Kickoff TNF Chiefs-Lions Advanced Review

A messy game where the absence of Travis Kelce was felt even more than expected

The Lions shock and banged up Chiefs team in Kansas City

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 and 2023 and historical Adjusted Scores and other site metrics are available in a downloadable format to paid subscribers via Google Sheet.

** For this initial run of the season there could be some issues with the data. Everything is pulled from the NFL website, which has yearly tweaks that need to be fixed. I manually changed a few things from the nflreadR data, but could have missed other stuff. Everything will be resolved going forward as the fixes are identified and codified.

I assumed it was going to be tougher sledding for the Kansas City Chiefs on offense without all-world tight end Travis Kelce, but the impact of that storyline was even more acute than high expectations. Part of the reason that Kelce’s absence was so noticeable was that all-world quarterback Patrick Mahomes pretty much did everything of which he was asked, i.e. carry the offense.

In a game that logically set up as a rare opportunity for the Chiefs to lean, at least somewhat, on their run game against a previously porous Detroit Lions run defense, they couldn’t get anything going on the ground. Instead, the Chiefs passed the ball at a rate 15.3% over expectation, as their non-Mahomes ground game generated 45 yards on 17 carries (2.6 YPC). The Lions, on the other hand, were able to lean on their ground game, with 177 yards on 29 carries (4.0 YPC), including grinding away the end of the game with a final first down.

The adjusted scores give the Chiefs a slight edge versus the actual score, but there are many factors pushing in either direction. The Chiefs offense played at a higher success rate, and threw at a significantly higher rate - two factors that normally equal an easy win. But the most impactful play in terms of EPA went against the Chiefs (Kardarius Toney drop to pick-6), costing them 8.5 points. I discount turnovers based on the inherent luck factor, but that one had to be one of the unluckiest possible, at least for the quarterback.

I’ve been an anti-Toney truther for a couple seasons now, but I still didn’t see this performance coming. Not only did he fully cause the most harmful play of the game, he had another two huge drops, depending on how you want to chart them. Toney’s final stat line: Five targets, one carry, zero yards and three drops.

The most impactful play in terms of win probability was the unsuccessful 4th down try for the Lions with a couple minutes remaining in the fourth quarter. I talk more about the non-sensical process employed by both teams on 4th downs later, but know that it was very much the right call for the Lions on this particular play, even if the results were poor.

Mahomes put up respectable per-play efficiency, considering Toney drop-to-pick-6. Taking out that one play would have boosted Mahomes’ EPA per play from -0.05 to 0.14, roughly top-10 type seasonal numbers. Mahomes was able to maintain efficiency without receiver support with strong scrambling numbers, and some help from the zebras. The Chiefs had a nearly four point advantage from penalties generally, with 41 yards and +2.5 EPA gained on a non-egregious defensive pass interference targeting Justin Watson down the field.

One number that jumped out at me from the game is Mahomes’ 9.1 aDOT, a number he only exceeded in one game last season (Week 11 vs LAC, 10.0). The absence of Kelce factored in, plus the need for the Chiefs to simply make something happen in chunks, without the peace of mind to know you could consistently execute smaller plays. Mahomes -6.2% CPOE looks bad, but I think it reflect smart calibration to force some balls into small-to-nonexistent windows and give recceiver who couldn’t get open chances to make plays (or draw penalties). Mahomes two most reliable wide receivers in the game were Watson and Marquez Valdes-Scantling, naturally stretching down the field where the Chiefs were going to target the ball.

Jared Goff was workman-like in his efficiency and performance, exactly what you hope for. A key for Goff was making minimal mistakes (just one sack for three yards), and letting the receivers do a lot of work after the catch. Goff can’t consistently make plays happen on clutch downs, but a 3rd & 12 conversion to Josh Reynolds under pressure stood out, netting the Lions +2.8 EPA.

I have no clue what the processes where for both teams on 4th down decisions. It looked completely vibes based, which was expected from the Chiefs (normally Reid’s negative vibes), but somewhat disheartening from the Lions, a team that was properly aggressive for most of last season. The Lions faked a punt on fourth down inside their own 20 early (a good decision by the numbers), but then passed up a chance to go on 4th and 3 in a no-brainer field position on the Chiefs 40 (3.9% win probability lost).

Andy Reid did his normal thing passing up chances to go for it twice with good win probability to gain, then was dead-set on punting on 4th & 20 (then 4th & 25 after a penalty) with the two-minute warning and three timeouts left. The numbers say that it was the wrong move to go-for-it then, but there might be a small sample issue with the uniqueness of the circumstances. I believe in Mahomes as much as anyone, and I’d probably still punt in that situation, or at least consider something creative, like taking the ball inside your own 10 and downing it to force the Lions to give you the ball back.

However you slice it, the preparation and process for these decisions was severely lacking in the game, which opens the door for more poor calls in the future, driven by whichever way the wind happens to be blowing.