We've introduced the greatness known as Football Outsiders a couple of times here at ROTB and we'll continue to incorporate their unique analysis of the NFL as time goes on, but for now the managing editor of FO, Bill Barnwell, has been kind enough to sit and answer some Arizona Cardinals related questions. First for those of you who don't know, Football Outsiders has published the Pro Football Prospectus in the past but this year they are calling it the Football Outsiders Almanac 2009 and a pdf version of the book can be purchased here.
For those who have never heard of FO at all, here is a quick description of their mission and other FAQ's:
Football Outsiders brings you a series of brand new, in-depth statistics you can't find anywhere else. With these stats, we will attempt to bring objective analysis to football that matches the revolution in baseball writing and analysis over the past 20 years. We have new methods for analyzing skill players, offensive and defensive lines, special teams, and total team efficiency.
We tried to focus on questions that not only effected the Cardinals on-the-field play but also some of the upcoming decisions that they'll have to make while also introducing as many of their different metrics as possible. Hopefully the result will give us all a good introduction to a site that is revolutionizing the way that some statistics are viewed. And for those of you who have been waiting to scream, "we get no stinkin' respect" from the hills, I think you'll have some new ammunition now......
ROTB - The Cardinals will face a major decision next off season when Karlos Dansby comes up for free agency. He certainly wants to be paid like a top tier linebacker in this league but how does he measure up with other elite middle linebackers according to FO's metrics?
Bill Barnwell - Football Outsiders keeps a stat called Defeats that basically totals turnovers, negative-yardage plays, and third- and fourth-down stops. In each of the last two seasons, Karlos Dansby has led the Cardinals in Defeats, and ranked in the top 20 of all players in the league. Most of the league leaders are sack masters, the DeMarcus Wares and Jared Allens of the world. Dansby, on the other hand, chips in with a little bit of everything. Last year, for example, he only collected four sacks, but added nine stuffs, two forced fumbles, and two interceptions, finishing with 26 Defeats. So we know he makes big plays.
Meanwhile, we also keep track of total Plays -- tackles, forced fumbles, recovered fumbles, passes defensed, interceptions. We then calculate percentage of team Plays for each defender. Dansby has been ridiculously consistent in this regard, ranking 17th among linebackers in 2007 and 16th last year. So we know he spends a lot of time cleaning up after his teammates.
We have other stats to measure Dansby's ability to play the run. We have a stat called Stop Rate that measures what share of a player's run tackles held opponents to short gains. We also have the average yards gained on each player's run tackles. Dansby's numbers here are up and down. He ranked 54th in Stop Rate in 2007, but 20th in average yards allowed on run tackles. Last year, though, those rankings virtually flipped; Dansby ranked 30th in Stop Rate, but 59th in average yards. It's a little hard to draw conclusions with numbers this erratic. However, given Dansby's high number of defeats and total plays, it seems like he either makes plays in the backfield, or after 7-yard gains, but not often anywhere in between.
Finally, we track how often each defender was targeted in pass coverage, Success Rate, which measures how often he held receivers to incompletions or short gains, and Adjusted Yards Per Pass allowed. "Adjusted" in this case means we take the quality of the receiver into account -- it's harder to cover Andre Johnson than it is to cover Andre Davis. There's no confusion here: Dansby has ranged from bad to awful in pass coverage, ranking 43rd or worse in both metrics in each of the last two seasons. Last year, he was 75th in Success Rate, and 86th in average yards allowed.
So we're left with a player who's something of a boom-or-bust defender, a guy who will chip in plenty of big plays that end drives immediately, but will also surrender first downs that keep a drive alive, especially in pass coverage. He's clearly the Cardinals' top linebacker, and he'd be the top linebacker on about half the teams in the league. At the same time, about half the teams in the league have at least one linebacker this good. Whether that makes him an elite player or not is, I guess, an argument of semantics.
ROTB - (after hearing Bill's answer we had to do a quick follow-up concerning Gerald Hayes) I've got one follow-up question concerning Dansby vs. Gerald Hayes. Hayes has long been an after thought amongst the Cardinals linebackers but after reading your description of the various defensive metrics I went back and compared the two. While it looks like Hayes doesn't make as many big plays (defeats), his stop rate is identical, his yards per play is less against the run and pass and his success rate is better against the pass. I guess my question would be two-fold, are Hayes' stats similar in past seasons and is it reasonable to assume that his stats might look better because he's on the field less, especially in passing situations (assuming that the average pass play goes for more yards than the average run play)?
BB - Here's how Dansby and Hayes have fared in our basic defensive stats over the past three seasons. I added in overall stop rate, which measures the share of each player's plays that held the offense to short gains, run or pass:
Dansby Team Pct Defeats Stop% Run Stop Rush Yds Success% Pass Yds 2006 11.7% 23 62% 73 2.9 40% 10.4 2007 15.2% 27 53% 61 2.7 43% 6.1 2008 15.8 26 59% 69 3.6 40% 8.6
Hayes Team Pct Defeats Stop% Run Stop Rush Yds Success% Pass Yds 2006 13.8% 15 53% 67% 3.6 45% 6.4 2007 12.4% 20 56% 64% 3.2 47% 7.5 2008 12.4% 15 59% 65% 3.2 47% 6.8 Note that in 2006 Dansby played a more traditional 3-4 outside linebacker role, playing close to the line. He had 8 sacks that year, and you can see the difference in his run defense numbers (and a corresponding drop in his Team Percentage -- he made fewer plays back then, but when he did, they were big ones). Note also that Success Rate is a measurement of how often the defender succeeds, so higher is better. Dansby has been awful in pass coverage for three seasons running now. Hayes hasn't been all that great himself.
Your second question is trickier to answer. Defensive stats are kind of nebulous, and are open to interpretation. Suppose the Seahawks, for example, run right at Hayes and he gets blocked out of the play, then Dansby comes over to make the tackle, but the runner has already gained seven yards. In a play like that, Hayes is the one who failed, but his numbers won't change (aside from a tiny dip in Team Percentage). Dansby, though, will see his Team Percentage rise, but his Stop Rate will go down and Average Run Yards will go up.
But going with the data we have, we can draw three conclusions:
- Dansby makes more big plays than Hayes.
- Dansby also makes more little plays than Hayes.
- On the whole, Dansby's average play is about as effective as Hayes' average play.
Based on that, it looks like Dansby is the superior player, but the gap may not be as wide as most Cardinals fans believe.
ROTB - Cardinals fans are ecstatic about the potential at cornerback with Dominique Rodgers-Cromartie and Bryant McFadden. Some of us thought that opposing teams tried to stay away from DRC down the stretch last season, proving that he's got 'shutdown potential' in his future. Do the stats back up that theory and how did DRC and McFadden rank amongst their peers last season?
BB - If that trend existed, it isn't reflected in our charting numbers. Let's divide the season into three parts: Weeks 1 to 3, when Rodgers-Cromartie was not starting; Weeks 4 to 11, Rodgers-Cromartie's first start up to his two-interception game against Seattle; and Weeks 12 to 17, all games after that Seattle game. Here's how often the Cardinals' top three corners -- Rodgers-Cromartie, Roderick Hood, and Eric Green -- were targeted in those timeframes:
Targets Week Hood Green DRC 1-3 16 11 7 4-11 41 19 31 12-17 31 9 36
At the end of the season, Hood and Green saw their targets go down, while Rodgers-Cromartie's targets went up. So if anything, teams were targeting DRC more often by the end of the season. Of course, that's only half the issue -- teams may have been picking on DRC, but were they successful? Like we did for Dansby, we can measure Arizona's top corners in Success Rate and Adjusted Yards Per Pass:
Success Rate Hood Green DRC 1-3 35% 55% 74% 4-11 56% 48% 52% 12-17 40% 46% 50%
Adj Yds Per Pass Hood Green DRC 1-3 8.6 5.3 4.9 4-11 6.1 13.4 5.5 12-17 10.5 7.4 7.7
In plain English, these numbers show that teams started throwing at DRC more often by the end of the year; that throws to DRC were not more likely to be completed; and that when they were completed, they were more likely to be big gains.
Breaking it down game by game, though, reveals some interesting details. The Giants threw 33 passes against Arizona, and 12 of them -- more than one-third -- were at DRC, usually when he was covering Domenik Hixon. But he allowed only seven completions for just 62 yards. New England threw at him seven times, usually to Randy Moss, and he allowed three completions for 100 yards, including a 76-yarder. In the other four games in that stretch, DRC was targeted only 4.3 times per game, allowing 2.5 completions for just 26 yards per game. Against Minnesota, DRC was targeted only once, an incompletion to Bernard Berrian. (He may have been the only Arizona defender to play well that day.) There were definitely some games when his receiver was a non-factor.
DRC's biggest problem may be his tackling issues. Even though he only started for a portion of the year, he allowed 200 yards after catch, the 17th highest figure in the league.
Over the course of the entire season, Rodgers-Cromartie ranked 34th among all cornerbacks in Success Rate, and 33rd in Adjusted Yard Per Pass. Considering there are 64 starting corners in the league, he was almost exactly average. That's a pretty good performance for a rookie, and he's likely to improve.
McFadden, meanwhile, marks a huge upgrade over either Green or Hood. He ranked 39th in Success Rate last season, but 26th in Adjusted Yards Per Pass. That's a similar overall performance to DRC, but slightly different in style; McFadden is more likely to allow completions, but less likely to allow big plays. He's a solid tackler, and that helps against the run as well. McFadden posted an 80 percent Stop Rate on rushing plays last year, tops among all cornerbacks in the league.
ROTB - FO attempts to differeniate the different positions on the offensive line and evaluate each position versus the run and pass. Given that, which positions need are better at run blocking vs. pass blocking and visa versa? If FO were to determine a 'weak link' in the Cardinals offensive line, who would it be?
BB - In a perfect world, you'd obviously want five guys on the line who each excelled at run and pass blocking. Realistically, you're usually looking for pass blocking from your tackles and run blocking from your center and guards. With that in mind, the Cardinals' offensive line appears to have been built inside out. We use a statistic called Adjusted Line Yards to measure run blocking. In a nutshell, this stat cuts down on the reward for long runs, but at the same time increases the penalties for runs that lose yards. The Cardinals were 31st in the league in ALY on runs up the middle, and about average on runs to either side. On the other hand, Reggie Wells, Lyle Sendlein, and Deuce Latui combined for just 5.5 blown blocks leading to sacks and hurries. That's a pretty good performance for interior linemen on a team that passes as often as Arizona does. Meanwhile, only ten left tackles had more blown blocks than Mike Gandy's 5.5, and Levi Brown's 8 blown blocks ranked second among right tackles. If I had to name one as a weak link, it would be Brown, because the Cardinals are a pass-first team. If they were a run-first team, I'd pick one of the interior guys. Regardless, plenty of room for improvement here across the board.
ROTB - If you had to pick one over rated player and one under rated player on this team, who would they be and why?
BB - Well, based on the first part of this essay, I'm tempted to name Rodgers-Cromartie as overrated. It's not like his playoff performance was off the charts. He held Roddy White to a bunch of short gains, picking off one pass. He played very well against Carolina, though his performance was overshadowed by Jake Delhomme's meltdown. But then he gave up a long go-ahead touchdown to DeSean Jackson in the NFC title game, and was covering Santonio Holmes for most of the day when Holmes was named Super Bowl MVP.
For underrated, I'll go with another member of the secondary, Adrian Wilson, who has been one of our favorite players for years. He was fifth among defensive backs last season in Stop Rate against runs, third in 2008, second in 2007. He's not as dominant in pass coverage, but he's still very effective. Consensus opinion is that Ed Reed and Troy Polamalu are the best safeties in the league, and I agree, but I'd put Wilson at the next level, and he badly outplayed Polamalu in the Super Bowl. He's an absolute stud, one of the best players in the league, but since the Cardinals have been so mediocre until last January, he wasn't the household name he deserves to be.
ROTB - I saw that according to FO's preseason statistics/rankings that the Cardinals were more likely to land a top three pick in next year's draft than a playoff spot, what exactly goes into those projections?
BB - I knew this was coming. Yes, the projection system thinks that Arizona is going to be very, very bad this year. When the first set of projections came out, we said that couldn't be right, and we kept going back to the drawing board, making tweaks to the formula, looking for ways the system may be misreading Arizona. But no matter what we did the Cardinals came out near the bottom.
To understand why, you have to think of Arizona not as the NFC Champions, and not as a team that won two playoff games at home and a third because the opposing quarterback collapsed. Think of them as a team that went 6-0 against the weaklings in the NFC West and 3-7 outside the division. Think of a team that scored one more point than they allowed over the regular season. Think of a team that was blown out by the Jets, Patriots, Vikings, and Eagles. Now that seems like a team ready to collapse.
Now consider that the key players to the Cardinals' offense all made it through the entire season intact. Arizona started the same five linemen in all 20 regular season and playoff games last year. That's ridiculous. For comparison, every one of Seattle's top five linemen were on injured reserve by season's end. It's hard to believe Arizona won't suffer a significant injury to a lineman this year, and if they do, it could be catastrophic -- there's nothing on the bench but a bunch of Day Two draftees lucky to have a job.
The other two most important cogs in the Cardinals machine are Kurt Warner and Larry Fitzgerald, and they also played in every game (as opposing cornerbacks are only too aware). If Warner goes down -- and he has never gone two seasons in a row without missing a game -- will Leinart be able to keep things on track? If Fitz misses time, is there anyone who can fill his shoes? (And I don't limit that question to Arizona's roster -- if Fitzgerald's not playing, I don't know if there's a human being on the earth who could take over.)
Speaking of Warner, he is now 38, and it's very doubtful he'll be able to match last year's 4,583-yard, 30-touchdown performance. Here is a complete list of all quarterbacks age 38 or older who have thrown for more than 3,600 yards or 25 touchdowns in a season: Warren Moon and Brett Favre. That's it. Joe Montana couldn't do it, Dan Marino couldn't do it, John Elway couldn't do it. It's possible Warner will defy Father Time like Favre and Moon, but it's more likely that his numbers will decline by at least a thousand yards and a half-dozen touchdowns. This team can't afford to lose that kind of production.
The final obvious harbinger of doom is the upheaval among the coaching staff. Teams that lose coordinators tend to decline, and Arizona lost theirs on both sides of the ball.
For what it's worth, despite all their faults, I have more confidence in the Cardinals than the projection system does. I expect them to win seven or eight games and contend with Seattle for the division crown.
-----------------------------------------------------------
There are obviously quite a few talking points in there so I'll just open it up to the floor. Agree on Dansby's not-so-elite status? Is DRC the most over rated person on this roster? Top three pick, really?