR team with their constant praise and assumption that the quarterback currently 18th in EPA per play was, indeed, the long-term answer for the Commanders. I think he still has a lot to prove, and their perception that his potentially record-breaking sack rate is mostly a blocking issue is probably off. That said, I think Howell does have something when it comes to making plays, and I respect his willingness to risk completely go down in flames and hurt his numbers at the end of almost totally unwinnable games. I’d be happy to have him compete for the 2024 starting job if I was running the Commanders, but I wouldn’t take myself out of the quarterback market next spring’s draft.

Another quiet, but deadly efficient game from Dak Prescott. Prescott now leads all non-Brock Purdy quarterbacks in efficiency and total EPA generated. I’m not sure what he needs to do to get more MVP hype (currently fifth in odds), other than get his defense to play a little worse and make his touchdown drives seem more meaningful.

SEA vs SF

This game got strangely “lose” in the second half by the scoreboard (11-point 49ers lead into the fourth quarter), but the Seahawks never got close to even 10% win probability by the numbers.

A horrific start to the game was too much for the Seahawks to overcome. They had zero net offensive yards in the first quarter, and trailed 24-3 at the halftime break. Those three points were only via a 66-yard Dee Eskridge kickoff return. The Seahawks didn’t convert a first-down until 12 seconds left in the first half.

The 49ers rushing efficiency deserves special attention in this game, generating 0.23 EPA per play, which is equivalent to an MVP-level passing game. A good omen for their playoff run for the 49ers to get that part of their offense back on track, after posting very negative rushing efficiency in three of the previous four games.

The only thing that kept the score optically close in the second half was a Brock Purdy pick-6, which was thrown off the frame of Christian McCaffrey, but still somewhat unlucky to end up as a tipped pick-6, the worst possible results (-5.8 expected points).

It wasn’t a great effort for Purdy, but weirdly his final 28-yard touchdown to Brandon Aiyuk could make this one of the better games in the minds of observers. At least I saw a ton of praise for it online, which reenforces and sometimes dictates what people think the next day. Purdy won’t be caught in efficiency this season, so he needs perceptions to match the numbers to get more respect.

Tough assignment for Geno Smith, but you’d like to see better than bottom-10th percentile efficiency even in tough situations. It’s similar to Smith’s collapse in Baltimore: the quarterback has to be able to adjust to poor circumstances and pressure to give his team a chance to win. Smith didn’t do that, failing to provide consistent positive plays (37.8% dropback success rate) while taking a ton of negative plays, including a 1st & 10 interception (-4.7 EPA) and six sacks for 48 yards (-11.6 EPA).

We can’t be sure how much Smith’s injury the week before affected his performance, being on the field didn’t give his team a better chance to win yesterday. Right now Smith ranks 20th in EPA per play, only slightly higher than quarterbacks who have been benched. He’s playing as a low-level starter right now, which isn’t good enough for the Seahawks to make a playoff run. The markets have their playoffs odds as a coin-flip, despite their 6-5 record. They face the Cowboys, 49ers (again) and Eagles the next three weeks.