FO Interview: Tony Dungy

FO Interview: Tony Dungy
FO Interview: Tony Dungy
Photo: USA Today Sports Images

interviewed by Aaron Schatz

Readers may remember that back in Week 16, I had the opportunity to visit NBC's Football Night in America and watch that Sunday's games with the whole FNIA staff, including former Indianapolis and Tampa Bay head coach Tony Dungy. Dungy is as friendly as everyone says, and was happy to indulge us in some questions about game strategy and his personal coaching philosophies. Thanks to various members of the FO staff for helping me come up with the questions.

Aaron Schatz: With your success in Tampa Bay and Indianapolis, a number of teams adopted similar coverage-based schemes to the "Tampa-2." Over the last couple years, though, it seems the scheme has been less popular. What do you think has changed? Do you think more teams should go back to more coverage-based schemes, or have changes in the league made it harder for teams to build good defenses around the "Tampa-2"?

Tony Dungy: To play that style of defense you need patience and you have to be committed to it. It's not flashy and it's not high risk/high reward. It's just basic, sound football. There are weaknesses that you have to understand and it takes a lot of discipline for the players to be able to play it effectively and have confidence in what they're doing. It's probably a system that's more conducive to the days before free agency where you could keep veteran players around and they would be familiar with it. But more than anything, just like any system, it just takes coaches who know it and believe whole-heartedly in it.

Aaron Schatz: For obvious reasons, people equate your defensive philosophy with Monte Kiffin's. However, our numbers show that while the defense was dominant against the pass both before and after you left Tampa Bay, the run defense was consistently better after you left. Was that a philosophical difference between you two, or was it merely the result of changing personnel?

Tony Dungy: It was somewhat personnel but more a philosophical thought. Monte plays more eight-man fronts and blitzes more. He is determined to stop the run and force more second-and-long and third-and-long situations. I was more tuned into taking away the big play and making offenses be perfect. The Bucs actually didn't play a lot of Cover 2 after I left.

Aaron Schatz: In Tampa Bay, you split carries with Alstott and Dunn, and the year you won the Super Bowl, you split carries with Rhodes and Addai. As teams seem to be moving to more reliance on a two-back system, does it make more sense to have players with different, complementary skills (like Alstott and Dunn) or a rotation of interchangeable backs who have good all-around skills (like Rhodes and Addai)?

Tony Dungy: I don't think it matters what type of backs you have. You have to maximize the talent of the guys on your roster. Sometimes you can do the same things with both backs and sometimes they may be completely different, as with Mike and Warrick. I do feel though, that you need at least two backs you can rely on in this day and age. Guys like Michael Turner or Edgerrin James, who can be every down runners and last a whole season, are few and far between. The defenders take a toll on running backs and splitting the carries to have your backs fresher in the playoffs is important.

Aaron Schatz: Your Colts team actually was fairly "vanilla," schematically, on both sides of the ball. We knew what we were going to see: the defense would play Tampa-2 and rarely blitz, while the offense would generally go three-wide and attack based on the stretch run and play-action off of that. You did what you did, and you did it really well. What are the advantages of using relatively simple tactics, as opposed to the variety you'll see in Sean Payton's offense or Rex Ryan's defense?

Tony Dungy: There are two schools of thought. You can be basic and spend all your time on what your opponents are going to do. Or, you can be very multiple and complex, put a lot of doubt in your opponents' minds and create problems for them with the volume of things you do. I always liked to be fairly simple because you could get more players ready to play quickly. If you lose players to free agency, injuries, etc., it is easier to get young players ready to play in a less complex system. However, you don't give your opponents as many problems when they face you. I guess I'm just old school, though. Coach Noll used to tell us the first thing you had to do to win was make sure you didn't beat yourself -- that always stuck with me.

Aaron Schatz: Since at least 2003, the Colts have played predominately no-huddle with extraordinary success. Why don’t more teams try to emulate this strategy? Are other quarterbacks not capable of doing what Manning does pre-snap, or are coaches not willing to cede autonomy and limit formations?

Tony Dungy: The no-huddle puts a lot of pressure on the defense and gives your quarterback extra time at the line of scrimmage to get to the right play. You're seeing more and more college teams going to this type of attack. I think you'll see more NFL teams use it as they get quarterbacks who've been comfortable doing it. The biggest thing you need to be successful with it is a quarterback who wants to be involved in the decision-making process and not just merely want to execute plays sent in to him.

Aaron Schatz: How must defensive coordinators adjust to the ever-increasing use of the shotgun around the league?

Tony Dungy: The shotgun really isn't that much of a strategic variance from being under center. Teams still have running plays from the gun, and they still do most of the things they do from a conventional set. The shotgun is a little tougher to blitz because the quarterback has a little bit of a cushion. However, we never treated the shotgun itself as anything different -- we always evaluated what the other team did out of it and tried to attack their philosophy, not their formation.

Aaron Schatz: You had great success coaching both a defensive-dominated team (Tampa Bay) and an offensive-dominated team (Indianapolis). How much did the relative strengths of these teams affect your game-management decisions, such as what plays to run on third-and-long, when to punt, playing at a quick or more deliberate pace, etc.?

Tony Dungy: Coaches absolutely play to their strengths and minimize their weaknesses. In Tampa, we were built around defense -- most of our money was spent on that side of the ball. We played a slower-paced game on offense, running the ball, taking the 40-second clock all the way down, etc. We wanted to be in low-scoring games and we wanted the other team's offense to get impatient. In Indy, it was just the opposite -- we played up-tempo and tried to pressure people into mistakes with our offense. So I think you do choreograph your team's style to fit your assets. In Tampa, we wanted to have the lead at the end of the game with our defense on the field or our offense trying to run out the clock. With the Colts, we never minded getting the ball back down by a score at the end of the game.

Aaron Schatz: Head coaches in the past were usually chosen from the ranks of coordinators, but it seems the success of Andy Reid, John Harbaugh, and now Jim Caldwell may indicate that coordinator experience might not be necessary to be a good head coach. After all, most head coaches do not actually have play-calling responsibilities. What did your long experience as a defensive coordinator teach you that a positional coach would not know?

Tony Dungy: I do not think being a coordinator is a requirement, but it does give you some advantages. You get used to working with half the team and you go through the game-planning and decision-making processes. You have to think about the good of the entire unit and blending the various skills of your players into a cohesive unit, just as you do as a head coach. You get the experience of running staff meetings and dealing with personalities of coaches who report to you. All of these provide great experience for things that will come up when you're a head coach. The biggest advantage is you interact with 20-30 players and not just the 4-10 you may have as a position coach.

Aaron Schatz: Which one player do you feel improved the most during your time as head coach in Tampa Bay, compared to how good he was when you first acquired him? Which player improved the most during your time in Indianapolis?

Tony Dungy: In Tampa, I'd have to say John Lynch. You could tell he was going to be a good player when we got there, but he had been a back-up defensive back who they experimented with at linebacker. He went from being a good situational player to one of the top safeties in the league. A lot of it was his training and studying, but some of it was due to the perfect marriage of a player's skills and the defensive system.

With the Colts, I'd have to say Reggie Wayne. Another good player our first year, Reggie developed into not only a No. 1 receiver, but one of the best in the league. Just repetitions and confidence. Probably the common ingredient for both guys was exceptional intelligence. They both really understand the game.

Aaron Schatz: Pierre Garcon had a breakout season in 2009, but he was drafted when you still coached the Colts. What did the Colts see in him that other teams missed?

Tony Dungy: Pierre played at a very small school and was extremely raw. He was so much better than the guys he was playing against in college that it was tough to evaluate. I don't think people missed him, though. We knew he was strong and fast but no one could have predicted how quickly he made the jump to big-time football -- understanding routes, coverages, adjustments, etc. That's the thing you never know.

Aaron Schatz: What's gone wrong for Tony Ugoh? Why didn't he develop as you and Bill Polian expected?

Tony Dungy: Tony is very athletic and has all the skills to play the position well. Hopefully he will get back to the form of his rookie year. I'm not sure what the problem was this year, but I know the Colts felt Charlie Johnson was playing better.

Aaron Schatz: In a playbook sense (i.e. game strategy, not management style), what is one thing you learned from Chuck Noll and Bud Carson that still applies to today's game? What's one thing that doesn't really apply today?

Tony Dungy: Back then, we didn't have to adjust to a lot of formations and personnel groups. If you studied, you could know what opponents were running and we had a lot of checks based on backfield sets or where different key players were lined up. Today, offenses do a much better job of not telegraphing their intentions. However, the key components of what they taught still are in play. You win with good, solid, fundamental football. Getting in position to tackle, playing good physical coverage and hitting receivers whenever possible are critical. And they always looked for speed and quickness over mass and bulk. It's great to have both, but if you can only have one, take the speed and explosiveness.

Aaron Schatz: What's it like watching football in television after all those years of not just coaching but also breaking down game film? Are you generally able to recognize coverages when you are watching games for FNIA, or do you feel as hamstrung by the standard TV camera angles as the rest of us?

Tony Dungy: I enjoy watching football on TV. Many times I can visualize the whole field, even on the TV screen, because I still know enough about what the teams are doing. Sometimes when I see a replay I'll realize I was wrong and say, "Oh, that's how that guy was open, or that's where that blitz came from." But most of the time I have a good idea. But I'm really enjoying my work with NBC -- especially trying to transmit some of that information to the fans and let them see some of the things teams do to take advantage of their strengths and hide their weaknesses.

Aaron Schatz: A lot of us wonder what could have happened if more enlightened coaches had left you at quarterback once you got to the NFL. What modern quarterback do you think has a style most similar to how you would have played as an NFL quarterback?

Tony Dungy: We had a lot of quarterbacks playing when I played who could have brought some things to the NFL. Quarterbacks like Seneca Wallace and Ben Roethlisberger, who could move around and put pressure on a defense by running or throwing on the run. The game has changed since the '70s as more coaches have looked at ways to make it tougher on defenses. But I'm glad my career turned out the way it did -- I would have never gotten a chance to learn the defensive side of the ball.


97 comments, Last at 27 May 2010, 5:02pm

2 Re: FO Interview: Tony Dungy

Thanks! This was an enjoyable read. None of the answers were big surprises, but this was still far more interesting than the offseason content from most sources.

Also, I thought that the questions were excellent, with one exception: asking about Tony Ugoh. I'd put the odds of Dungy (or most other people who haven't been out of the game for long) giving specific criticism about a specific player below 1%.

11 Re: FO Interview: Tony Dungy

I disagree regarding the Ugoh question. I think it's telling that Dungy's response wasn't something like "he just needs to focus on fundamentals and keep working hard," but instead was "I don't know." That's a lot more worrisome.

You don't get specific criticism when you ask questions like that, but the responses they give can be helpful. You just have to read through the lines.

20 Re: FO Interview: Tony Dungy

In a sense, though, that was his response, wasn't it? Ugoh is athletic and has the skills, so it must be something else ... "fundamentals and hard work" sounds like what you could take away from what Dungy didn't say.

I think I agree more with dmb, though. I have a difficult time picturing Dungy saying something specific and negative about a player during an interview.

I think it's near to hear him say that he can usually picture the rest of the players based on what he sees on a regular camera.

25 Re: FO Interview: Tony Dungy

You're right that sometimes you can read between the lines for that sort of response, but that also leaves one prone to simply reading too much into it.

56 Re: FO Interview: Tony Dungy

Coaches these days know how to answer questions like a good politician and Dungy's answer to the Ugoh question is a great example. Dungy praised Ugoh's abilities, offered hope for the future, and avoided any criticism. Hillary Clinton could have given the exact same answer, all the while thinking to herself: "Who the hell is this guy Ugoh?".

4 Re: FO Interview: Tony Dungy

I could have kept reading Dungy' answers for a long time. So good, and he did not do the football jargon that those of us who did not play do not always understand. Also, I loved the lack of "jock speech" in his answers.
great interview

5 Re: FO Interview: Tony Dungy

That was as interesting a conversation with a coach as I've read in years.


6 Re: FO Interview: Tony Dungy

I remember when he got passed over for several jobs when he was an assistant with the Vikings because he was "too cerebral". Well, the reason that he's one of the best on TV right now is because he's more cerebral than the rest of those clowns.

Kudos to Coach Dungy.

Sports talk radio and sports message boards are the killing fields of intellectual discourse.

64 Re: FO Interview: Tony Dungy

Yep, that was why he was passed over for so many jobs, because he was "too cerebral". He was definitely considered too something (do I need to say what it really was) back then, but I don't think it was cerebral.

65 Re: FO Interview: Tony Dungy

I assume you mean being black?

I don't think that was the only reason. He doesn't fit the old school image of a football coach, barking out instructions, spittle spewing from his mouth, steaming mad when something goes against him.

I bet if he had the personality of Ditka or Cowher he would have got an opportunity earlier.

7 Re: FO Interview: Tony Dungy

That was very well done. Good questions and interesting answers.

8 Re: FO Interview: Tony Dungy

Nice interview.

Now if you could find a medium who can contact Bill Walsh, Vince Lombardi and George Halas for you then you would have some headline interviews. Until then Dungy is a good start.

9 Re: FO Interview: Tony Dungy

Seriously, those questions were great. I often think people underestimate how important the questions are to eliciting a good interview.

I particularly like the one about how he switched from Tampa to Indianapolis, teams with opposite strengths. I wonder how many coaches have done well with both kinds of teams?

Now if only Chris were around in the off-season. I'm sure he'd light this board up.

12 Re: FO Interview: Tony Dungy

Shula to me is the guy who really stands out as the "willing to switch your philosophy based on your players" rather than "I have my scheme and we will run it whether or not we're able". Shula had his power running teams in the 70s, then became pass-first with Marino. Whatever worked.

I'm not a huge fan of Dungy on TV, just because he's so restrained. I keep thinking that they need to combine him and Rodney Harrison and divide them again, and you'd then have two good commentators. Harrison's just trying to be obnoxious and confrontational, but Dungy needs a little bit of that to make his points more forcefully. That being said, he does come off really well in this interview, would love to see him writing regular in-depth pieces on football just to see what's going on in his head.

26 Re: FO Interview: Tony Dungy

Seriously, those questions were great. I often think people underestimate how important the questions are to eliciting a good interview.

Agreed. Often you can learn nearly as much about the interviewer as you do about the interviewee.

10 Re: FO Interview: Tony Dungy


Typically, aren't the questions in italics and the answers in normal text? I found this way really hard to read, actually.

15 Re: FO Interview: Tony Dungy

That didn't occur to me while reading it, but you're right. Though typically the questions are much shorter than the responses so it tends to make sense to have the longer part be in a normal font. here the questions and the responses seem to be closer to equal in length. perhaps other method should have been used to visually differentiate question from answer.

16 Re: FO Interview: Tony Dungy

We've never really established a format around here. Sometimes we do it one way, sometimes the other. We probably should go back and standardize the past interviews, and those in the future, especially if we are going to do these more often -- which I would certainly like to do.

22 Re: FO Interview: Tony Dungy

You could go the William Faulkner route and italicize entire chapters... even if they are one sentence long... one very long sentence long. Or his stylistic heir, Garcia Marquez, and ask one sentence questions that take fourteen minutes to complete.

Glad I can be so helpful.

27 Re: FO Interview: Tony Dungy

I'll echo the sentiments of others by expressing my hope to see more interviews with current and former coaches and front office staff. And when you think about it, many of those people would probably rather field questions from FO than from most other media outlets; surely these questions were more interesting to Dungy than the typical "is Mike Vick really a changed man???" drivel.

14 Re: FO Interview: Tony Dungy

I have a complaint: There's no more interview to read. :(

All joking aside, you could've made this twice as long, and I would've loved to have read every single bit of it. Thanks for doing this!

17 Re: FO Interview: Tony Dungy

I have always loved Tony Dungy. I remember how, in pregame interviews, most coaches would give answers like, "We're going to try to establish the run and stop them from running, and we're going to try to pass and stop them from passing." But Tony Dungy would give answers like, "We're going to try to isolate their safety against our TE."

Great interview!

19 Re: FO Interview: Tony Dungy

Terrific content, great questions, and awesome of Coach Dungy to do this.

Kudos, F.O.

21 Re: FO Interview: Tony Dungy

Interesting bit about Wayne and Lynch. I've often wondered about paired teammates at big schools getting drafted high (USC LBers last year, DL men, etc)--Santana Moss and then Wayne about 10 slots later in the first round. Which was better? Which team knew more? Which went to a better place to develop?

31 Re: FO Interview: Tony Dungy

Steve Smith and Dwayne Jarrett; Mario Williams and Manny Lawson; Michael Haynes, Jimmy Kennedy, Anthony Adams (Penn St DL); Shaun Rogers and Marcus Stroud; Vilma and DJ Williams (Miami) and Dontarrious Thomas and Karlos Dansby (Auburn) in the same draft. There are loads to look at for your study and that is before you look at players taken one ot two years apart where the presence of some of the guys that later turn into superstars probably got their team mates drafted (eg, Buchanon and Rumph drafted having played in the same secondary as Ed Reed and Shaun Taylor).

36 Re: FO Interview: Tony Dungy

Here's a comprehensive list (I think) of "unit mates" who were drafted in the first round in the same year, courtesy of P-F-R's drafty query tool. I'm defining "unit mates" somewhat loosely; basically, any pair who could plausibly have a direct impact on each others' performance: QB-WR, QB-RB, QB-TE, WR-WR, WR-TE, RB-RB, OL-RB, OL-OL, DL-DL, DT-LB, LB-LB, or secondary mates. I did NOT include DE-LB, though there were several instances of those (perhaps most notably Courtney Brown and LaVar Arrington going at the top of the 2000 draft). QB-T is another omission I decided to make. It's certainly debateable, but I personally felt that the two generally operate independently enough that someone watching much film would be able to separate the contributions of the two players.

If the same player falls into more than one of these groupings, it's listed separately, unless it seems likely that all three would effect each other substantially. (QB-RB-RB is listed together, but QB-RB-TE is listed as QB-RB and QB-TE.)

Since 1993, there have been 43 of these pairs/groups drafted in the first round. (Again, a few players are double-counted.) This year was actually the first since 1994 that a pair of "unit-mates" was not selected in the first round...

1993: John Copeland (DE, 5, CIN) / Eric Curry (DE, 6, TB) -- Alabama
1993: Rick Mirer (QB, 2, SEA) / Jerome Bettis (RB, 10, LA Rams) -- Notre Dame
1993: Rick Mirer (QB, 2, SEA) / Irv Smith (TE, 20, NO) -- Notre Dame
1995: Ki-Jana Carter (RB, 1, CIN) / Kerry Collins (QB, 5, CAR) -- Penn State
1995: Kerry Collins (QB, 5, CAR) / Kyle Brady (TE, 9, NYJ) -- Penn State
1995: Kevin Carter (DE, 6, STL) / Ellis Johnson (DT, 15, IND) -- Florida
1995: Tyrone Wheatley (RB, 17, NYG) / Trezelle Jenkins (T, 31, KC) -- Michigan
1996: Terry Glenn (WR, 7, NE) / Ricky Dudley (TE, 9, OAK) -- Ohio State
1996: Regan Upshaw (DE, 12, TB) / Duane Clemons (DE, 16, MIN) -- California
1996: Jeff Hartings (C, 23, DET) / Andre T. Johnson (T, 30, WAS) -- Penn State
1997: Peter Boulware (LB, 4, BAL) / Reinard Wilson (LB, 14, CIN) -- Florida State
1997: Walter Jones (T, 6, SEA) / Warrick Dunn (RB, 12, TB) -- Florida State
1997: Ike Hilliard (WR, 7, NYG) / Reidel Anthony (WR, 16, TB) -- Florida
1997: Kenard Lang (DE, 17, WAS) / Kenny Holmes (DE, 18, TEN) -- Miami
1998: Peyton Manning (QB, 1, IND) / Marcus Nash (WR, 30, DEN) -- Tennessee
1998: Greg Ellis (DE, 8, DAL) / Vonnie Holiday (DE, 19, GB) -- North Carolina
1998: Tebucky Jones (CB, 22, NE) / Donovan Darius (S, 25, JAX) -- Syracuse
1999: Jevon Kearse (DE, 16, TEN) / Reggie McGrew (DT, 24, SF) -- Florida
2000: Chris Samuels (T, 3, WAS) / Shaun Alexander (RB, 19, SEA) -- Alabama
2000: Ron Dayne (RB, 11, NYG) / Chris McIntosh (T, 22, SEA) -- Wisconsin
2001: Richard Seymour (DE, 6, NE) / Marcus Stroud (DT, 13, JAX) -- Georgia
2001: Dan Morgan (LB, 11, CAR) / Damione Lewis (DT, 12, STL) -- Miami
2001: Steve Hutchinson (G, 17, SEA) / Jeff Backus (T, 18, DET) -- Michigan
2001: Santana Moss (WR, 16, NYJ) / Reggie Wayne (WR, 30, IND) -- Miami
2002: Julius Peppers (DE, 2, CAR) / Ryan Sims (DT, 6, KC) -- North Carolina
2002: John Henderson (DT, 9, JAX) / Albert Haynesworth (DT, 15, TEN) -- Tennessee
2002: William Green (RB, 16, CLE) / Marc Colombo (T, 29, CHI) -- Boston College
2002: Phillip Buchanon (CB, 17, OAK) / Ed Reed (S, 24, BAL) / Mike Rumph (CB, 27, SF) -- Miami
2003: Jimmy Kennedy (DT, 12, STL) / Michael Haynes (DE, 14, CHI) -- Penn State
2003: Jerome McDougal (DE, 15, PHI) / William Joseph (DT, 25, NYG) -- Miami
2004: Jonathan Vilma (LB, 12, NYJ) / D.J. Williams (LB, 15, DEN) / Vince Wilfork (DT, 24, NE) -- Miami
2005: Ronnie Brown (RB, 2, MIA) / Cadillac Williams (RB, 5, TB) / Jason Campbell (QB, 25, WAS) -- Auburn
2006: Mario Williams (DE, 1, HOU) / Manny Lawson (DE, 22, SF) / John McCargo (DT, 26, BUF) -- NC State
2006: Reggie Bush (RB, 2, NO) / Matt Leinart (QB, 19, ARI) -- USC
2006: Ernie Sims (LB, 9, DET) / Brodrick Bunkley (DT, 14, PHI) -- Florida State
2006: Kamerion Wimbley (DE, 13, CLE) / Brodrick Bunkley (DT, 14, PHI) -- Florida State
2007: JaMarcus Russell (QB, 1, OAK) / Dwayne Bowe (WR, 23, KC) / Craig Davis (WR, 30, SD) -- LSU
2007: Michael Griffin (S, 19, TEN) / Aaron Ross (CB, 20, NYG) -- Texas
2008: Darren McFadden (RB, 4, OAK) / Felix Jones (RB, 22, DAL) -- Arkansas
2008: Sedrick Ellis (DT, 7, NO) / Keith Rivers (LB, 9, CIN) -- USC
2008: Sedrick Ellis (DT, 7, NO) / Lawrence Jackson (DE, 28, SEA) -- USC
2009: Matthew Stafford (QB, 1, DET) / Knowshon Moreno (RB, 12, DEN) -- Georgia
2009: Brian Cushing (LB, 15, HOU) / Clay Matthews (LB, 26, GB) -- USC

40 Re: FO Interview: Tony Dungy

Huh. It looks like Sims and Bunkley in Philadelphia will be the first time any of these "unit mates" have re-teamed in the pros (as far as I can tell.) I expect, nay demand, magic from them.

45 Re: FO Interview: Tony Dungy

Shaun Alexander briefly reunited with Chris Samuels on the Redskins in 2008. I don't know how many of Alexander's 11 carries or 24 yards came behind Samuels (probably most ... the Redskins led the league in combined percentage of rushes to left end and left tackle that year), but it should suffice to say that the results weren't exactly glorious.

51 Re: FO Interview: Tony Dungy

To be clear, I don't think the play of Samuels and Alexander have anything to do with that of Sims and Bunkley. I just wanted to point out that the latter won't be the first pair I listed to reunite in the pros.

37 Re: FO Interview: Tony Dungy

Several of the players you named (Smith, Jarrett, Adams, Rogers, Thomas) were mid- or late- 2nd round picks, which I suppose is somewhat "high," but not extremely so. Of course, that makes your ability to remember Adams and Thomas that much more impressive!

Buchanon and Rumph were taken in the same year as Reed, not before him. Sean Taylor did not start for the Hurricanes in 2001 -- he came in as an extra defensive back and also played special teams -- which means he probably wasn't on the field with the other three as much as you suggested.

48 Re: FO Interview: Tony Dungy

I find it interesting to see good players who played together in college and then turn out to be great players in the pros as opposed to them simply being first round picks. I never saw what happened when some poor college lineman had to try to block Casey Hampton and Shaun Rogers (whether at the same time or platooning) but it must have been interesting to watch. Or some college defensive coordinator having to come up with a defense that contains Chad Johnson (as was) and Houshmanzadeh using college standard defensive backs.

I suppose it is also good to try to spot trap picks, ie players boosted by strong supporting casts but that is a fairly time consuming process.

49 Re: FO Interview: Tony Dungy

I think Albert Haynesworth and John Henderson as a DT pair might be the ultimate example of that. Then again, Charles Grant, Richard Seymour, and Marcus Stroud were all at Georgia in the early 2000s -- Grant was drafted in 2002; Seymour and Stroud in 2001. I admit I don't know how they were actually utilized, but it had to have been interesting.

In any case, I agree with you that it's interesting to revisit college units that ended up producing multiple top-level pro players. However, I think Bobman's original post expressed an interest in addressing the second issue: the scouting of college players whose true ability might be difficult to separate from their teammates. I posted the list to allow any interested readers to draw their own conclusions.

23 Re: FO Interview: Tony Dungy

Dungy should have won more than one Superbowl with the personnel he had . He lucked out playing the Bears with Rex Grossman or he still wouldn't have a ring . I'm a big Colt fan and have no idea why people consider Dungy a great coach . His replacement almost went undefeated . One Superbowl appearence in seven years with Manning , Harrison , Freeney , James , Wayne , etc , that's underachieving , I don't care what they did in the regular season .

24 Re: FO Interview: Tony Dungy

Your comment might have some traction if he didn't have great success in TB. He turned around a former laughing stock of a franchise and the team he built won a SB. Very few coaches have that kind of success with 2 different teams, even more rarely with 2 such wildly different types of teams.

Also, I've disputed the idea elsewhere that his Indy teams have been stacked - I think the Colts have had more success with fewer outstanding players than any team I can think of... (my controversial presumption being Manning is the greatest QB of all time and all of his WR's are therefore over-valued. Or rather, I'm not convinced you can say they are great because Manning makes everyone he throws the ball to look like a superstar. Same goes for what he does for his o-line. Who else has Indy had? A one-dimensional pass rusher who has cost them several notable games by being a ridiculous liability against the draw play? An oft-injured safety? A pretty good, but far from HOF caliber RB?) Anyway, I think Indy's success with many mediocre players, especially on defense, is a testament to Dungy. Also, even if Indy is having success now, they're still very much using Dungy's template - would anyone deny that?

38 Re: FO Interview: Tony Dungy

I think Dungy is an excellent coach. But the point remains that he would not have a ring if the Bears had a competent quarterback. Rex Grossman was the reason the Colts won.

As for the interview, I wish Aaron had asked some follow-up questions. Particularly Dungy's answer about QBs not calling their own plays. I'd never heard that some QBs just don't want the added responsibility of calling plays. I'd have loved more on that.

41 Re: FO Interview: Tony Dungy

That's a pretty ridiculous "if." If Seattle hadn't been one of the worst Superbowl participants of the last 20 years (or, if FO would argue, the refs had been on the up-and-up) then Cowher wouldn't have a ring.

And you're also working from the assumption that the Colts couldn't have beaten a good team in Dungy's Superbowl year, which is a theoretical I am in no way willing to concede.

It's patently ridiculous to speculate and try to rewrite history to minimalize a coach's accomplishments...

44 Re: FO Interview: Tony Dungy

Ha! Rex Grossman was the reason the Colts won the Suber Bowl. Are you aware that teams actually have to win games to GET TO the Super Bowl? So what does that say about all the other NFC teams the Bears beat on the path to the SB?

28 Re: FO Interview: Tony Dungy


I think first both the TB and Ind teams Dungy coached were defined by outstanding player growth. I would credit Dungy for this. He built a system which allowed players to develop and that made two franchises into consistent winners.

Second the simplicity approach Dungy speaks of (vs Payton or Ryan complexity) leads to rarely beating oneself but also having difficulty beating others who are at the top of their game. This certainly always felt true.

Also despite Manning's amazing skills there have been times where the no huddle exposed him against the finest defenses minds. LeBeau for instance on the sideline is very likely gonna beat -the play calling- of Manning who is on the field getting hit frequently.

So coaches do many things and I'd say moreso than some others the Dungy's greatness is achieved by how he developed players day to day into stars. While not a liability on gameday by any stretch even those who are naysayers cant deny his record from mon-sat so to speak.

29 Re: FO Interview: Tony Dungy

I think you're overlooking something else: the quality of the upper echelons of the AFC during the time that Dungy was head coach in Indianapolis.

The 2003-04 Pats won I've-forgotten-how-many consecutive games — an all-time record. The Steelers were also a dangerous team that got hot in 2005. The playoff loss to San Diego in 2007 is probably the worst exit the Colts suffered under Dungy, but those Chargers weren't bad either. If Indy had been in the NFC, they'd have probably made another couple of Superbowls.

The 1980s Bears get criticized for the same reason — that one championship was a poor return. Again, look at the context. The '84 Niners, '86 Giants, '88-'89 Niners: them's good teams.

Perhaps the Colts' biggest error was to win the Superbowl in a year when they 'only' won 12 regular season games. If they'd done it the year previously, and slipped up in 2006, they'd have their all-time, 18-win season, as the Bears did, and people like buckturgidson might be happier (though I suspect not).

32 Re: FO Interview: Tony Dungy

I also think it's disingenuous at best to act like James, Harrison, and Wayne were great for all of Dungy's 7 years in Indy.

Take James - he was only even on the roster for 4 of Dungy's 7 years. And in 2002, he was a below average back, clearly not fully recovered from his knee injury. In 2003, James was a good back - 8th in rushing DYAR, but 20th in rushing DVOA, indicating that a lot of his value was from volume; and 46th in receiving - not the great all-round player he was prior to his knee injury. In 2004-2005, James was a great back again, top 3 in both rushing and receiving DYAR. Then he was gone.

Harrison - great from 2002-2006 but mediocre in 2007 and clearly done in 2008.

Wayne - was very good in 2002-2003 and then great for the rest of the Dungy years.

I think it's fair to criticize Dungy's failures with James, Wayne, and Harrison all playing great in 2004-2005, but they were not all-great, all-the-time during the Dungy years.

58 Re: FO Interview: Tony Dungy

Mathis , Bethea , Sanders (when healthy) , Glenn , Saturday . All played in the Pro Bowl . That's at least 10 Pro-Bowlers in the Dungy era , not a mediocre cast . Those are the facts .

60 Re: FO Interview: Tony Dungy

Oops , I almost forgot Dallas Clark , another of the " handful of playmakers " Polian assembled for Mr. Dungy . That makes 11 Pro Bowl selections .

68 Re: FO Interview: Tony Dungy

Really? Bethea and Mathis? That's like arguing the Eagles are good at drafting because of Michael Lewis and Lito Shepard.

Again, I'm not sold on any of the WR/TE's - Manning has made literally every player he throws to look All-World. Let me know when a Colts WR or TE has success on another team or with Jim Sorgi pitching the ball...

34 Re: FO Interview: Tony Dungy

Maybe I have a warped view of how much criticism Dungy gets because of a few certain posters on this site, but it's always been surprising to me how much of it there is when other coaches with similar levels of "underachievement" (e.g, Cowher, Shanahan, Vermeil, Parcells, et. al.) get comparatively little.

35 Re: FO Interview: Tony Dungy

In thirteen years, he has had ONE losing season as a head coach, and that was his first year in Tampa.

He's coached 208 regular season games and 19 playoff games. However, you choose to focus on a sample that is less than one tenth the size of the majority of his body of work.

I'm sure that you are one of those people who believes that if Scott Norwood makes that field goal, Marv Levy suddenly becomes a great coach.

Sports talk radio and sports message boards are the killing fields of intellectual discourse.

39 Re: FO Interview: Tony Dungy

He was a great coach. There was a pretty good GM there too, that might be the real key, as well as the reason to annoint Bellicick and Walsh as the true greats of the modern era, and Walsh did it fair and square ;-)

46 Re: FO Interview: Tony Dungy

Personally I believe the issue with Dungy's teams in the postseason was that they were too vanilla. It's a similar criticism to the one levelled at Marv Levy's four-time Super Bowl Bills.

Basically in the regular season you can win by playing and executing well. But once you reach postseason you're up against the best of the best. All the teams are good and everybody is motivated. If they know what you're going to do it's easier for good teams to stop. I believe the Colts needed to throw in some wrinkles for the postseason.

52 Re: FO Interview: Tony Dungy

If they're not so vanilla, maybe it doesn't come to a last second kick?

We're discussing a game with only 16 regular season games. All sample sizes are going to be small. There is no harm in theorizing why teams win or lose, based on small sample sizes. As long as we don't take it too seriously.

42 Re: FO Interview: Tony Dungy

A great man with a great mind. Well done on the interview, FO! It was wonderful to read.

43 Re: FO Interview: Tony Dungy

I second (third, fourth, whatever) the comments for MORE INTERVIEWS! Former players turned broadcasters would also be fine, too. For my two cents, it might be easier to interview a former player/coach who: doesn't do TV work; or who does radio/TV LOCALLY vs. nationally.
BIIIG coup if you could get it: a former player/coach who would work with FO to do a weekly feature breaking-down film. In other words, Tanier's diagrams with VIDEO. (Not that MT isn't good--he's great; a former player/coach who used to break down film for a living will notice things that even good commentators don't.)

RE: Dungy--The reason he didn't get more SB rings is simply going against great teams in his conference. With TB, he had to go through the greatest show on turf (2 SB appearances in the 3 yrs. before TB won it). With Indy, there have been Brady's Pats, Pittsburgh, Baltimore, and SD--who seems to match up well with Indy.
Dungy mentions that he always designed his teams to not beat themselves. IMO, this leads to a good number of regular-season wins--beating the bad teams with solid execution + talent, and playing well against the good teams--winning some, losing some. The problem is that, in the playoffs, the opposing teams are all good teams who have a similar make-up/strategy (ie, solid execution/talent). That's where the other team had an advantage--they did not have to prepare as much for the Colts/Bucs--they were simple to game-plan for.

55 Re: FO Interview: Tony Dungy

I second this. But please, find men who will give actual insight into the games. More than the typical "try our hardest, give props to the other team, bring our A game" crap that masquerades for intelligent sportstalk. I agree with the local types, too. Personally, I would love to hear Bruce Matthews or Mike Munchak break down how to block tough defense. Or Manning's insights (after retirement, of course) into why he audibles what particpular play. I seem to remember Steve Tasker discussing special teams pretty well, until he shut down by his game calling partner. As Mr. Dungy showed, intelligent discourse without coachspeak or overburdening us with technical football jargon is very insightful.

53 Re: FO Interview: Tony Dungy

Another skip on the broken record - Aaron, you asked good questions, and
Tony rewarded you with actual answers instead of 'coachspeak'. It's also
nice to hear that he is as good a person as he's made out to be.

57 Re: FO Interview: Tony Dungy

In a 3 year span under Dungy the Colts had a first-round bye and homefield clinched twice . Both times they were favored , heavily favored against San Diego . Both times they lost at home . If Dungy was a great coach he would have game-planned and succeeded at least one of those times . As a fan it's hard to get excited over beating the Falcons in week 4 only to see your team choke year after year in the playoffs . And he had many quality players I didn't mention , ie: Mathis , Stokely , June , Thornton , Bethea , Glenn , Saturday , etc. The Colts deserve the title of Atlanta Braves of the NFL .

59 Re: FO Interview: Tony Dungy

Great stuff!

Was reading through hoping we might be able to find the secret behind the Colts' ability to exceed after-the-fact "estimated wins" in every year of the Dungy era (plus last year with a Dungy protege). Delta Whiskey had posted some notes about the ability of the best "drive stat" teams to outperform projections in a prior thread. This comment from Dungy kind of jumped out at me:

"In Tampa, we wanted to have the lead at the end of the game with our defense on the field or our offense trying to run out the clock. With the Colts, we never minded getting the ball back down by a score at the end of the game."

To a degree that's master of the want to be running out the clock, or have your strength on the field if a game is being decided. But, I wouldn't be surprised if a percentage of teams have figured out some strategies for "possession management" with the clock that gives them a percentage boost in the close ones. So, the general stathead theme that the close games break out evenly over the long haul isn't actually correct. If you're looking at 32 teams over 10 years, or 30 years, it seems like they break out evenly. But, maybe there are a handful of teams who are good at possession management...a handful of teams who are horrible at it (inexperienced coaches and quarterbacks, mistake-prone styles), and those cancel out in way that sustains the big picture illusion.

Possession management would be similar to the NBA "two for one" deal in the final minute of a quarter, where the smart teams try to squeeze out an extra quality shot attempt. For NFL teams, it would be halves obviously, with peak priority for the second half of a close game.

Given what we saw with the first DVOA look at the Bills era (when Aaron posted the numbers from the final Bills dynasty season as DVOA research is working its way back season by season as time permits)'s possible that this was something Bill Polian was on to well before the rest of the league.

These components are in play:
*High quality quarterback
*Proven ability to drive the field
*Can drive the field quickly through the air if needed
*Consistently outperforms "estimated wins" formulas that are based on big picture statistical tendencies

You have to recognize the opportunity for a strategic have to have a quarterback who can do it...and you have to have a teamwide commitment to valuing possessions. This seems to be a hallmark of the Dungy/Polian era in Indy, and the Polian era in Buffalo (though future research may not end up showing that once it's done).

I'm not suggesting the Colts are the only one's who do this. Other teams are obviously scary when they've got the ball in the final minute! But, perhaps this is why they overachieve formula expectations that are based on overall production. They don't have a defensive style that creates production mismatches (as Dungy explained). They focus on getting the game over with when ahead rather than piling up blowout stats. If you have an edge in possession management, this style will win more games than production differentials say it should.

Or, if you DON'T have an edge in possession management (maybe because of a young QB), you'll go 9-7 when DVOA stats say you're the most efficient team in the NFL (which turns the meaning of "efficiency" on its head)...and you'll go 0-2 vs. the Colts and lose to veteran QB's like Carson Palmer (twice), Big Ben, Brett Favre, and Tom Brady.

I think when the riddle of ranking teams as best as possible is finally solved, the role of drive stats and possession management is going to be in the mix...and retro-calculating the 2009 season won't show Baltimore at #1 and Indy at #8. This great interview from Aaron helps provide some talking points for that possibility.

67 Re: FO Interview: Tony Dungy

As an Eagles fan, I am under the impression that a good 2-minute drill involves throwing a 5 yard pass to the middle of the field on 1st and goal at the 9. Preferably, with no timeouts and 25 seconds on the clock.

63 Re: FO Interview: Tony Dungy

I'm showing, for 2002 to the present (Dungy years plus 1 of Caldwell), regular season and the playoffs...

73-22 in games decided by 6 or more after regulation
35-14 in games decided by 5 or less (close + OT)

Winning Percentages:
76% in 6 or more
71% in 5 or less (using all OT results in this category)

Does anyone have bigger picture studies for the league? I know it's standard for closer games to be more randomized. That seems like a slighter drop-off than normal for Indy. Pretty impressive to go 35-14 in close games. Would guess this is the main reason that Indy has overachieved "estimated wins" every year. Now we have to decide if fortune has been smiling on them, or if there's a skill involved that ratings systems should be trying to incorporate.

Quick addendum...went to look at the Polian/Kelly years. When they hit their stride from 90-93 (Polian left after '93 because of an internal dispute according to wiki).

17-6 in games decided by 5 or less
41-13 in games decided by 6 or more

Was using pro-football reference here, not sure if they put a star by OT games, so that might be off a smidge.

74% in close games
76% in six points or more

Could obviously be interpreted as "lucky teams post overachieving records," or "there may be a skill set that helps you win more than your share of close games." If it's fortune, it sure seems to follow Polian around when he has QB's who can drive the field.

70 Re: FO Interview: Tony Dungy

Went to look up McNabb's career stuff off the Eagles comments. Did it by hand, so it might not be perfect.

20-23-1 in games decided by 5 or less
78-28 in games decided by 6 or more

Didn't count his rookie season because he was new and the team wasn't very good. Only using games he played...and I left out one where he threw 3 passes before sitting out when the Eagles were resting up for the playoffs late in a regular season.

A numerical expression of the frustration Eagles fans are feeling. From the time since McNabb lost much of his mobility---2007 to 2009---it's 4-8-1 and 25-10. Small sample size, but it might suggest an ability to create in crunch time has been diminished (not that crunch time was great before then). McNabb suffered a groin injury in 2005 and a torn ACL in 2006.

Stark contrast between the numbers of Indy and Philly, which I guess is at the heart of a lot of discussions at the site. The two are very similar in the samples above in games decided by 6 or more...

Colts through the Dungy era +LY: 73-22
Philly in McNabb's games since Y2: 78-28

Colts in close games during that period: 35-14
Philly in close games with McNabb playing: 20-23-1

Again, you may get something slightly different because I did this by hand (and yours is probably right!). Think that lays out the framework pretty well.

71 Re: FO Interview: Tony Dungy

Interesting, would like to look at some of this a bit more; also digging out drive efficiency data...again.

Nitpick - would suggest looking at this as Dungy/Manning Era v. Reid/McNabb era. In this article, Dungy outlined his philosophy w/ respect to the Colts (Clots?) and having Manning to execute. I don't know if we have a similar philosophical view from Reid as framework and I don't enough about the Eagles to say whether the above record can be hung solely on McNabb given Reid's difficulties with clock management, anymore than Dungy's success can be separated from Manning's ability.

73 Re: FO Interview: Tony Dungy

Interesting but several things worth mentioning:
a)Philadelphia has lacked the consistent three receiver threat that Indy exploited.

and perhaps more important

b)Philadelphia plays a high risk/high reward blitzing style while Indy basically plays a prevent defense all game, so presumably they're better at defending against a fourth quarter comeback. Obviously this could be checked. Like, what was fourth quarter scoring for Indy and Philly opponents in games decided by 5 or less?

74 Re: FO Interview: Tony Dungy

This highlights the point I think Jeff Fogle is making - the potential importance of scheme and how some schemes, with the right people to execute it, are consistently coming out ahead in the NFL. It relates to an area of interest I've had regarding DVOA. Overall, DVOA seems to draw a pretty good picture of how well teams are playing or played over the course of a season. Seventy-eight percent (78%) of the time from 2000-2009 teams miss their DVOA projected wins by more than 1 game and only eight percent (8%) of the time do they miss it by more than two games. What is interesting is to try and figure out what, if anything, can account for this.

26/32 Have exceeded projections by at least 2 games at some point during this period during (overacheivers):

ARI, IND and STL are the only teams to do so at least 4 times, and only STL did it 5 (2001, 03, 04, 05, 06).

21/32 underperformed projections by at least 2 games at some point during this period during (underachievers):

BAL, KC, MIA, NYJ, PHI, SD and TB are the teams that accomplished this feat at least 4 times, and KC, MIA and SD achieved it 5 times.

These tend to be the teams that people rally around to claim that DVOA is broken, despite it's consistent correlations with wins. When it comes to over achieving projected wins, Jeff Fogle may or may not be on to what is going on with IND, but it this group of over and under achievers that is most interesting, not from the standpoint of whether DVOA is or is not broken, but what are these teams doing differently and/or better/worse than the other teams in the league.

I suspect that random factors play a role for many teams, especially at the +/- 1-2 game level, but over/under achieving by 3 or more games suggests the possibility of some other factor at play (perhaps drive efficiency?).

77 Re: FO Interview: Tony Dungy

Great stuff Delta.

I would add in the evolution of "possession management" in addition to scheme/execution/right guy in charge. Meaning part of the mix is squeezing out that extra possession per game whenever possible. I get the sense this is part of the issue just from hearing various comments in interviews...the comments from Dungy above...and even hearing Mack Brown down here in Austin constantly talking about trying to have the last possession of the first half (he's big on trying to have the last possession of the first half, and the first possession of the second half). I think those things are all tied together...meaning the smart QB/coach that can drive the field also has an eye on limiting any comeback time for the opponents.

From my standpoint, I wouldn't say DVOA is "broken." More of a "behind the market" theme to me. I mean, it's not like they have the Lions in the top five and the Saints in the bottom five or something. I don't think, though, that this past season's week-to-week rankings compared well at all to other places doing rankings.

Putting it all together, I guess I'm seeing it this way:

*DVOA data worked in describing reality in the past.
*Reality is changing
*DVOA isn't changing

Imagine a scenario where Bill Polian is kind of an "agent zero" who puts the mix of "possession management" and the drive stuff we've talked about together in his Buffalo years. There's some sort of baseball saying that "When Branch Rickey caught a cold, the rest of the National League sneezed." Meaning, Rickey would innovate and everyone else would eventually try to copy it. Let's say gradually, over the next 10-15 years, those Polian concepts starting being adopted in other places.

DVOA was developed using numbers when relatively few teams were using those concepts. But, now, those concepts are wider spread, so the difference between the have's and have not's is getting bigger (smart approach/right schematic/veteran QB who can move the ball has a BIG edge over generic approach/less than ideal schematic/young or mistake-prone QB)This "BIG" theoretical edge doesn't show up any more than usual in volume numbers during the first 26 minutes of a half, but shows up dramatically in the last 4 minutes of a half in a way that helps determine who wins or loses close games.

Imagine further that the tipping point for this is around 2005. I've heard that year often in FO discussions/articles/etc... Aaron or Bill or somebody will say, "when we measured this is 2005, it showed..." Could it be that DVOA was capturing football reality extremely well pre-tipping point, but is now watching the game evolve away from it? I've made that point before. The betting markets aren't using DVOA, and are staying in touch with reality based on all the studies I've seen. DVOA is not in my view.

It's kind of like watching an explosion in slow mo...or maybe computer simulations of the big bang. DVOA measured reality in "the third second" and said "this is reality!" But, in second four, and second five, things were drifting away from "the third second."'s a theoretical timeline (I'll assume it will turn out to be false, or extremely over-simplified...but it expresses at least a possible explanation for various things that have been discussed this year).

*Polian, or somebody he learned from, is agent zero.

*The rest of the league very gradually adopts what worked for Polian. This goes at a slow pace through the 90's and early 00's because Polian isn't a well-publicized guy and he's not publicizing what he's doing. Eventually, through game films, personnel moves, etc...possession management and drive importance take hold.

*DVOA "defines" football in its mind during a time when these keys hadn't reached a tipping point.

*The tipping point is reached around 2005 or soon after.

*NFL results continue to drift away from DVOA thanks to this kind of mini-strategic "big bang."

*By 2009, preseason expectations based on the "reality" of football as defined by DVOA are so far off that they go 5-16 against Las Vegas "regular season win" estimates.

*In 2009, FO spends much of the year saying variations of "Indianapolis and New Orleans aren't as good as their records" while two teams who are great at driving the field and have confidence that they can score on the final possession if needed are working their way to a Super Bowl showdown.

*At the end of the 2009 season, FO has a 9-7 team with a second-year quarterback ranked as the best team in the NFL...even though they went 0-2 vs. their division winner, had a poor record vs. other good teams, and lost at home to the Colts without scoring a TD. In the playoffs, that 9-7 team would lose to the Colts again, completing a second game where they couldn't even score a TD.

I think this is at least a feasible "spine" for the body of what's happening.

The danger of using data for all 32 teams during the 2000's is that we're looking at various stages of the pre-tipping point years of the big more and more teams are adopting possession management strategies. There are also personnel changes on the sidelines and at QB that could influence what's happening. Personally, I think this is a skill (or an inability) that follows coaches around, and improves with the experience of the QB. I'd be more interested in tracking that +1/-1 and +2/-2 information of yours by coach and QB rather than franchise...but that would be a bear to figure out.

Anyway...wanted to get that down on paper before this article got pushed off the front page. Thanks Delta for sharing your analysis and data with everyone. It's great to have somebody to talk through these things with. Thanks to others who have commented as well.

78 Re: FO Interview: Tony Dungy

Some somewhat Random and disjointed thoughts;

- For me the most easily recalled "revolution" is the 1970s Dallas Cowboys. IIRC, the Cowboys under Tom Landry and Tex Schram(m?) were seen as innovators because of their use of computers to analyze their opponents' tendencies.

- I don't know if Polian has latched on to something that others will mimic and copy and make successful, or if he has found one of several ways to increase the likelihood of success. It is interesting to me that in the last decade, that 3 teams comprise 8/10 of the last AFC Super Bowl representatives.

*DVOA data worked in describing reality in the past.
*Reality is changing
*DVOA isn't changing

I haven't seen evidence, but I'll have to check, that DVOA is working less well. Consistently, even on a week-to-week basis, the correlation between DVOA and Wins has remained fairly consistent.

All of the above why I think at present the outliers are interesting, especially teams that are successful and are outliers. As I've done some of these preliminary analyses, IND is very interesting b/c of there "over achieving," and SD is interesting b/c of their "under-achieving." I'm also becoming more interested in MIN, b/c they have been the best at "over-achieving."

79 Re: FO Interview: Tony Dungy

Sorry for the double post, but after posting above I couldn't help myself and did some calculations.

Correlation of DVOA with wins 1994 to present: r = 0.854
Correlation of DVOA with wins 2000-2009: r = 0.853

Correlation of DVOA with Wins by Season (2000-2009):

2000 0.876
2001 0.871
2002 0.834*
2003 0.837*
2004 0.866
2005 0.861
2006 0.841*
2007 0.900**
2008 0.881
2009 0.847*

Average correlation 2000-2009: 0.861
Standard Deviation: 0.021

* More than one SD below the decade mean (note none more than 2 SD below the mean)
** More than one SD above the decade mean (note none more than 2 SD above the mean)

Perhaps I could/should analyze further and do a year-to-year analyses, but I think the above data points to this as an exercise in futility. I don't see evidence of a tipping point, as much as I see evidence that some teams outperform their DVOA and others underperform and a lot tend to hit the mark. Almost all teams at some time in the last decade managed to do both; however, only a few have done it consistently.

80 Re: FO Interview: Tony Dungy

Thanks Delta! Figured I should do a quick study to even out the workload after all your contributions. Part of my thesis is that experience trumps inexperience in close games more than DVOA gives it credit for (ie, Peyton Manning can drive the field and win the game for you but Joe Flacco hasn't figured out how to do that yet).

Went through this morning and grabbed:
*All teams the L3 years who had missed their "estimated wins" by 1.1 or more
*The names of the QB who started the most games for them that season
*The QB's in season "age" as listed by pro-football reference

If you go to the standings of any year at pfr, you can click on that team to get the players stats...and each guy's age at the time is listed. So, that's the number I'm using.

Exceeding estimated wins by 1.1 or more (from low to high in age): Cutler 26, Rivers 28, Brees 30, Palmer 30, P. Manning 33, Hasselbeck 34, Favre 40 (median 30, with only two guys under 30)

Failing to reach by 1.1 or more: Freeman 21, Flacco 24, Henne 24, Roethlisberger 27, Orton 27, Campbell 28, Bulger 32, Brady 32 (median 27)

Exceeding by 1.1 or more: Ryan 23, Cassell 26, Schaub 27, Romo 28, P. Manning 32, Pennington 32, Delhomme 33, Collins 36, Frerotte 37, Warner 37 (median 32)

Failing: Thigpen 24, Rodgers 25, Orlovsky 25, Anderson 25, Fitzpatrick 26, Rivers 27, Wallace 28, Garrard 30, McNabb 32 (median 26)

Exceeding: Rivers 26, E. Manning 26, Romo 27, Brady 30, Kitna 35, Warner 36, Favre 38 (median 30)

Failing: Boller 26, Palmer 28, Lemon 28, Harrington 29, McNabb 31, Huard 34, (note here that the Jets were also a failure, and split time almost evenly between Pennington 31 and Clemens 24. Those straddle the midpoints of 28-29, so won't influence the median. Also, Huard only had 14 career starts as a 34-year old, so he's kind of a false read for that age. Personally, I'd prefer 28 as the median here because Huard had the in-game inexperience of a younger player).

Overachievers: 30-32-30 as medians, with some prominent quality on the list
Underachievers: 27-26-28 as medians, with the woes of inexperience prominent

Again, we're not just looking at good vs.'s doing MORE than DVOA would have suggested with estimated wins, or LESS than DVOA would have suggested.

To me this is at least consistent with the theory that there is a "knowing how to perform" skill in close games that comes with experience, and generally isn't there with inexperience (Ryan is an exception for Atlanta in 2008 on the good side, McNabb shows up twice on the bad side despite his experience).

I'm guessing using "career starts" rather than age would show something similar. That's much more of a bear to figure out for a quick and dirty project like this.

Not sure how this might or might not impact your "win correlation" data Delta. I do think the elements of QB experience, ability to drive the field, ability to drive the field with the game on the line, and the ability to make the most of possession management strategies are at the heart of the disconnect between DVOA and reality, to the degree there's a disconnect. Maybe that's at the heart of what prevents near perfect correlation.

Can we at least agree that a 9-7 team with a 24-year old QB is very unlikely to be the best team in the NFL? Though, VERY likely to improve as he matures?

81 Re: FO Interview: Tony Dungy

Can we at least agree that a 9-7 team with a 24-year old QB is very unlikely to be the best team in the NFL? Though, VERY likely to improve as he matures?

I would agree to this, but you can't judge DVOA on how it grades a single team.

82 Re: FO Interview: Tony Dungy

We're getting to a point where some serious hypothesis testing and formulation is warranted.

I took your data and ran a quick T-test comparing the mean ages of the overachievers v. underachievers. There is a statistically significant difference between the means of the two groups.

This finding raises an interesting question for me. Previously, I've been arbitrarily underwhelmed by the +/- 1 group and have only been interested in the +/- >1.5 (i.e two or more games). The fact that on this variable (QB age), segregated narrowly (+/- 1.1 wins) with a small sample size there was a statistically significant difference leads me to believe that a clearer definition of under/over achieving is required. To date, I've referred to teams as over/under achievers if they exceeded or failed to achieve their predicted wins. The question from the above analysis is where to properly set the bar for over/underachieving. Furthermore, defining terminology in a meaningful way is important for clarity.

Next, I think it's useful to identify other potential influences that may account for under/over achievement. Obviously, one possible variable is QB experience. I think another one could be penalty discipline. We're all mystified by the 2009 Ravens, but throughout the season there was consistent talk about penalties being a problem. Another factor that may be of importance is Luck (randomness). Luck has been my justification for only considering +/- 1.5 teams. I've assumed that most teams probably have won or lost at least 1 or more games during the year due to luck. Speaking of luck, I began compiling a list of factors that have at least some degree of randomnes involved - fumble recovery, penalties, injuries, interceptions, etc.

Another area that has to be considered involves the context of over/under achieving. I think that the predicted 3-13 team that goes 5-11 may be very different than the predicted 11-5 team that goes 13-3. Additionally, in the QB data I think a consideration has to be made for the quality of team around the QBs. Quick glance suggests that more of the failures (underachievers) were teams that were not that good and many of the exceeders (overachievers) were teams that were good.

Anyhow, there's a lot to get my head around and organized. I think that hypothesis testing is going to require more sophisticated analyses than what I've been conducting.

84 Re: FO Interview: Tony Dungy

Perhaps, but I would expect that instead of rankings that go beyond DVOA, we would have statistically validated explanations of what the rankings reveal.

For example:

1. BAL DVOA 30.5% Rec 9-7 EST Wins 11.8 - Baltimore underachieved this year by almost a full 2 games, costing them a division title. This failure seems to be primarily due to poor penalty discipline and effects of having a young QB.
2. GB DVOA 30.4% REC 11-5 EST WINS 11.0 -
3. NE DVOA 29.1% REC 10-6 EST WINS 11.3 - NE underachieved by one win this year. As with most one win under/over achievers, luck appears to be the primary factor behind this. Despite having an older/wiser QB, they were not able to overcome poor fumble recovery luck and injury luck.
4. PHI 28.4% 11-5 10.9 - Despite all the bitching and moaning about PHI and DVOA, at seasons end, DVOA predicted EST Wins and actual Wins lined up almost perfectly.
5. DAL 25.5% 11-5 11.2 -
6. NO 23.5% 13-3 11.6 - NO overachieved by more than one game this season. It appears that this was due to a combination of QB Wisdom and the accompanying efficiency and some good penalty luck. This team is the other side of the BAL coin
7. MIN 18.4% 12-4 10.2 - MIN overachieved by 2 wins and this represents one of our most robust findings about wise and efficient QBs, additionally, the indoor dome effect comes into play in a rather robust manner as well.
8. IND 17.9% 14-2 11.0 - IND's overachievement this year is not surprisingly almost completely due to the wise/efficient QB effect. Remove Peyton Manning from the equation and the Colts (Clots) are at best 11-5 and that may be stretching it a little as they are a dome team.
9. etc.

85 Re: FO Interview: Tony Dungy


Do you have a statistically validated study on the "indoor dome effect" as you called it? Could see how that could come into play...meaning it's easier to drive the field for a late score in ideal conditions rather than outdoors in the wind, potentially in cold if the Colts play a lot of close games, and the Ravens play a lot of close games, the Colts have a better chance to pull more out of the fire with late possessions. Not sure if that was just an example of what the future might hold, or if you were referencing something you had shown with your numbers that I missed in prior comments. Almost feel like we have to say "and they were dealing with an "outdoor non-dome effect" for NE, Philly, etc...too.

Ideally there would be a way to channel what we learn into future DVOA methodology. You can't really adjust for "QB age" because it's apples and oranges...sticking an age into a percentage won't cut it. But, if we establish that there's a REAL factor here in terms of QB experience, or penalty tendencies, or dome environments, then we can at least try to find ways to emphasize certain elements of the game within DVOA (meaning you "double count" fourth quarter performance when teams are within 8 points of each other...something like that). There has to be an acceptance from FO that DVOA isn't all the way there yet, then a pursuit of logical ways to deal with the significance of additional influences as they're discovered. Or...there IS an ability beyond luck to win close games (or an inability that causes losses), and those MUST be included in any proper ranking of teams.

(Quick may be true that it's luck in some sports, but less-so in others. I think most agree that results in one-run games in baseball are mostly luck...but it's hard for hitters to "control their own destiny" in the same way an NFL QB can. AROD has to hit a 95 mph pitch with movement late in a close game. A veteran QB may have more control over driving the field percentage-wise than AROD does of putting that elusive pitch into play...and the seed that got planted from SABR-metrics about luck determining close games has influenced thinking about other sports a bit too much).

Not sure if the statistically significant data from the last three years with QB ages is something that goes back further. That's about the time they liberalized holding...which may be a trigger point that allowed experience to really thrive late in close games (and may be an influence in my thinking that DVOA is drifting away from a reality it was better reflecting in the 2005 references). I'll try to do some work with that when I get a chance (not imminent, lol). I think there is a general consensus that "football people" have always felt that QB experience mattered in close games, and that inexperience was a big negative. And, that might be part of why Bill Polian's most supportive comments to FO were "you're on the right track" rather than "you guys have nailed it."

In terms of "ideal" rankings...the betting markets are pretty close to having the degree they're possible. The "wisdom of the crowds" appears to be taking QB experience into account given how teams are rated and priced (veteran QB's get respect in the line, newcomers don't). At least over the last three years, it looks like the formula for estimated wins isn't fully incorporating what causes wins into the mix. Most of it's there, but not all.

87 Re: FO Interview: Tony Dungy

Well, age is pretty good. Not ideal because you occasionally get an older guy who's been a back-up all his life (like Huard in an earlier post), and some younger guys look like they're clicking in around 27-ish (Romo). Think games started is the very best way to do it...but it's time consuming as heck in terms of going back and doing earlier studies. I used age because pro-football-reference had each guy's age on each year's stat page.

Age gets you most of the way there. Would assume starts and pass attempts would do the job just as well or a little better. Might be concerned about pass attempts in that bad QB's may pile up attempts in comeback efforts that aren't really about winning close games...while good QB's may be handing off to sit on leads. Might create some illusions as the QB's are climbing the ladder experience-wise (a guy throwing 35-40 passes per game on a bad team looks like he's getting more experience than a guy throwing 25 passes per game on a good team, but it's not "learning how to win close games" experience). Once somebody's an established starter after X number of starts, or at age X on the calendar, they've established a level of quality that's likely to be relevant.

Age does the heavy lifting, and did show statistical significance according to DW. Would be interesting to see if game starts and career attempts did the same...

90 Re: FO Interview: Tony Dungy

I think it would all even out because good QBs will throw the ball more early.

I mean really when is the last time a bad QB averaged over 30 passes per game over more than 4 consecutive games?

94 Re: FO Interview: Tony Dungy

Tuluse, I'm hoping you're being sarcastic or something.

From this year:

Stafford had a QB rating of 61.0
Passes by start: 37-30-36-36-33-42-51-43-43

Freeman had a QB rating of 59.8.
Passes by start: 31-28-33-29-44-33-26-31-31

Quinn had a QB rating of 67.2, which is well below average
Passes by start: 35-31, then a long layoff...then 31-33-34-45 in his first four starts back.

The kid QB's aren't in there to learn how to hand off. They're in there to learn how to run the offense, which is largely a mix of long and short passing plays against aggressive defenses (with runs mixed in once there's movement).

Not all teams throw that often. Oakland looked fairly restrained with Jamarcus this year: 30-24-21-33-13-28-11-22-23.

I think you'll find that it's not uncommon for bad QB's to average 30 passes per game over more than 4 consecutive games...particularly the youngsters who are learning the ropes. That's how they learn the ropes. Depends on how much clock opponents are running with leads in the blowouts...and late season weather is an issue when bad QB's are playing for Buffalo, Cleveland, etc...

David Carr got 75 starts and about 2,000 passes before Houston gave up on him, which is close to 27 per game over 5 seasons of disappointing play.

91 Re: FO Interview: Tony Dungy

“Do you have a statistically validated study on the "indoor dome effect" as you called it?”

- No I don’t. In that example, I was pulling thoughts about what may/may not contribute to a team over/under achieving. As you can see, one thought I had was that dome teams might have an advantage. I hadn’t thought through whether it was due to increased efficiency associated with having perfect weather (interesting given recent NY/NJ Super Bowl decision) or some other factor like crowd noise.

“and they were dealing with an "outdoor non-dome effect" for NE, Philly, etc”

- Essentially we’re talking about weather here, and I think FO has tried to model this in the past and hasn’t had success. This could be due to the fact that it is not an effect, but also could be due to lack of statistical sophistication on the part of FO – not saying I could do any better, just wondering,

“Ideally there would be a way to channel what we learn into future DVOA methodology.”

- Maybe, I think you’re operating from an assumption that there are only a few ways to over/under achieve (e.g. drive efficiency). I suspect that are a larger number of ways that don’t always weigh out the same because they can have synergistic effects of cancelling effects. These effects may be able to be statistically modeled; however, it may require rather sophisticated techniques (beyond my ability). Alternatively, there may be idiosyncratic reasons for the effects to act the way they do. Consequently, I think the first step may be to try and identify some of these factors, such as QB age (I know this is flawed measure, but still think it was an interesting find).

Interaction effects and unsophisticated statistical analyses may be what is behind some of the failed attempts to improve DVOA over the years (e.g. “That includes adjustments based on climate rather than just indoors/outdoors, "blowout" adjustments for plays before the fourth quarter, and adjustments for home-field advantage. That last one really surprised me when it didn't actually improve correlation from year to year or from the first half of the season to the second half, but it didn't, so it doesn't go into the new DVOA.” - I should have recalled earlier that the indoor/outdoor phenomena has been tried along with specific weather effects, guess I’m not as clever as I thought).

Ultimately, improving DVOA may have as much to do with knowledge of statistical techniques as it does with identifying variables. In the past I’ve complained about not having enough access to the methodology behind DVOA’s development to assess whether such incorporation is feasible. Without this access, all we can do is guess. For example, I understand the success point concept, but I have no idea how those weights were established – empirically, rationally, or a combo of both? Perhaps DVOA can be strengthened simply by tweaking success points. In the end, I suspect that if Brian Burke keeps plugging away, his system may exceed the work at this site.

“and the seed that got planted from SABR-metrics about luck determining close games has influenced thinking about other sports a bit too much”

I think luck (randomness) is a large part of the phenomena we’re looking at; however, I think there are different flavors of luck. There is pure luck (i.e. fumble recovery), there is temporal luck (e.g. when something happened), there is magnitude luck (e.g. severity of injury), there is personnel (?) luck (i.e. who got injured), act of God Luck (i.e. weather), human factors luck (i.e knowledge, cold weather hardiness, etc). It is worth noting, that many of these factors can be mitigated to a degree and controlled by human actions.

“Age gets you most of the way there. Would assume starts and pass attempts would do the job just as well or a little better”

Absolutely concur; on the surface, what the data did suggest is that there may be some measurable variables to account/explain the over/under achievement phenomena.

Peyton already had 64 starts by his age 26 season

The idea that starts leads to experience, efficiency and “clutchness” (that ought to set off a lot of bells and alarms at FO Poster Central – danger, danger, unmeasurable phenomena being mentioned) seems reasonable and dovetails with the Lewis Forecast models.

92 Re: FO Interview: Tony Dungy

Stat-layman here, is it possible that home field advantage doesn't improve year-to-year correlation because teams play the same number of home games every year?

72 Re: FO Interview: Tony Dungy

How many times did he blink? Did he remind himself to smile at the end?

88 Re: FO Interview: Tony Dungy

Went through tonight (Wednesday) and did the rest of the 2000's. Far too many names to type up. In short:

*The phenomenon with "age" didn't kick in until 2006. That was the year of the "kid quarterback" so to speak, where all of the following showed up on the minus 1.1 or more list: Leinart 23, Gradkowski 23, Walter 24, Frye 25, E. Manning 25. The median age was 28 out of 11 qualifiers. It was 30 out of nine qualifiers on the good side, basically in line with years 2007-2009 reported in an earlier post.

*Prior to that, it was common to see 30-year old (plus) mediocrities in the spot. It's as if teams preferred to go 6-10 with a veteran rather than risk the embarrassment of 3-13 with a kid. In 2006 that switched and teams thought it best to throw young QB's in the deep end to let them learn how to swim. Earlier in the decade Peyton Manning and Brady had made the good list in their 20's, showing that it was possible. Teams took their shots with youngsters because going 6-10 with Jeff Blake or Jay Fiedler in their early 30's just wasn't worth the energy. Might as well see if you've got the next Peyton Manning on your roster (then pick up a high draft pick if you don't!).

*There may still be earlier correlations with games started. That's a bear to do for everyone. 2002 looked interesting because there were some young successes and some old failures. The median of the successes was only 28, with guys like P-Manning, McNabb, and Jake Plummer on the list. The median age of the six failures was 31-32. If you look at games started though, the "good" guys had 79 as a median, while the "bad" guys had 35-48. Peyton already had 64 starts by his age 26 season. Among the guys in their 30's who struggled were Jim Miller of the Bears (18 prior starts), Jay Fiedler of Miami (32 prior starts), and Trent Green of KC (35 prior starts). 2006 the flood of kid quarterbacks were given shots, and that's continuing to happen as a general rule. I think it was 2007 when defensive players started complaining about holding being allowed on the offensive line except for the obvious pulldowns. Don't know when the league acknowledged that had been done and it was generally accepted. Was either 07 or 08 I think. So, we have two potential evolutionary ticks in play here (ages skewing younger for borderline QB's plus more time for veterans to make decisions in the pocket) possibly contributing to the statistical significance with ages that DW measured.

Age in and of itself doesn't show up as a difference-maker from 2000-2005. Games started might given additional study. And, there still may be the "Branch Rickey Sneeze" factor involving drive ability possession management coming into the mix more dramatically of late as well (particularly after liberalized holding). Discussions for another time. Gotta go write about the Orlando Magic's three-pointer barrage...

95 Re: FO Interview: Tony Dungy

Agree about the influence of Big Ben, Jerry (no ice cream pun intended). Michael Vick was also making headlines in 2004 (Ben 22 that year, Vick 24). Think gradual experimentation with young guys who ended up doing well led to the eventual 2006 floodgate that saw guys like Gradkowski, Frye, Walter, Leinart get in and struggle.

Maybe the best description is that early experiments were with "cream of the crop" guys. Those went well, so more experimenting further down the ladder ended up leading to some guys being in way over their heads, at least in terms of making the most out of their DVOA components as they relate to expected wins.

Fun discussion. Agree with DW about the heavy theoretical work ahead. Looking forward to thinking through and working through the process...

97 Re: FO Interview: Tony Dungy

Hmm. Not only did Roethlisberger have an unimaginably successful rookie season, but a sixth-round draft choice was leading the Patriots to championships. Maybe some less astute front offices put those two ideas together and decided to take a shot with their less highly regarded young QBs.

One nitpick: Leinart was a first-round pick, while Frye and Walter were thirds and Gradkowski was a six. The Cardinals might reasonably have had higher expectations for Leinart than the other teams (should have) had for their guys.

Save 10%
& Support Aaron
Support Football Outsiders' independent media and . Use promo code SCHATZ to save 10% on any FO+ membership and give half the cost of your membership to tip Aaron.