Let's begin with the conclusion of last week’s Watchtower:
The numbers tell me that the most likely outcome will be a 20-24 loss.However, I just can’t believe it. All of the factors besides the raw data point towards the Lions both scoring more, and allowing fewer, points against the Rams. My instincts tell me that the team that fell two points shy of beating the Packers in Lambeau has the edge over the Rams at home—and Vegas agrees with me; the Lions are three point favorites. Either way, it will be a closer game than we thought it’d be going into the season—but I’m going to depart from the numbers. I’m predicting a 27-17 win for the Lions.
As commenter Angus Osborne said:
Ty: this Watchtower proves that despite the diligent detailed work you put in you're more a fan than a statistician.
That’s true, I am a fan more than a statistician. In fact, I’m no kind of statistician; I’m terrible at this stuff—but I’m so fascinated with the ways statistics can reveal hidden truths in football, that I’m researching and learning and gleaning everything I can to try and play with the statistical big kids. However, I also believe that all the numbers have to be grounded in reality, and my eyes and instincts revealed the numbers’ lies.
In this case, the numbers were telling me that Shaun Hill was going to play like an average of his performances against Chicago, Philly, Minnesota, and Green Bay—and I knew the Rams weren’t going to put up that kind of a fight. Second, the numbers were telling me that the Rams were going to play like an average of their performances against Arizona, Oakland, Washington, and Seattle—and I knew the Lions were going to put up more of a fight.
The “mean” (average) of a given data set isn’t always the most telling number; sometimes it’s the “median” (the one in the middle), or the “mode”) the most frequent number. If there were a defense that gave up seven points in each of its first five games, then 42, then seven again, the average would tell you that team typically allows 13 points—when really, it typically allows seven. As I said in the Watchtowering of the Rams:
The Rams have faced the Cardinals, Raiders, Redskins, and Seahawks, and allowed 17, 16, 16, and 3 points in those games. I’m inclined to believe that allowing about 16 points to a mediocre offense is where the Rams’ defense “really” is, and the Lions have a slightly-better-than-mediocre offense. They also have Calvin.
As it turns out, I was right that the Rams were an "alllow 16 points to a mediocre offense” kind of defense, but I was wrong about the Lions. They aren’t a “slightly-better-than-mediocre” offense, they’re the sixth-best in the NFL; now scoring 25.2 PpG. That Lions offense against a middling Rams defense (with a likely systemic advantage on the ground) projects out to 27-30 points; 30 offensive points were exactly what the Lions scored. Of course my work last week could only take into account Weeks 1-4:
Given theoretically-lesser-but-probably-really-equal talent, and a nonexistent-but-probably-really-existent systemic advantage, especially vis-a-vis the run game, I project the Lions will score 15-to-20 points, pass for 6.25-to-6.50 YpA, and rush for 4.5-to-4.75 YpC. I have low-to-medium confidence in this projection.
. . . and yeah. Not what happened. What about the defense?
Given no real systemic advantage or disadvantage, I project the Rams’ offense to meet expectations against the Lions’ defense: 23-26 points, 6.5-7.5 YpA, and 4.00-4.25 YpC. I have low-to-medium confidence in this projection.
I got nothin’ here. The Lions only sacked the Rams once in 74 dropbacks (though they did score countless pressures and hits on the quarterback). They forced three turnovers, but one was immediately returned for a score, so snaps and drives weren’t terribly depressed. The Lions simply shut the Rams down, regardless of averages. I’m trying to nail down exactly what is was that enabled this kind of performance—but for the meantime, just be happy about it.