Wednesday, October 19, 2011

Statistical Analysis: Turnovers

One of the big topics of discussion this week has been the Steelers inability to generate turnovers. Ed Bouchette wrote a decent article in the Post-Gazette about this issue today, and Tunch and Wolf talked about it on their radio show. Bouchette brings up some good points, including citing injuries to Hampton, Smith, and Harrison as a cause for lack of turnovers.
"Harrison has missed the past two games with an eye injury and will miss more. In the past, he was a turnover machine -- he forced 24 fumbles over the past four seasons that all wound up with him starting in the Pro Bowl. He has one this season, and that helped the Steelers beat Indianapolis when Troy Polamalu scooped it up and ran 16 yards for a touchdown in a 23-20 victory."
Bouchette also cites the stat that over the last 10 years, the 7 times the Steelers have had a positive turnover differential (creating more turnovers than they have committed), we made the playoffs every year. Conversely, the 3 years we had a negative turnover differential we did not make the playoffs. Currently, the team has a -10 turnover differential, due in large part to the 7 turnovers in the opening game against Baltimore. Since that game, the Steelers have committed 5 turnovers and created 2. We took a look at the data over the last 3 seasons to see if there was any difference between what the Steelers have done this year and in the past.


The first number that one needs to look at is interceptions. Troy Polamalu (despite missing half the season in 2009) has led the team in interceptions each of the last 3 seasons. In 2008 and 2010 he had 7 and had 3 in his injury-shortened 2009 season. In the last 3 years, the only other players to record more than 2 interceptions in a given season are Ryan Clark (3 in 2009) and Tyrone Carter (3 in 2008). The way the Steelers defense is set up, it is predicated on the safeties creating the interceptions, not the corners. When you look at other teams like Philadelphia or New England, their corners tend to play more of a short zone coverage that enables them to undercut sideline routs and come up with a lot of interceptions that way. The Steelers tend to give more of a "3 deep" look where the outside corners drop into a deeper zone, giving up the 9-yard out-route, or man coverage with only 1 or no safeties over the top, which forces the corners to give a little more cushion.

So the question should be asked - why haven't the safeties created any interceptions this year?

The easy answer would be, "They're not catching the ball." If I was writing for the Post-Gazette or the Trib, I would probably use this excuse and cite a number of examples where Troy and Ryan have come close to having picks this year, including an easy pick-6 against Seattle that hit Troy right in the hands. However, I'm a blogger and I hold myself to a little higher standard than simply citing plays that weren't made - I want to know if there's anything fundamentally different about our defense this year. 

The first stat that jumped out to me is the number of pass attempts by opponents. Over the past 3 years, the defense has been on the field for an average of 972 plays per season. On average, 558 of those have been passing plays (which is 57.4% of the total plays). This season, only 194 of the 357 plays the defense has played have been passing plays, 54.3% of the total. Since teams have had more success running the ball against us this year than in the past, they have thrown the ball less, which means there has been less opportunities for interceptions.

Another aspect to take into account is the types of plays opponents are running. There are two stats we can look at for this: Yards Per Attempt and Yards Per Completion. The following chart shows how these two statistics have changed for the Steelers opponents over the last 3 years and how this season stacks up.

As you can see, both Yards Per Attempt and Yards Per Completion are the lowest they have been since 2008. From this, it is safe to conclude that the Steelers opponents have been throwing shorter passes and not taking as many shots downfield. Simple football logic says that the shorter the throw, the more accurate a quarterback will be. Since opponents are throwing shorter passes this year and taking less chances downfield, this could be another contributing factor in why there have been less interceptions this year.

But is this enough? Can we simply place the blame on there being less pass attempts and those attempts being shorter passes? In short, the answer is no. Over the past 3 seasons, the Steelers have 53 interceptions on 1,674 pass attempts (3.17% of attempts). This season however, we have only 1 interception in 194 attempts (0.52% of attempts). If we had the same interception rate this year as we had the past 3 years, we would have 6 interceptions at this point in the year. That is a significant difference from the 1 we currently have.

Considering that Troy and Ryan account for most of the team's interceptions, one has to ask - has something changed about how they are playing? In my opinion, yes. Given that teams have had success running the ball against us, we have seen Troy and Ryan lining up in the box a lot more this season than in previous seasons. In particular, Troy has taken to lining up on the line of scrimmage a lot more (especially with Harrison out) on first and second down and playing backside pursuit against the run. This has afforded him the opportunities to make some big plays in run defense, such as his 3rd and 1 stop of MJD this past week. However, when Troy is playing on the line or down in the box, this limits his ability to make plays against the pass. Over the past 2 weeks, the Steelers have faced two very good running backs in Chris Johnson and Maurice Jones-Drew, both of which have been heavily utilized in the passing game. To combat this, the Steelers had Troy spying the running backs, even in passing situations, to cut down on screen passes and swing passes. By playing against running backs rather than wide receivers, Troy has had less of an opportunity to make interceptions than in the past.

Does this mean that Troy hasn't had his chances to make plays? Absolutely not. As I said at the beginning, Troy has had a few balls bounce off his hands or been in his vicinity that he just hasn't caught. I would say the combination of missed chances, facing less pass attempts, the schematic change of playing in the box more and opponents attempting shorter passes have been a big factor in Troy not making any interceptions yet.


The first aspect of the fumble stat that needs to be addressed is the difference between forced fumbles and fumbles lost. A forced fumble is recorded any time a ball is knocked free. A fumble lost is when the other team recovers the ball. Last season, there were 730 fumbles in the NFL, 349 of which were lost (47.8%). Over the last three years, the Steelers have fared much better than this, recovering 9 of 12 in 2008 (75%), 10 of 14 in 2009 (71%), and 14 of 28 in 2010 (50%). This season, they have only recovered 1 of the 4 fumbles they forced (25%). From 2008-2010 the Steelers forced 54 fumbles, 33 of which we recovered (61%). This is an average of 1.125 forced fumbles per game. If the Steelers had continued at this pace this year, they would have 7 forced fumbles right now as opposed to the 4 they have forced.

Over the past 3 seasons, James Harrison has led the team in forced fumbles every year (7 in 2008, 5 in 2009, 6 in 2010). Only 2 other players have forced more than two fumbles in a season (Timmons forced 4 in 2009 and Woodley forced 3 in 2010). Between the three (Harrison, Woodley, Timmons), they have forced 31 of the team's 54 fumbles (57%). I was surprised to find that of all the splash plays Troy has made, his only forced fumble in the last 3 years was the strip-sack of Joe Flacco last year.

Now, considering that Harrison, Woodley, and Timmons have been our best fumble generators, let's take a look at why these three guys aren't producing as they have in the past. For Harrison, the answer is obvious - he has been injured. In the 3.5 games he played, he forced 1 fumble which was nothing short of the biggest turnover of the season. As for LaMarr Woodley, he has forced the least amount of fumbles of the three linebackers (6). 

This can be explained from a schematic sense. James Harrison plays right outside linebacker, which means that for right-handed quarterbacks, he is rushing the passer from the blind side. Most of the time, quarterbacks don't see him coming, making it easier for him to tomahawk the ball out of their hand. 

Conversely, Woodley as the left outside linebacker rushes from the front side (of right-handed quarterbacks). This means that most quarterbacks will see Woodley coming and can either step away from the pressure, throw the ball away, or at least secure the ball before the sack.

The third player in the equation is Lawrence Timmons, who has forced 7 fumbles in the last 3 years. With Harrison's injury, Timmons has been moved from his natural position at inside linebacker to playing Harrison's spot on at right outside linebacker. In two games, Timmons has been less than impressive. For the guy who led the Steelers in tackles last year, he missed several last week against the Jaguars and was shut out on the stat sheet, finishing with 0 tackles. Timmons hasn't gotten pressure from the backside and hasn't really been a factor in creating turnovers. 

At this point you might be wondering about the fourth linebacker in the unit - James Farrior. As surprising as it might be for how productive the other 3 linebackers have been, Farrior has only produced 3 forced fumbles and 1 interception over the last 3 seasons. Larry Foote, who was inserted into the starting lineup when Timmons moved to the outside, has only 2 forced fumbles since 2008, including his stint in Detroit in 2009. 

Finally, since we began by discussing turnover differential, let's take a look at how the Steelers offense has done with turning the ball over. In the past 3 years the Steelers have fumbled 50 times, with 30 of those being recovered by opposing teams. This 60% rate is a bit higher than the league average, but is still below this season, where opponents have recovered 6 of 7 fumbles (86%) by the offense. In fact, the only fumble the offense did recover was when Trai Essex fell on a Roethlisberger fumble against Jacksonville. If we had recovered offensive fumbles at the same 60% rate we had over the past 3 years, we would have only lost 4 fumbles as opposed to 6. 

In conclusion, there are a number of factors that have created the current situation for the Steelers defense and their lack of turnovers. To begin with, the injury to James Harrison has not only taken our best fumble-producer out of the lineup, but has also shifted Timmons to the outside where he is nowhere near the same player he is on the inside and inserted Larry Foote into the lineup. Foote is a good tackler and solid against the run, but he, like Farrior, is not a turnover producer. Without Harrison in the lineup, we haven't gotten backside pressure on the quarterback, which is a key facet of producing turnovers. On top of that, the offenses we have faced this year have utilized a shorter passing attack, which has limited the number of deep balls and thus the potential for interceptions. Due to teams having more success running the ball against us, Troy has moved down into the box and keyed more on running backs rather than sitting in zones and undercutting passes. All of these elements have contributed to the Steelers current situation where they have only generated two turnovers in six games.

1 comment:

Webothhatetheravens said...

I really enjoy your blog! Good stuff!