Last week's defense/special teams rankings column was not my best moment. I really wanted to include a chart plotting defenses' average scoring against their consistency rating (as created and explained by me here several weeks ago). Unfortunately, as it turns out, I'm bad at that stuff. The graph was there, but it was (a) poorly labeled, and (b) illegible.
Other than that, though, I crushed it!
Anyway, I had a request in the comments to make a more visible version of the graph. I tried. I failed. So instead, I set out to make it into a table this week, thinking that would be equally helpful, with the added advantage of being something I actually have the technical acumen to accomplish.
As I worked on it, I came up with what (I think) is an interesting system for ranking defenses. So I'm going to try it this week, with the up-front caveat that I'm somewhat figuring out this particular metric as I go, so there is the possibility that I will be completely, totally wrong. But you get the bonus of learning with me, and that's valuable.
What I'll do, then, is explain my methodology, and go to the results. At the end of the column, I'm going to list this week's rankings as my system dictates, but I'm also going to list the rankings I came up with beforehand. We'll see how they compare; we'll see how they turn out this weekend. It might be an exercise in futility, but an exercise in futility is still an exercise, you know?
Here we go.
The easiest way to rank defenses, if you aren't interested in significant research, is average scoring. It's tidy, and it's reasonable reliable. The problem, though, is consistency. Would you rather have a defense that averages 12 points a game, but scores in the negative once every three weeks, or a defense that averages 11, but is always between nine and 13? I'd gladly sacrifice the one average point for the greater reliability.
That is why I came up with my defense consistency ranking, a system that, once I get it ironed out here, I plan to apply to all positions eventually. Without considering actual value of performance, the system simply calculates how consistent a specific defense is.
Now, I'm not crazy - I'd rather a high scorer that is moderately inconsistent than an awful team that scores exactly one point every week. I know the average score needs to be more weighted in this system (a problem I was running into with my simple graph last week). So how to properly scale the two measurements?
The forgotten concern
One thing I've stressed since the preseason is that there are always two major factors to consider in picking a defense - the defense playing (obviously), and the offense it is playing against. The Kansas City Chiefs have been the best defense in the league this season, but few people would rank them first this week going against the Denver Broncos. Honestly, a lot of the time, I think the opposition is at least as important as the specific defense, and it doesn't always get enough attention.
So in addition to finding average scores and consistency ratings for all the defenses, I found average scores and consistency ratings for all the opposing teams (in other words, the data versus each team). I'm now to four total data points for this.
Below, I've included the charts of the average score and m rating for each defense and each opponent, with the rankings in parentheses:
Now, the easiest way to do this would be to add a unit's ranking in average score to its average ranking in consistency, and the lowest number wins. That, though, forgets the problem I mentioned several paragraphs ago - consistency is important, sure, but actual score is moreso.
To correct for this, I double-weighted average score, then took the average ranking. So, if a team was first in average scoring, but 13th in consistency, it would be 1+1+13=15. Divide that by three to get five, and that is that unit's overall score. Ranking those scores - lowest best - should yield an approximately accurate ranking for which defenses would fare best against a neutral opposition.
Of course, no one is facing a neutral opposition (except perhaps the Bills, who are facing the uber-neutral Jets this week). Which brings me back to the opposition scores - perform the same tactics, and you get the ranking of opponents you would want your neutral defense to face.
So what I did then was combine scores of Week 11 opponents. Seattle is playing Minnesota - I took the Seahawks' first-ranked defense and combined it with the Vikings' seventh-ranked opposition score, for a composite of 8. That brought that matchup in at first this week. In my initial just-looking-at-them ranks, I had the Seahawks in first, as well, matching those numbers up rather well, actually. Of course, those rankings go both ways; the Vikings' defense rank 28th right now, while the Seahawks are the 18th-ranked opponent, for a composite of 46, which is 27th this week; I had them at 28th.
(The numbers and my opinion didn't match up nearly this well in all the matchups, but I won't lie; that was nice to see.)
Okay, I hope that was clear. If it wasn't, or if you want more information, or if you can find some fatal flaws in my methodology (which would not remotely surprise me), feel free to get down in the comments. I'll respond to everything relevant.
With that, here are the rankings. In the case of tiebreakers in the numbers, I leaned toward the higher-scoring defense before the better consistency rating. As a reminder, the first column of rankings is the list I made before I started my calculatin', while the second column is the rankings that resulted from all this figgerin'. I'm definitely going to revisit this whole thing next week, and we'll see if I've cracked some sort of code:
|Rank||My rankings||Opponent||System rankings||Opponent|
|1||Seattle Seahawks||Minnesota||Seattle Seahawks||Minnesota|
|2||Arizona Cardinals||at Jacksonville||Arizona Cardinals||at Jacksonville|
|3||Carolina Panthers||New England||Cincinnati Bengals||at Cleveland|
|4||Cleveland Browns||Cincinnati||Buffalo Bills||NY Jets|
|5||Cincinnati Bengals||at Cleveland||Chicago Bears||Baltimore|
|6||New York Jets||at Buffalo||Green Bay Packers||at NY Giants|
|7||Oakland Raiders||at Houston||New York Jets||at Buffalo|
|8||San Francisco 49ers||at New Orleans||New England Patriots||at Carolina|
|9||Tennessee Titans||Indianapolis||Detroit Lions||at Pittsburgh|
|10||New England Patriots||at Carolina||Carolina Panthers||New England|
|11||Baltimore Ravens||at Chicago||Tennessee Titans||Indianapolis|
|12||New Orleans Saints||San Francisco||Houston Texans||Oakland|
|13||Houston Texans||Oakland||San Diego Chargers||at Miami|
|14||New York Giants||Green Bay||New Orleans Saints||San Francisco|
|15||Tampa Bay Buccaneers||Atlanta||Cleveland Browns||Cincinnati|
|16||Kansas City Chiefs||at Denver||Oakland Raiders||at Houston|
|17||Green Bay Packers||at NY Giants||San Francisco 49ers||at New Orleans|
|18||Denver Broncos||Kansas City||Denver Broncos||Kansas City|
|19||Indianapolis Colts||at Tennessee||Jacksonville Jaguars||Arizona|
|20||Detroit Lions||at Pittsburgh||Tampa Bay Buccaneers||Atlanta|
|21||Chicago Bears||Baltimore||Indianapolis Colts||at Tennessee|
|22||Buffalo Bills||NY Jets||Atlanta Falcons||at Tampa Bay|
|23||Miami Dolphins||San Diego||Baltimore Ravens||at Chicago|
|24||San Diego Chargers||at Miami||Kansas City Chiefs||at Denver|
|25||Atlanta Falcons||at Tampa Bay||Philadelphia Eagles||Washington|
|26||Pittsburgh Steelers||Detroit||Miami Dolphins||San Diego|
|27||Philadelphia Eagles||Washington||Minnesota Vikings||at Seattle|
|28||Minnesota Vikings||at Seattle||Pittsburgh Steelers||Detroit|
|29||Washington Redskins||at Philadelphia||Washington Redskins||at Philadelphia|
|30||Jacksonville Jaguars||Arizona||New York Giants||Green Bay|