Statistical analysis of the greatest defenses in NFL history

Re-running this from March 2009. - nc

Some of you who have been on the site for a while remember my statistical analyses on QBs, RBs (Part 1, Part 2), and WRs from last year. Well, recently, we had a Ravens fan arguing that the 2000 Ravens unit was the greatest defense in the history of the National Footall League. Sure, that's to be expected of partisans of Baltimore much like there was plenty of talk amongst Steelers faithful this year that the 2008 Steelers defense was the greatest to ever take the field. And if you were a fan of the Chicago Bears, particularly one who was old enough to vividly remember the dominating fashion in which the 1985 defense suffocated opponents, then your vote probably would go there. And on and on.

That got me interested in looking into a comparative analysis of a number of the greatest defenses in NFL history. If there was ever a time to do it, now is the time considering the great performance this year's Steelers defense registered. There are a handful of defenses that are generally argued as the greatest defensive statistical seasons, so I thought I would focus on those. I also wanted to include a couple of other great defensive seasons to add depth and context to the comparative analysis.

The best ever? Let's try to find out after the jump.

The seasons that I will analyze in this post are:

• 1973 Los Angeles Rams
• 1976 Pittsburgh Steelers
• 1977 Atlanta Falcons
• 1985 Chicago Bears
• 1991 Philadelphia Eagles
• 2000 Baltimore Ravens
• 2008 Pittsburgh Steelers

I think it's fairly obvious to most NFL fans that the '76 Steelers, '85 Bears, and '00 Ravens are the defenses most commonly discussed by pundits as the contenders for the greatest ever. However, I also included the '73 Rams and '91 Eagles because they were the teams that were mentioned throughout the course of the 2008 season as the Steelers continued to hold opposing offenses under 300 yards of total offense for the first 14 contests of the regular season. The '73 Rams were the only team to hold every last opponent under 300 yards each week (14 games), while the '91 Eagles were the last to hold 15 of 16 opponents below 300 yards total offense prior to Dick LeBeau's defense doing so in 2008.

I also included the '77 Falcons because they only gave up 129 points that season. I had never heard of this season until the NFL Network had a show that included the '77 Falcons "Grits Blitz" defense as one of the top defenses in NFL history. But, I believe they had the fewest total points allowed in a 14 game season, which obviously merits their inclusion in this analysis.

To begin, it makes the most sense to look at the raw numbers of each of these defenses. The following chart is not an exhaustive list of statistics, but rather, it includes what I thought were the most pertinent statistics. Sacks are not included because the league did not officially track them for all of these years, and more importantly, they were not included in the data that I got from pro-football-reference.com.

RAW DATA OF BEST SINGLE SEASON DEFENSES

	YDS	PTS		TOTAL	TOTAL	PASS	PASS	PASS	RUSH	RUSH	TOTAL
TEAM	Rank	Rank	G	Points	Yards	Yds	TD	Int	Yds	TD	FR
1973 Los Angeles Rams	1	4	14	178	2951	1681	10	20	1270	5	18
1976 Pittsburgh Steelers	1	1	14	138	3323	1866	9	22	1457	5	24
1977 Atlanta Falcons	2	1	14	129	3242	1384	9	26	1858	5	22
1985 Chicago Bears	1	1	16	198	4135	2816	16	34	1319	6	20
1991 Philadelphia Eagles	1	5	16	244	3549	2413	16	26	1136	4	22
2000 Baltimore Ravens	2	1	16	165	3967	2997	11	23	970	5	26
2008 Pittsburgh Steelers	1	1	16	223	3795	2511	12	20	1284	7	9

It's nice to have all this data in one place, but it's far too early to make any sort of conclusions based on it. Basically all this tells us is that the defenses we're looking at are good. Very good. Those pass and rush TD numbers are ridiculous in most instances. Still, beyond the simple truism that these Ds were stingy, it's hard to make any sort of insightful and convincing comparisons, primarily because three of the teams played 14 game schedules, while the other four played 16 game schedules.

And that's before mentioning changes in the way the game has been played on the offensive side of the ball, and the rule changes facilitating a trajectory towards more entertaining, offensive-oriented contests. For that reason, I wanted to look at the data on a per game and per play basis. You can see that data below, as well as the Offensive Ranks for each of these teams. I included the Offensive Ranks because the efficiency of a team's offense (particularly rush offense) is widely believed to impact the defense. A good offense keeps the defense off the field more, which keeps the defense fresher and limits the opposing offenses' opportunities with the ball.

PER GAME AND PER PLAY DATA

WITH TEAMS' OFFENSIVE RANKS ADDED

	TOTAL	TOTAL	TOTAL	PASS	PASS	RUSH	RUSH	O_Rank	O_Rank	O_Rank	O_Rank
TEAM	Pts/G	Yds/G	Yds/Play	INT/G	TD/G	Yds/Rush	TD/G	Points	Total	Pass	Rush
1973 Los Angeles Rams	12.7	211	4	1.43	0.71	3.5	0.36	1	1	13	2
1976 Pittsburgh Steelers	9.8	237.4	3.8	1.57	0.64	3.2	0.36	5	9	22	1
1977 Atlanta Falcons	9.2	231.5	3.7	1.86	0.64	3.7	0.36	25	25	27	17
1985 Chicago Bears	12.4	258	4.4	2.13	1	3.7	0.38	2	7	20	1
1991 Philadelphia Eagles	15.25	222	3.9	1.63	1	3.0	0.25	18	26	22	21
2000 Baltimore Ravens	10.3	248	4.3	1.44	0.69	2.7	0.31	14	16	22	5
2008 Pittsburgh Steelers	13.9	237	3.9	1.25	0.75	3.3	0.44	20	22	17	23

As in my previous statistical comparisons, I have sorted the teams according to each of the per game and per play statistics as well as for their team's offensive rank. For interceptions and fumble recoveries per game, I sorted so that defenses with higher numbers would rank higher. For Pts/G, Yds/G, Yds/P, TD/G, and Y/A, I sorted so that those defenses with higher numbers would rank lower. The idea being that defenses that give up more yards are worse, but those with more fumble recoveries and interceptions are better. I also sorted the offensive rankings for each team so that a defense with a higher offensive ranking would be given a lower ranking in this table. Again, the idea here is that a defense with a better offense does not have as hard of a job, but a defense with a worse offense would have a harder job. (Please Note: This table is not saying that the 2008 Steelers had to top rated rush offense. It is saying that the defense is ranked 1st in terms of having the worst ranked rush offense of any of the defenses in this analysis) Again, as in my other posts, I took an average of the teams' rankings to see each defense's average ranking. The table is shown below.

DEFENSES' RANKINGS: PER GAME, PER PLAY AND OFFENSIVE RANKINGS

TOTAL

TOTAL

TOTAL

PASS

PASS

RUSH

RUSH

Pass+Rush

Offense

Offense

Offense

Offense

Team

Pts/G

Yds/G

Y/P

TD/G

INT/G

TD/G

Y/A

FR/G

Pts

Total

Pass

Rush

Average Ranks

'73 Rams

5th

1st

5th

4th

6th

5th

5th

5th

7th

7th

7th

5th

5.17

'76 Steelers

2nd

5th

2nd

1st

4th

4th

3rd

1st

5th

5th

3rd

6th

3.42

'77 Falcons

1st

3rd

1st

1st

2nd

3rd

6th

3rd

1st

2nd

1st

3rd

2.25

'85 Bears

4th

7th

7th

7th

1st

6th

2nd

6th

6th

6th

5th

6th

5.25

'91 Eagles

7th

2nd

3rd

6th

3rd

1st

7th

4th

3rd

1st

4th

2nd

3.58

'00 Ravens

3rd

6th

6th

3rd

5th

2nd

1st

2nd

4th

4th

2nd

4th

3.50

'08 Steelers

6th

4th

4th

5th

7th

7th

4th

7th

2nd

3rd

6th

1st

4.67

Based on just this data, we would conclude that the rankings of these defenses should be as follows:

1. 1977 Falcons
2. 1976 Steelers
3. 2000 Ravens
4. 1991 Eagles
5. 2008 Steelers
6. 1973 Rams
7. 1985 Bears

However, now we're getting into what I think is the most important data in this analysis. I have compiled the offensive rankings of each of these teams' opponents, and then I created a table showing a distribution of the strength of offenses that each of these teams have faced. As shown below, the table gives a count of the number of top 3, top 5, top 10, top half, bottom half, bottom 10, bottom 5 and bottom 3 offenses that each defense faced in its season. I decided to break it down by top half and bottom half (as opposed to, say, top and bottom 15) because the number of teams in the league has increased at various times. The 2000 Ravens were the only team that had an odd number, so I gave them the benefit of having anything in the top 16 out of 31 being a top half offense.

DISTRIBUTION OF OPPONENTS' OFFENSIVE RANKINGS

	1973	Rams	1976	Steelers	1977	Falcons	1985	Bears	1991	Eagles	2000	Ravens	2008	Steelers
	Pts	Yards	Pts	Yards	Pts	Yards	Pts	Yards	Pts	Yards	Pts	Yards	Pts	Yards
Top 3 Offenses	1	0	1	1	0	1	0	1	3	2	0	0	1	1
Top 5 Offenses	1	1	2	1	2	3	2	2	4	4	0	0	2	2
Top 10 Offenses	4	4	5	6	3	4	4	5	7	6	2	2	4	4
Top Half Offenses	4	4	8	8	4	6	7	8	7	7	5	6	8	7
Bottom Half Offenses	10	10	6	6	10	8	9	8	9	9	11	10	8	9
Bottom 10 Offenses	8	7	5	4	7	8	6	6	7	6	8	8	6	4
Bottom 5 Offenses	5	6	2	3	3	3	1	2	3	4	5	5	5	4
Bottom 3 Offenses	3	3	1	1	2	2	1	2	3	3	5	4	4	4

Just a cursory review of this data seems to favor the '76 Steelers, '85 Bears, '91 Eagles, and '08 Steelers because the other three teams faced 10 or more opponents in the bottom half in either points or yards, or both. But, it's still hard to draw really solid conclusions from the data. So, I decided to borrow an idea (or at least something similar to it) that I've seen at pro-football-reference.com's blog. I decided to develop what I'm calling the Opponent Offensive Factor. It is nothing more than the equation shown below.

T3*10 + T5*5 + T10*2 + TH - BH - B10*2 - B5*5 - B3*10

where

• T3 = Number of Top 3 Offenses Faced
  
• T5 = Number of Top 5 Offenses Faced
• T10 = Number of Top 10 Offenses Faced
• TH = Number of Top Half Offenses Faced
• BH = Number of Bottom Half Offenses Faced
• B10 = Number of Bottom 10 Offenses Faced
• B5 = Number of Bottom 5 Offenses Faced
• B3 = Number of Bottom 3 Offenses

The resulting numbers don't really have a quantifiable defined meaning, but what they do give us is an idea of how tough each defense's overall offensive schedule was relative to the other defenses. Because of the multipliers, the number of top 3 offenses has a weight of ten times that as the number of top half offenses. This means that top 3, 5 and 10 offenses are counted more than once, but that's fine, because it is the same for every defense that is being analyzed here.

So, using this simple equation, I came up with the following Opponent Offensive Factors for each of these defenses (both for offensive points rankings and offensive yards rankings).

OFFENSIVE OPPONENT FACTOR

Points

Yards

Points

Yards

Points

Yards

Points

Yards

Points

Yards

Points

Yards

Points

Yards

'73

Rams

'76

PIT

'77

ATL

'85

CHI

'91

PHI

'00

BAL

'08

PIT

Opp Offensive Factor

-54

-67

2

-4

-39

-20

-11

-12

3

-12

-93

-81

-49

-42

Based on these numbers, we can rank the teams in terms of which had the toughest offensive schedules, as determined by the final offensive ranks that each team played. The rankings are:

1. 1976 Steelers
2. 1991 Eagles
3. 1985 Bears
4. 1977 Falcons
5. 2008 Steelers
6. 1973 Rams
7. 2000 Ravens

Also, if we were to group those defenses who played similar levels of offensive schedules, it would make sense to me to break them up into 4 different groups:

1. '76 Steelers and '91 Eagles - These are the only two defenses with a positive Offensive Opponent Factor, indicating that they are the only teams with at least as many good to really good offenses on their schedule as bad to really bad offenses.
2. '85 Bears - I almost included them in the first group, but they didn't have a positive Points Offensive Opponent Factor in either yards or points, unlike those in group 1.
3. '77 Falcons, '08 Steelers, and '73 Rams - Each of these teams had Offensive Opponent Factors significantly below zero, which indicates that either most of their opponents were in the bottom half offensively (Rams) or they had more teams in the bottom 3 and bottom 5 than the in the top 3 or top 5 (Falcons and Steelers)
4. '00 Ravens - This indicates that for the Ravens, most of their opponents in the bottom half and they had more teams in the bottom 3, bottom 5, and bottom 10 than in the top 3, top 5 or top 10.

At this point, looking at the ranking lists that we have so far, the front runner for greatest defense of all time appears to be the 1976 Steelers. Their composite ranking in per game, per touch, and offensive statistics is second to the 1977 Falcons, but they also had more games against better offenses than the Falcons. Also, as we're getting ready to see, they played in a tougher average offensive season than the Falcons.

Before I end the analysis, I also wanted to look at one more thing: the average offensive stats for each of the years that we are looking at. The table with those statistics is shown below.

AVG OFFENSIVE STATS FOR RELEVANT YEARS

Year	Games	Points	Pts/Game	Yards	Yds/Game	Plays/Game	Yds/Play	1stDown/G
1973	14	272.3	19.45	3994.3	285.31	62.2	4.6	16.57
1976	14	268.1	19.15	4237.9	302.71	65.67	4.6	17.66
1977	14	240.5	17.18	4000.7	285.76	64.77	4.4	16.88
1985	16	344.5	21.53	5271.1	329.44	65.55	5	19.18
1991	16	303.8	18.99	4908.6	306.79	60.76	5	17.75
2000	16	330.8	20.68	5110.8	319.43	62.96	5.1	18.42
2008	16	352.5	22.3	5235.8	327.24	61.87	5.3	18.42

To make it a little easier to analyze, I ranked each of the years in each category. The resulting table is shown below.

RANKING OF AVG OFFENSIVE STATS

Team	G	Pts	P/G	Yds	Y/G	Plays	Y/P	1stD	AVERAGE RANK
1973 Rams	14	5th	4th	7th	7th	7th	6th	7th	6.14
1976 Steelers	14	6th	5th	5th	5th	5th	5th	5th	5.14
1977 Falcons	14	7th	7th	6th	6th	6th	7th	6th	6.43
1985 Bears	16	2nd	2nd	1st	1st	1st	3rd	1st	1.57
1991 Eagles	16	4th	6th	4th	4th	4th	4th	4th	4.29
2000 Ravens	16	3rd	3rd	3rd	3rd	2nd	2nd	3rd	2.71
2008 Steelers	16	1st	1st	2nd	2nd	3rd	1st	2nd	1.71

As we would expect there is a trend showing increases in offensive numbers as the years go by. One thing that is especially interesting, though, is that there appeared to be a peak in offensive production in the mid-'80s, as the year 1985 ranks in the top 3 in all offensive categories, with the highest average offensive rank of any of these years. There was a decline in offensive production in the '90s, followed by a resurgence of offense in the '00s.

At first that surprised me, but then I thought back to the '80s, and I remembered the Bill Walsh offenses, the Dan Marino Dolphins, and the Run and Shoot offenses that the Oilers and Lions used. That would explain the offensive peak in the '80s. The year 1985 was also the one with the most plays, which would also fit in with the style of Marino and the Run and Shoot teams. In addition, the NFL made some rules changes in the '90s to shorten the game, which resulted in fewer plays per game. They also made rules changes to favor offense, and have continued to do so, in the '90s and '00s. The resulting increases in points, yards, yards/play, and first downs after the '90s are to be expected. The resulting decrease in number of plays is also to be expected with the rules changes intended to keep most games under 3 hours.

Now, back to the defensive analysis. Based on these numbers, offensive production clearly increased in the '80s, '90s, and '00s. In particular, the years 1985 and 2008 are very similar in overall offensive production. In the '80s, it was probably more due to innovations brought on by the rules changes of 1978 (the 5 yard contact rule). In the '00s, it was probably more due to the continued rules changes to favor WRs over DBs. Not only has the NFL continued to change rules to favor the offense since 2000, they have also continued to increase the emphasis of calling pass interference to officials with each offseason. Anyway, regardless of the reasons, it's clear that 1985 and 2008 are the two years from our analysis in which offensive productivity were the greatest.

If you made it this far, I'm sure you're wondering what conclusion can we really draw from all of this data? Pure defensive numbers combined with the strength of offensive opponents gives the edge to the 1976 Pittsburgh Steelers. But, when looking at the average offensive statistics from each year, it's clear that offensive production as a whole has been higher since the '80s. Of the '70s, 1976 was the highest in terms of average offensive output, which means we can safely conclude that the 1976 Steelers were the best defense of the '70s. But were they the best of all time?

Having eliminated the '73 Rams and '77 Falcons from the discussions, let's see if we can determine which of the post  '70s teams can be eliminated as well. Of the remaining teams, the 2000 Ravens had the best raw numbers. But, IMO, their argument for best defense of all time is severely weakened by the extreme weakness of their offensive opponents. Their one clear argument for best defense is points given up in a season (33 fewer than the next closest 1985 Bears). But, looking at their Opponents Offensive Factor, they played the worst offensive schedule of the 4 remaining teams, by far. So, I would eliminate the 2000 Ravens from consideration.

Next in the raw numbers statistics comes the 1991 Eagles. They have many arguments to support them as the best of the remaining 3 defenses. They were the only defense (post '70s) to have a positive Opponent Offensive Factor. In fact, not only did they play a very balanced schedule in terms of top half and bottom half offenses, but they played 3 games against top 3 scoring offenses and 4 against top 5 offenses (both scoring and yards). Both of these are more than any other defense.

Add to this the fact that the 1991 Eagles offense ranked 26th in yards and 21st in rushing yards, and the 1991 Eagles accomplishments become even more impressive. The only possible knock on them in this analysis is that the year 1991 had the 4th overall average offensive statistics.

That brings us to the 1985 Bears and the 2008 Steelers, 7th and 5th place, respectively in terms of raw statistics. But, they also played in the top 2 seasons in terms of average offensive output (1985 1st and 2008 2nd). The Bears had a better Opponent Offensive Factor, which means that they faced a tougher offensive schedule than the Steelers. But, they also had the number 1 ranked rushing offense for their season, while the Steelers rushing offense ranked 23. As discussed earlier, this means that the Steelers offense was less likely to provide the defense with help in terms of limiting the opposing offenses' opportunities. So, it seems that for every reason to argue for one, there is a counter-argument to argue for the other.

Having looked at everything, and kicked it around for a while, I would say that the two greatest defenses of all time are the 1976 Steelers and the 1991 Eagles. Neither team won a championship that year, which of course is going to be a knock on them, but both played considerably tougher offensive schedules than the others, based on the Offensive Opponent Factors. And being that it was the days before the 1993 Collective Bargaining Agreement that brought Free Agency to the game, the Steelers defense did win championships with many of those 1976 defensive players.

Regardless, I would actually give the '91 Eagles the nod, because they had a much less competent offense to complement them, particularly rushing the football, the safest and most tested way to keep your defense fresh, off the field, and limited in their exposure to the extremely capable offenses of professional football. Even though they gave up the most points per game in this analysis, their rush offense was terrible. Given the number of top 5 offenses that they faced, and their lack of a good rushing attack to keep those offenses off the field, the fact that they gave up the most points in this analysis doesn't bother me much.

In fact, after checking, I discovered that their offense had 43 giveaways that season. So, not only did the offense not help in terms of limiting offensive opponents' opportunities, it actually gave the opponent's more offensive opportunities. I was not able to find how many of them lead directly to points, but when you turn the ball over that often, it's amazing that they only allowed 244 points.

After looking at all of this, although I would favor the Eagles in this analysis, there are so many valid arguments for the 1976 and 2008 Steelers, as well as the 1985 Bears, that I would not argue strongly against any of them. I do believe, though, that the data does pretty well rule out the 1973 Rams, 1977 Falcons, and 2000 Ravens. All three played over half their schedules against bottom half offenses. Granted, the Offensive Opponent Factor for the 2008 Steelers was among the bottom three. But, when reviewing the data more closely, there are several points that argue in their favor.

1. First, they played in one the top two seasons in terms of average offensive performance, second only to 1985.

2. Second, while their Opponent Offensive Factor was low because they had an above average number of bottom 5 and bottom 3 offenses, they also had an above average number of top half offenses and average numbers of top 10, 5, and 3 offenses.

3. Third, like the 1991 Eagles, they had a bottom 3rd rush offense. And, as we Steelers fans were well aware this season, the offense often went through long periods of stagnant production.

4. Finally, I think we can safely say that 2008 was a season in which officiating and rules favored the offense over defense more than any other year in history. So, the 2008 Steelers defense certainly had several obstacles to overcome.

With all of that said, I'm certainly interested in hearing any feedback you have.