NFL DEs and 3-4 OLBs ranked by production, with Stats

[PonyExpress]

NFL Defensive Ends and 3-4 OLBs
Statistical Analysis of 74 NFL Defensive Ends, 2004 and 2005 seasons.

Here are the production rankings for 4-3 Defensive Ends and 3-4 OLBs. I combined the 2004 and 2005 season statistics to rank these players over a two year window. The two year window mitigates against once great players past their prime and also against one year wonders. I ranked them against the pass, against the run and combined.

I. Defensive End Rankings: Impact in the Passing Game, 2004 and 2005 Seasons
Abbreviation Key:
PR: Position Rank
S: Sacks
FFP: Forced fumbles in the passing game
PBU: Passes Broken Up

DE Statistical Notes:
For an NFL 4-3 DE or 3-4 OLB,
(a) a Forced Fumble is worth roughly 1/2 a turnover.
(b) a PBU is worth roughly ¼ of an interception.
(c) a PBU is worth roughly ½ of a forced fumble.
(d) a PBU has roughly the same value as a sack.
(e) The First Tie breaker in Position ranking for DEs is S + ((FFP + PBU/2) x 2). The second tie breaker is (FFP + PBU/2).


Pos. Rank, Player S(PR) FFP + PBU/2 (PR) PR aggregate
1.Peppers, Car 22(4) 11(2) 6
2.Freeney, Ind 27(1) 8(6) 7
3.Taylor, Cin 22(4) 10.5(3) 7
4.Rice, TB 26(2) 8(6) 8
5.Porter, Pitt(3-4) 18(8) 11.5(1) 9
6.Mathis, Ind 23(3) 7.5(7) 10
7.Schobel, Buff 20(6) 8(6) 12
8.Umenyiora, NYG 22(4) 6.5(9) 13
9.Foley,SD(3-4) 15(11) 10.5(3) 14
10.Kearse, Phi 16(10) 9.5(4) 14
11.Ellis, Dall 17(9) 9(5) 14
12.Kerney, Atl 20(6) 7(8) 14
13.Abraham, NYJ 21(5) 6.5(9) 14
14.W. Smith, NO 16(10) 8(6) 16
15.Hayward, JK 20(6) 6(10) 16
16.Allen, KC 20(6) 6(10) 16
17.A. Brown, Chi 12(14) 10.5(3) 17
18.Berry, Ar 21(5) 5(12) 17
19.Fisher, Sea 18(8) 6(10) 18
20.McGinest,NE(3-4)16(10) 6.5(9) 19
21.Hall, Det 17(9) 6(10) 19
22.Suggs, Balt 19(7) 5(12) 19
23.KGB, GB 22(4) 3(16) 20
24.Grant, NO 14(12) 6.5(9) 21
25.Vrabel,NE(3-4)***15(11) 5.5(11) 22
Burgess, Oak 19(7) 3.5(15) 22

***Because Vrabel changed to ILB in 2005, I used his 2003 and 2004 seasons as an OLB here.

II. Defensive End Rankings: Impact in the Running Game, 2004-2005
Abbreviations:
PR: Position Rank
STF: Stuffs For Loss against the run
TT-S: Tackles minus Sacks

Statistical Notes.
For an NFL DE:
(a) a tackle for loss in the run game has roughly the same value as 10 conventional tackles.
(b) a decade of tackles is considered the same value (72 or 79 is statistically negligible)
(c) the tie-breaker for pos. rank for DEs is (STFs x 10) + (TT– S). The second tie breaker is TFL-S.

Pos. Rank, Player STF(PR) TT-Sacks (PR) PR aggregate
1.Grant, NO 17(1) 128(2) 3
2.Schobel, Buf 17(1) 124(2) 3
3.Suggs, Balt 15(2) 109(4) 6
4.Kampman, GB 12(5) 138(1) 6
5.Taylor, Mia 13(4) 119(3) 7
6.Umenyiora, NYG 14(3) 107(4) 7
7.Okeafor, Ar 15(2) 87(6) 8
8.Spires, TB 12(5) 96(5) 10
9.Brown, Chi 13(4) 85(6) 10
10.Porter, Pitt 12(5) 93(5) 10
11.Fujita, Dall 7(10) 134(1) 11
12.Smith, Cin 8(9) 123(2) 11
13.Carter, Mia 11(6) 90(5) 11
14.Brock, Ind 12(5) 78(7) 12
15.Peppers, Car 10(7) 93(5) 12
16.Strahan, NYG 9(8) 101(4) 12
17.Kerney, Atl 9(8) 99(5) 13
18.Little, StL 10(7) 85(6) 13
19.Hall, Det 9(8) 91(5) 13
20.Smith, NO 9(8) 86(6) 14
21.Kelsay, Buff 10(7) 74(7) 14
22.Young, SF 10(7) 74(7) 14
23.McGinest, NE 8(9) 92(5) 14
24.Ellis, NYJ 9(8) 81(6) 14
25.Wistrom, Sea 8(9) 82(6) 15



III. Overall Defensive End Rankings, 2004 and 2005. Passing Game and Running Game Combined.
Abbreviation Key
PR: Position Rank
PR@Pass: Pos Rank against the Pass
PR@Run: Pos Rank against the Run

Statistical Note:
(a) Tie breaker for pos rank among DEs is impact against the pass.

Pos. Rank, Player PR@Pass PR@Run PR aggregate
1.Taylor, Cin 7 7 14
2.Schobel, Buff 12 3 15
3.Peppers, Car 6 12 18
4.Porter, Pitt(3-4) 9 10 19
5.Umenyiora, NYG 13 7 20
6.Freeney, Ind 7 16 23
7.Grant, NO 21 3 24
8.Suggs, Balt 19 6 25
9.Kerney, Atl 14 13 27
10.A. Brown, Chi 17 10 27
11.Rice, TB 8 21 29
12.Abraham, NYJ 14 16 30
13.W. Smith, NO 16 14 30
14.Foley,SD(3-4) 14 17 31
15.Okeafor, Ar 23 8 31
16.Mathis, Ind 10 22 32
17.Allen, KC 16 16 32
18.Hayward, JK 16 16 32
19.Hall, Det 19 13 32
20.Ellis, Dall 14 19 33
21.Berry, Ar 17 16 33
22.McGinest, NE(3-4)19 14 33
23.Kampman, GB 27 6 33
24.Fisher, Sea 18 16 34
25.Brock, Ind 23 12 35
26.Kearse, Phi 14 22 36
27.Little, StL 23 13 36



Observations.
1. When Brady said that Jason Taylor was the best defensive player he had ever played against, he may have been telling the truth, and not just buttering up the opposition.

2. I’m going to pay more attention to Aaron Schobel this year. He may be one of the more under appreciated players in the NFL.

3. I have always thought Peppers was an overrated player, stemming from his poor Super Bowl performance against the Pats. Now I have to give more credit to the Pats O-Line than even before. What an amazing performance they put on in that game.

4. McGinest may have been undervalued for his production against the run.

5. Players on the upswing: W. Smith NO, Umenyiora NYG, Allen KC. I think Smith could become one of the dominant DEs in the game going forward. He has a lot of leadership skills for a guy stuck on NO, and is excellent against the run and the pass. Any thoughts?

 

[VT_NY_Patsfan]

Where is Seymour Ranked

Pony,

I find it hard to belive that the NFL's highest paid DE isn't on your list. I've search and I can't find him.

 

[PonyExpress]

Pony,

I find it hard to belive that the NFL's highest paid DE isn't on your list. I've search and I can't find him.

Seymour is ranked among Defensive Tackles. I did a ranking last week for 3-4 DEs and 4-3 DTs.

 

[arrellbee]

Since 3-4 DE is somewhere between a 4-3 DT and a 4-3 DE ?? if you have the data where it would be relatively straightforward to include Seymour in this ranking, could you take the time to do that ? I, for one at least, would really like to see how that comes out.

 

[Box_O_Rocks]

Observations.
1. When Brady said that Jason Taylor was the best defensive player he had ever played against, he may have been telling the truth, and not just buttering up the opposition.

2. I’m going to pay more attention to Aaron Schobel this year. He may be one of the more under appreciated players in the NFL.

3. I have always thought Peppers was an overrated player, stemming from his poor Super Bowl performance against the Pats. Now I have to give more credit to the Pats O-Line than even before. What an amazing performance they put on in that game.

4. McGinest may have been undervalued for his production against the run.

5. Players on the upswing: W. Smith NO, Umenyiora NYG, Allen KC. I think Smith could become one of the dominant DEs in the game going forward. He has a lot of leadership skills for a guy stuck on NO, and is excellent against the run and the pass. Any thoughts?

Schobel is quite good. Yes, the O-line was extraordinary in the playoffs, and generally unappreciated. Willie's run defense was noted by the handful who did the tape breakdowns or read them last season .

 

[PonyExpress]

Since 3-4 DE is somewhere between a 4-3 DT and a 4-3 DE ?? if you have the data where it would be relatively straightforward to include Seymour in this ranking, could you take the time to do that ? I, for one at least, would really like to see how that comes out.

Arrellbee,
I can understand your interest in this subject; after all, Borges has hinted that the Pats were intending to cheat Seymour by franchising him as a defensive tackle. This is another example of how slithery a character Borges is. If the Pats had done so with Seymour, they would have had every right to, because Seymour is a DT in everything other than name. Seymour functions as a defensive tackle and should be compared only to other defensive tackles. The best indicator of this is his stat line: Over the last two years Seymour has 9 sacks and 5 pass break ups, total. In comparison, Jason Taylor had 12 and 10 in ’05 alone. Needless to say, if I had included Seymour in this survey he would have been in the lower third of the league. That is why I believed Seymour should be paid along the lines of John Henderson and not Jevon Kearse.

 

[primetime]

Jason Taylor plays for the Bengals?

 

[PonyExpress]

Jason Taylor plays for the Bengals?

sorry, typo. should read Mia., not Cin.

 

[PatsFanInVa]

Pony, I appreciate as much as anybody your effort in gathering this material. but I want to make a couple of points:

1) As we chatter on endlessly about here, impact players can be statistically underrepresented, and Seymour is a prime example. The Pats' 3-4 is designed to free up linebackers, not make stars out of ends, and I'm pretty sure you know this (this has to do with the Seymour ranking in the last exercise)

2) I don't like all your assumptions going in. Particularly galling is the idea that a tackle for a loss is the equivalent of 10 conventional tackles. Essentially, it has the same value as a QB sack, the difference being that the QB typically loses more yardage than a RB. You "10 tackle equivalent" TFL bears no resemblence to the importance of these plays - the impressiveness, perhaps. But not the importance. Your guy can stuff a runner a yard shy of the goal line, and sit on his butt "outperforming" a guy with 9 solo tackles.

I haven't looked too much deeper than this - my first thought is you re-do the stat tables, assuming a TFL gets you:
A tackle plus,
Another tackle plus,
An additional tackle for each four yards behind the line of scrimmage.

The logic: the average NFL running play goes for four yards. The TFL gets you the stop for the running play, that's your tackle.

You get the second tackle because on average it takes another extra play to get those four yards. A tackle equals a play stopped. Sure, be generous. Give him another for a second play's impact prevented.

You get the quarter point per yard going backwards to represent this progression going backward in space; I doubt very many TFL actually get you more than a third tackle out the whole scenario.

To my mind, that's more than generous in the TFL category. These only become significant in first down and goal line situations, and if they cannot be correlated specifically to those situations, the above more than adequately accounts for them, to my way of thinking.

At any rate this compilation results in an increased depth of knowledge of one very important data set, to wit, what Ponyexpress thinks is the best statistical formula for ranking at a given position.

Some of it seems pretty questionable - but again, I'm reacting primarily to the unbelievable importance attached to the TFL in this representation.

PFnV

 

[maverick4]

re

I agree with PFnV.

One example I can think of is Ted Washington. The guy is huge, consistently takes on double teams, and makes it very hard for teams to get good yardage between the tackles. He is without question a force in the middle. However, his annual stats under your evaluation system would make him seem like an average player.

Also, I think in general, teams that are extremely deep in talent will have statistics spread out evenly over all of their players, even if all the players are playing at a high level. For example, the Patriots defense has only sent one or two players every year to the pro bowl, but I think several of them could shine with better numbers on other (less winning) teams. Also, on offense, no receiver will shine because Tom Brady spreads the ball around so well; lesser QB's will mostly target a few guys, causing those players to accumulate huge stats.

 

[PonyExpress]

I can see that this 10:1 ratio appears arbitrary, so let me explain my method. Against the run, I considered that an active and durable player will accumulate tackles; that a “playmaker†will accumulate Stuffs; and that those two aspects of a defensive end are equal in importance. The 10:1 ratio is a byproduct of that equivalence, and based on the frequency of those statistics among all Defensive ends. Maybe I am overrating tackles for loss; but maybe you are overrating “tackles-for-gain†(re:Ray Lewis). As far as certain team philosophies spreading out statistics and thus reducing the stats of good team players, I think this is a fallacy, and I’ll explain why. Rodney Harrison I’m sure would be at the top of the safety list for his 2003-4 seasons, and Ty Law would do very well among corners for 2002-3. Seymour’s 2003 season, when he established himself as a great DL in football, was monstrous statistically. Also, Bruschi’s 2003-4 seasons would do very well in a numbers contest. As far as the system being prejudicial against 3-4 OLBs, Joey Porter did very well for himself, and McGinest showed up well also. I expect Merriman will do nicely after the coming season.
As far as Ted Washington being “devaluedâ€, I did a DT ranking last week, where Ted showed up 4th among 3-4 nose tackles while playing for Oakland over the last two seasons. Historically, the reason Ted might be devalued is his lack of durability (he missed half a season while with the Pats), his tendency to cruise when he’s not interested, and the fact he’s a one or two down player. To ignore these aspects of Ted would be silly. That is why Ted is a role-player and not an all-pro.

 

[PatsFanInVa]

I can see that this 10:1 ratio appears arbitrary, so let me explain my method. Against the run, I considered that an active and durable player will accumulate tackles; that a “playmaker†will accumulate Stuffs; and that those two aspects of a defensive end are equal in importance. The 10:1 ratio is a byproduct of that equivalence, and based on the frequency of those statistics among all Defensive ends. Maybe I am overrating tackles for loss; but maybe you are overrating “tackles-for-gain†(re:Ray Lewis). As far as certain team philosophies spreading out statistics and thus reducing the stats of good team players, I think this is a fallacy, and I’ll explain why. Rodney Harrison I’m sure would be at the top of the safety list for his 2003-4 seasons, and Ty Law would do very well among corners for 2002-3. Seymour’s 2003 season, when he established himself as a great DL in football, was monstrous statistically...

A few items here. First of all, once again, good work on the collecting, and I only take issue (right now) with the TFL issue. Your first point is that a TFL reflects two other positive abilities for D linemen, aspects also reflected in stuffs and total tackles. What you've just shown is that you've further skewed the stats in favor of TFL leaders. Consider: they're placed above peers by virtue of the total tackles and/or stuffs. You've predicted that if they will accumulate both statistics, they will be placed further above their peers - even though they are already "rewarded" in this system by the stats themselves! I get your point - that the combination of skills needed for TFLs is rare, but rare to the tune of 10 tackles for gain?

My point is to value the player for what he brings to the game situation at hand (which admittedly is inherently subjective and difficult to model.) But I think you've gotten "too clever by half" here, in trying to identify one "super stat" that really separates the wheat from the chaff... just my two cents worth.

Oh, and about "overvaluing" a tackle for gain? I'm not sure you can do it. Without the tackle, the play is a touchdown, or someone else's tackle (or several someone else's tackle.) The plain old vanilla tackle is, to my mind, the heart of these statistical analyses. That's the point of a defense, to get the other guy on the ground ("not one more yard" is always fine... for every solo tackle a guy makes, ten guys don't. It can only be inflated by horrible tackling among surrounding cast.)

I made this point unsuccessfully here in regards to tackling cornerbacks (Law is obviously one, Warfield was the subject of the discussion.) I maintain that if you can't shut 'em down, there is a place in the world for making sure you get 'em down - even for DBs. A tackling corner is fine by me, so long as he isn't getting torched for 100 yards every game (and Warfield wasn't... just this one time against Oakland.)

Okay, enough for now. Thanks again for doing the work... and don't be surprised if I go after other weightings down the road.:violent:

 

[arrellbee]

PatsFanInVa - I understand your question about whether PonyExpress' formula is 'appropriate' or 'valid' or a 'meaningful' way rank DLs. But there are a couple perpectives. One is that PonyExpress has simply chosen a formula with a reasonable way to put numbers to performance - not necessarily 'THE' most valid or best measure. It's just one formula and can provide some interesting rankings to consider - perhaps very interesting. You may have a different formula that you would consider to give a more appropriate weight to certain areas of performance. Great - that would give us a different perspective, but probably no 'better' or 'righter' than some other formulas. Because there likely is no one 'best' or 'rightest' formula - as you may perhaps agree. So, if you have the time, it would be of considerable interest to see how rankings might come out with a formula that you like.

 

[PatsFanInVa]

PatsFanInVa - I understand your question about whether PonyExpress' formula is 'appropriate' or 'valid' or a 'meaningful' way rank DLs. But there are a couple perpectives....

So, if you have the time, it would be of considerable interest to see how rankings might come out with a formula that you like.

Not a bad point, Arrell, and a challenge I'd respond to, if I did have the time. I've done this sort of spreadsheet exercise before, though not here, and I know just how much time is involved, at least for me. I also have a day career and "night school" via distance learning, so it would be foolish to go beyond my "few minutes here and there" level of contribution, certainly until after May.

Pony, if you view my input into the weighting of your stats as "sideline sniping", feel free to disregard it. It was meant as honest critique, so if I went past that in my tone, I do apologize.

PFnV

 

[gomezcat]

That's some interesting and valuable work, Pony Express. However, IMHO, to really measure a DL's effectiveness, you would have to look at what he contributes to "yards not gained" (for want of a better term) as a whole. As others have said, Sey spends the entire game being double teamed, which must be worth a lot of "yards not gained" to the defence. Think of the big gains that aren't made because he has closed a gap or two at the line, and then think of the pressure he allows to be caused because he attracts a double team on every passing play. Warren seems to do the same on the left. In the games I saw last year, he seemed to stuff the run a hell of a lot, or at least force the running back towards a pile of tacklers.

 

[drpatriot]

Have you ever seen http://www.footballoutsiders.com, Pony? A site that does similar statistical analysis, except with offensive players and defenses on the whole. I bet they would be very interested in your ranking system.

 

[PonyExpress]

Not a bad point, Arrell, and a challenge I'd respond to, if I did have the time. I've done this sort of spreadsheet exercise before, though not here, and I know just how much time is involved, at least for me. I also have a day career and "night school" via distance learning, so it would be foolish to go beyond my "few minutes here and there" level of contribution, certainly until after May.

Pony, if you view my input into the weighting of your stats as "sideline sniping", feel free to disregard it. It was meant as honest critique, so if I went past that in my tone, I do apologize.

PFnV

PatsfaninVA,
I am grateful for all responses and input. In hindsight I think I should have titled the thread "How good was McGinest". I think that would have elicited more responses. All of the critiques people are making have some validiity, and whenever a statistical model is made, all intelligent people should question its methods. As far as DEs getting credit for "team play", in other words, getting credit for their teams overall defense against the run, I dispute this and I'll explain why. One gap 4-3 defensive are tasked as playmakers, and over a two year period I believe their excellence is expressed as production, i.e. by how many plays they make. The same can be said for 3-4 OLBs. These are playmaking positions, and should be judged on how many plays they make. I prefer simple statistical models, some might say overly simplistic. But I think the two year window allows me to exclude certain stats from my model. First, I don't have to worry about varying levels of competition, because I feel over a two year period all teams in the NFL face generally equal opposition. Second, a two year window mitigates against statistical anomalies. Third, I am not concerned with "Performance on a given play vs an average replacement" like footballoutsiders, because I am not concerned with talent, but production. A tremendously talented player who plays rarely but when he plays dominates in my model is less productive, and therefore less valuable than a steady average player who is durable enought to stay in the lineup. As a final note to your argument that "tackles for gain" are being undervalued, I would like to point out that if a defense performs poorly, and is unable to get off the field, more "tackles for gain" are likely to be distributed throughout their defense, whereas a great defense will accumulate fewer "tackles for gain" and comparitively more tackles for loss. True, the players making many tackles for gain are often very active and durable. The players making tackles for loss are often dominant playmakers.Therefore I believe the 10:1 statistical ratio, which emerges organically from my equivalence of those aspects of an NFL defensive end against the run, is not unrealistic.

 

[PonyExpress]

That's some interesting and valuable work, Pony Express. However, IMHO, to really measure a DL's effectiveness, you would have to look at what he contributes to "yards not gained" (for want of a better term) as a whole. As others have said, Sey spends the entire game being double teamed, which must be worth a lot of "yards not gained" to the defence. Think of the big gains that aren't made because he has closed a gap or two at the line, and then think of the pressure he allows to be caused because he attracts a double team on every passing play. Warren seems to do the same on the left. In the games I saw last year, he seemed to stuff the run a hell of a lot, or at least force the running back towards a pile of tacklers.

Gomezcat,
Seymour and Warren both showed up as excellent playmakers in my ranking of Defensive Tackles, which I did last week. In fact Warren emerged as one of the top DTs in the league over the last two years, just as your eyes were telling you. Seymour would have placed in the top 5 but his durability issues hurt his standing.
http://www.patsfans.com/new-england-patriots/messageboard/showthread.php?t=33865

I think as Pats fans we should be less sensitive to statistical studies. I sense a reluctance to scrutinize our own heroes with numbers, and I think it is mostly unjustified. As my studies have shown, all the Patriots involved have been in the upper third among their peers. Keep in mind that the Defense was sub-par last year, and the players suffered injuries. That is bound to be reflected somewhere.

 

[gomezcat]

Gomezcat,
Seymour and Warren both showed up as excellent playmakers in my ranking of Defensive Tackles, which I did last week. In fact Warren emerged as one of the top DTs in the league over the last two years, just as your eyes were telling you. Seymour would have placed in the top 5 but his durability issues hurt his standing.
http://www.patsfans.com/new-england-patriots/messageboard/showthread.php?t=33865

I think as Pats fans we should be less sensitive to statistical studies. I sense a reluctance to scrutinize our own heroes with numbers, and I think it is mostly unjustified. As my studies have shown, all the Patriots involved have been in the upper third among their peers. Keep in mind that the Defense was sub-par last year, and the players suffered injuries. That is bound to be reflected somewhere.

Thanks for your post on your DTs as I hadn't seen that. I take your point about not wanting to see Pats in terms of their numbers. I guess I don't want people to just be judged on "numbers". As you know, they can be misleading, in terms of who is single teamed, double teamed etc.

 

[ilduce06410]

the pain hasn't stopped

Schobel is quite good. Yes, the O-line was extraordinary in the playoffs, and generally unappreciated. Willie's run defense was noted by the handful who did the tape breakdowns or read them last season .

i don't think they'll be able to ask anyone else to carry willie ray's responsibilities. ergo, he defined that position and it vanishes when he goes.
IMO the D will heve to be reconfigured with bruschi, warren and the OLB on that side splitting stuff up.
but then there's the 'presence' thing. opponents had to watch out for mcginest every play he was on the field, which gave big sey some light.
BB/SP must believe they've got something good working to let an asset go like that.
i'll check out more browns games tho.