Draft Needs According to 2010 EPA - Defense

Measuring defensive players is trickier than offensive players for a couple reasons. Most notably, +EPA (Positive Expected Points Added) captures half the story at best. But the idea is that the part of the story we see will correlate well with the part we don't see. In other words, "playmaking" defenders, in most cases at least, are often making unseen impacts as often as they make prominently visible plays.

The table below lists the 2010 regular season +EPA totals for each team by position. Although it doesn't consider free agent losses or injuries, this might be considered a good starting point for determining defensive draft needs.

Unfortunately, defensive positions are not so cut-and-dry. Depending on the base scheme, there are varying numbers of DEs, DTs, and LBs. (The 49ers don't even have a DE position.) Plus, players can sometimes be designated different positions from week to week. Advanced NFL Stats ultimately classifies players according to their most frequent designation in each game's official playbook.

So take these numbers with a heavier does of salt you otherwise would. In fact, it may be wiser to consider the DE and DT positions as one whole, at least for the 3-4 teams. Further, a team's scheme may use its linemen as space-eaters, allowing LBs to make plays, while some teams may expect their linemen to penetrate more often and make more plays.

The table lists each team's league-wide rank for each position. The top third is shaded in green and the bottom third is shaded in red. Defenses that ran a base 4-3 were a minority for the first time as of 2010. There were only 15. To make apples-to-apples comparisons easier, an asterisk leads the 4-3 team abbreviations, so sorting the Team column will group like base defenses together.


Compared with the table for offensive players, there aren't as many basket-cases. A sieve of an offensive line or a poor QB could ruin the entire offense, but perhaps individual defensive positions are not as inter-dependent on their performance. In other words, a good safety can make a big impact despite (or because?) of poor line play. Just a thought.

  • Spread The Love
  • Digg This Post
  • Tweet This Post
  • Stumble This Post
  • Submit This Post To Delicious
  • Submit This Post To Reddit
  • Submit This Post To Mixx

15 Responses to “Draft Needs According to 2010 EPA - Defense”

  1. Jacob Stevens says:

    Wow look at Arizona. Folks seem to think Dockett and especially Wilson had bad years. The loss of Damsby was noticeable. But they really are a Half-decent QB away from being pretty talented again and could have quite a turnaround year.

  2. Anonymous says:


    You wrote:
    "Further, a team's scheme may use its linemen as space-eaters, allowing LBs to make plays, while some teams may expect their LBs to penetrate more often and make more plays."


    In the second part of the sentence, did you mean "...expect their DE/DT to penetrate more often..."? Or did you mean that BOTH the D line and LB would penetrate and make plays? Both are possible, but it looked like you were talking about D Line...so I wasn't sure...

  3. Brian Burke says:

    Oops. Yes, that's what I meant.

  4. Brian Burke says:

    Sorting by team, you can see that the top DE teams are all 4-3 schemes. But that doesn't mean we should just through out the table. For example, notice that Jacksonville is 20th in the league despite running a 4-3. Tampa bay is 18th. Those teams should be looking for upgrades at DE.

    On the flip side, notice that the 3-4 teams lead at LB. But BUF is 16th in the league. They should be looking to upgrade there.

    By the way CIN gets classified as a 3-4 at nfl.com, but I don't think that's right. Could someone set me straight? Their EPA profile looks very 4-3 and clicking around google tends to indicate they were still 4-3 last season.

  5. Dr Obvious says:

    Brian - Isn't Buf going back to a 4/3.

    I find it interesting that very few teams have green for both CB and S. Don't know if it means anything, but it is interesting.

  6. Anonymous says:

    so oakland, have one of the greatest CBs to play the game and they get 31st in the league, surely the rest of them werent that bad.

    and with 1st in the safety category surely it might be better to take the secondaryof each team as a whole ratehr than individualy?

  7. Joseph says:

    Brian, to piggyback on Anonymous' post above, I think he may have a point--group the secondary.
    For my Saints, I would think the S & CB category are flip-flopped. Cause I can say, if we need to upgrade anyplace, it's not the CB's--it's SS. Porter & Greer are solid starters, and last year's 1st rounder Robinson is improving, but was a green rookie last year. Jenkins played good FS in his 2nd year after replacing Darren Sharper. But SS is the weakest point back there.
    From what I have read, Oakland's #2 CB, Stanford Routt (sp?) didn't play too bad either--it was their #3 CB that got picked on.

  8. Brian Burke says:

    I agree. The CB position in particular poses big problems for +EPA. There's this paradox where the better a CB is, the less often he is thrown at, and the fewer opportunities he has to 'make plays.'

    On the other hand, are those OAK or NO CBs really that good? Could they be overrated? (Not saying they are, just posing the question.)

    Also, a weak nickelback who gets picked apart *does* mean CB is a need, no matter how strong your top 2 guys are.

  9. Anonymous says:

    While the Seahawk defense is technically a 4-3, it does look a lot like a 3-4. In the base defense there are 3 DT-sized linemen and one OLB-sized lineman. Both the 1-tech and 5-tech positions play a lot of 2-gap while the 3-tech and the LEO are the dedicated pass-rushers.

    Later in the year, after injuries to multiple 300lb+ 5-techs, Raheem Brock played the position as a more traditional 4-3 end and ended up producing sacks at a relatively rapid pace.

  10. David Myers says:


    I look at these numbers and I wonder. And in this, I'm talking purely as a Dallas Cowboys fan and not an analyst, as I haven't broken down any play by play numbers yet.

    Wade Phillips was fired in the middle of last year because the defense was a mess, and perhaps more so, the idea that opposing coaches were inside Wade's OODA loop. Perhaps he went to the well once too often, perhaps a player was tipping off plays, perhaps the loss of a DC left him with too little time to actually shuffle his defensive game plan around to the point it seemed different.

    So I look at the Dallas numbers, especially for DT, where Jay Raitliff mans the position, and I can only wonder: to what extent do these numbers reflect the player, and to what extent does EPA reflect intelligent coaches getting players into position to make plays? Can these effects ever be separated?

    David Myers.

  11. Joseph says:

    Replying to Brian's most recent comment--
    1. I don't know OAK's CB's that well (as a Saints' fan, I'm more an NFC guy), but Nnamdi is of course great, and they just extended Routt. My guess is that he really is good since they're letting Nnamdi go. The 3rd CB (I think his name is Johnson) isn't bad, from what I understand--he's just not Nnamdi. IIRC, OAK keeps their CB's on specific sides--and teams can scheme to get their #1 WR in the slot against OAK's #3CB.
    2. Regarding the Saints' CB's, Greer had Revis/Nnamdi-type stats in the Super Bowl year. Last year, he was just average. Porter had pretty decent stats both years. The difference was Sharper 2 yrs ago vs Jenkins last year. Jenkins will be good, but Gregg Williams didn't give Jenkins the freedom to cheat like he did Sharper. Their pass rush wasn't as good, nor was the offense quite as unstoppable. Two yrs. ago, with the prolific offense, teams were forced to play catch-up, which allowed the D-line to pin their ears back, and with Williams blitzing like crazy, the CB's knew that the QB was going to get the ball out in a hurry. Thus, they (and Sharper) could sit on shorter routes, and jump them frequently. Last year, well, not so much. Plus, the rush defense was horrible 2 years ago, and last year's was top-half (of the league). So, teams ran less, threw the ball a little more, and the stats reflect that.
    Having explained that, both are solid starters, and would start on probably every team in the league. Their #3 CB is last year's 1st round pick--and, like most rookie CB's, had pretty bad stats. According to the team's beat writers, he improved over the course of the season.
    So, I don't expect the Saints to take a DB this weekend, unless someone really falls--and even then, I think it would be a strong safety in the 4th round or later. Most mocks expect them to take a DE, and it seems to be the biggest need on the team at this point.

  12. Anonymous says:

    How does San Diego rank so low in four of the five categories despite having such a good overall defensive EPA?

  13. Alvin says:

    I really didn't like Nick Fairley's demeanor when the Lions picked him up. He was on his phone for like 3 minutes after he was selected. I haven't heard other people talk about this, so maybe I'm over-exaggerating.

  14. Anonymous says:

    Wow...this is just plain absurd.

  15. The Wizard says:

    Hello Brian,
    Regarding defensive player ratings, and as I recall, you said that if they were involved in the play (specifically, "Defensive players are credited for WPA when they tackle or sack the ball carrier, are credited with an assisted tackle or sack, cause a fumble, defend a pass, or are flagged for a penalty"),

    However, wouldn't this cause some players to be either over valued?
    For instance, defensive lineman when making tackles are likely to make them after short gains. But when they miss a tackle, they are not even 'penalized'.

    it might be harder to do (like looking at each play on film) but wouldn’t it make more sense to involve the whole 'side' of the defense in the wpa? or epa.. or something along those lines.

    Here is a proposal. Have intervals of yardage of say 0-4, 4-8, and 8 plus (you can decide what to do with values on the border). Let’s start with DL’s. if they make a tackle, they are clearly in on the play. If a play is run to their side, and they do not make a tackle and it is for less than 4 yards, there is no effect on them. If it is run for more than four yards, they get ‘credit’ for being in on the play.

    For linebackers, it works similar, only if the play goes for more than 8 yards, they get ‘credit’ for being in on the play as well as any tackles they make.. etc etc.

    I know there are drawbacks to this, but it seems to me there are bigger drawbacks to the way it is done.

    Just my thoughts.

    Thanks and keep up the great work.


Leave a Reply

Note: Only a member of this blog may post a comment.