As we look forward to what the Raiders defense has to offer, I decided to look and analyze the defenses around the League. In particular, I wanted to see what the best defenses, elite defenses, were doing and how they were performing in a numerical and statistical fashion. This would help to put some of the Raiders' numericals in perspective and give context to some of the values that i was generating.
It also forms a basis for what we some of the hallmarks of a successful defense. The Steelers' defense was fantastic in 2011 and so was a natural choice (plus some buddies of mine are steelers' fans so they would enjoy an analysis of their defense).
So if you are stil interested, here it is (if not, forward this to your neighborhood Steelers' fan)The full blog entry with full text and graphs and charts is here : http://www.raiders1.ninjagoro.com/?p=1306
The opening game of the 2011 season had the Steelers visiting their modern day arch-rivals, the Baltimore Ravens. The game ended up as a 35-7 thrashing that created panic and caused an uproar in the Black and Gold. 15 games later, the Steelers finished the season 12-4--losing the tie-breaker to the Ravens in part due to that initial loss--with a defense that was dominant. Truly Dominant, a defense whose worst day (excluding the opener) was better than many NFL teams' good day.
Here is what I refer to as the "Defensive Signature," the Points per drive graph. This is the mathematical model as explained above and the shape and location of the graph describes the defensive scoring performance.
The nature of playing defense in the NFL means that the fundamental shapes of the Defensive PPD graphs are similar. It is basically easier to score with a shorter field, naturally. But particular defenses perform at different levels and have curious anomalies at various places, resulting in distinct shapes that describe--and possibly define--the defense.
For the Steelers, the distinctive features of their signature are two fold. (1) Relatively flat, high-performing (low ppd value) from the 10 yard line to the 60 yardline. and (2) the sharp ramp up from the 60 yard line to the end zone (99 yard line).
In the first domain (10-60), the Steelers' defense is surrendering slightly more than 1.0 points per drive. For example, at the 20 yard line, the PPD value is 1.36 and at the 35 yard line, it is 1.15. This means that (on average), for every 6 drives that start from the 35 yard line, the Steelers defense gives up 6.9 pts (1.15 x 6), basically 1 TD every 6 drives. That the graph is relatively flat indicates that the defense is performing about the same from the 20 to the 60. This indicates that the defense is quite impressive. 60 yardline (opponents' 40 yard line) is just beyond field goal range and requires the opponent to get at least 10 yards to attempt a field goal. The fact that this value is so far below 3.0 indicates that the Steelers often do not surrender those 10 yards, holding the opponents to being just outside of field goal range and punting (or the Steelers forcing a turnover). If an offense struggles to gain 10 yards, it's about as difficult to score from the 60 as it is from the 30. The little hump near the 20 indicates that some teams managed big plays at longer distances.
From inside the 60 yardline, though, the graph turns upwards very sharply, going from 1.36 at the 60 to 3.04 at the 78 yard line. That is more than double the PPD rate in 18 yards. But those 18 yards are critical yards. The 78 yard line is the Steelers' own 22 yard line, well within field goal range for even the most challenged kickers. For drives that start at that yardline, opponents SHOULD score near (or above) 3.0. (note that the NFL average is 4.0). What this value shows is that even when the Steelers give up terrible field position (up to the 78 yardline), they manage to keep scoring to field goals instead of TDs. Closer than that and the Steelers surrender points even faster than the NFL average, allowing TDs at a high rate.
A third minor zone is inside the 10 yard line. Opponents score < 1.0 at that range, which is expected.
The Steelers perform far better (lower graph) than the NFL average at nearly the entire range; it is only at the 85 yardline that the Steelers surrender points at the higher rate. But the majority of the time, the Steelers are better than the average and not by just a little bit. They are performing ridiculously better than it.
The key zone is the 20-35 because a majority of the drives start at that range. The Steelers surrender 1.36 at the 20 yard line while the NFL average is 1.58 (14% better). At the 35 yard line, it is 1.15 v 1.85 (38% better) and you can see this very clearly on the graphs. This disparity increase until the 60 yard line; once past the 60 yard line, the Steelers' graph quickly converges to the NFL average, as could be expected.
Recall that the Raiders' defensive signature is also relatively flat, but far worse, 1.86 at the 20 (43% worse) and 2.3 at the 35 (100% worse).
Note the strength of the defense occurs right around the 45 yard line and that the performance at the 60 yard line is about equivalent to the performance at the 20. According to this graph, the Steelers should kick the ball out of bounds at the 45 yard line every time! In addition what is interesting is with the Steelers' dominance at hte 50-60 yard range, coupled with the new kickoff rule that moves the kickoff up to the 65 yard line, the risk-reward of an onside kick swings far in favor of the Steelers. In other words, mixing in a few Sean Payton-esque onside kicks throughout the course of a game might be a valuable tactic! The benefit is the offense getting the ball at the 55 yard line. The risk is that the defense is on the field at the 55 yardline, though at that range, they surrender points at about the same rate as on drives that start at the 20! Interesting indeed. (Note, I understand it's mostly enjoyable and interesting from a fantasy/intellectual perspective and note a viable gameplan). Oh, the scenarios you can come up with when you have a truly dominant defense.
Steelers' surrender more points than the average NFL team at the 85 yard line; the Steelers' PPD matches the NFL PPD but the Steelers gave up quite a few drives that started in very poor field position (85+ yardline). These are likely turnovers (Mendenhall fumble?) or big kick off/punt returns. This hump in the total points graph would vanish if the drive distribution were fixed, meaning the defense were not put into poor position by either the offense or the special teams. Although, in fairness, the valley just to the left where the Steelers points surrendered and drive distribution are near 0 is a testament to the offense and special teams.
We all understand that Offenses and Special Teams affect Defensive performance by affecting starting field position. For the Steelers, this is very evident. Fewer drives meant fewer opportunities to score and also meant fewer total points scored. The fewer drive phenomenon can occur for a number of reasons: a ball-control offense with a strong running game will often cause fewer overall drives in a game (for both teams), a bend-but-don't-break defense that allows drives to move "between the 20s" in small chunks will also cause fewer overall drives, Turnovers and punt/kick returns for TDs will also give fewer defensive drives. Without looking at the actual game data, I can't be exactly sure why the Steelers had so few defensive drives.
The Steelers defense had 167 drives while the NFL average was 184. On those drives, the Steelers' defense surrendered 222 points while the NFL Average was 343. "Better" is a tremendous understatement. "Much Awesomer" might be a better way of saying it.
The Steelers' graph is a beautiful graph, indeed. For a Raiders fan, we can only look on with jealousy, hoping that the Allen/Tarver revamp will eventually provide us with performance somewhere in the neighborhood of the Lebeau defense.
There's one last exercise left. For comparison purposes, we would like to "normalize" the values. This is an attempt to isolate the DEFENSIVE performance against outside factors like Special Teams and Offenses. Obviously, a team's 3 main components are inextricably intertwined, but for some reasonable sense of assessment, especially as we compare defenses against each other, it is desirable to try as much as is possible, to "compare apples to apples."
This is the Normalized Points graph.
This is a description/view of the Steelers' defense production given its PPD but mapped against the NFL's average drive distribution.
Dotted Blue is the PPD. Dotted Green is the total points. Dotted Red is the NFL Average for points. These are all carryovers from the previous graph. The solid Purple line is the normalized Steelers' points graph.
The drive distribution is shifted to the 20 yard line and falls off smoothly (instead of plateau-ing at the 30). The distribution then flattens at about the 50 yard line instead of falling then rising.
So here's the finale:
Steelers' total points surrendered was 222 according to my PPD model and calculated against the Steelers' acutal drive distribution.
When normalized against the NFL distribution average, this values shifts up 10% to 245. This value appears to be affected almost entirely by the number of drives. 10% more drives = 10% more points.
For comparison purposes, the Steelers' Defensive PPD Total Points value = 245. (this is about 100 pts better than the NFL Average).
Splits By Half
The Steelers' are consistent; the distinctive feature to their performance signature is the low flatline in the range of 10 to 40, which holds true for both 1st and 2nd halves. While there is some variation (natural), the performance graphs have very similar shapes.
Combining the performance with drive distributions that are fairly similar and very close in number (1st half had 83 drives, second half had 84 drives) results in total points that are close. 1st half points : 116. Second half points : 124. A difference of 8 points (6%). Basically the Steelers' defensive production was nearly equal independent of half.
Splits By Win-Loss
The Steelers' defensive performance and production during the wins was ridiculous. In 12 wins, 130 points surrendered projects to a mere 174 over 16 games, which is 10.9 points per game. Wow.
In 4 losses, 92 points surrendered projects to 368 points over 16 games (2.1x as much during the wins). 368 points. That's not very good. While it's not near the Raiders' number (over 400), 368 points would have ranked the Steelers' defense 21st overall, ahead of San Diego (377) and just behind Washington (367).
Another way of thinking about it is that The Steelers' defensive production during their losses was equivalent to having the Washington Redskins playing defense for them.
Now, how much did that poor field position play? Taking the loss values and normalizing it against the NFL Average drives, which factors out poor field position, and we get 346 total points, about 22 points difference.
So we might say that the field position factor accounted for just over 3 touchdowns in the defensive production during the four losses.
Their four losses were :
Week 1 : 35-7 Baltimore
Week 4 : 17-10 Houston
Week 9 : 23-20 Baltimore
Week 15 : 20-3 San Francisco
Other than that week one total of 35 points, those were hardly blowout losses by normal standards. The Steelers gave up 17 or 20 points in a win five times during the season. (Additionally, held opponents to 0 or 3 points in a win four times. They gave up 9 or 13 points in a win three times.), so the points allowed production during the three latter losses were still at the level of a Steelers' win; the offense just could not get enough scoring.
So, while the graphs show that the Steelers in losses appeared to be dominated, it is skewed substantially by comparison to the Steelers' defense during wins.
You can definitely see the difference and the effect that both performance and field position have on the overall point production for the Steelers' defense. In the wins graph, the points curve is mostly BELOW the drives graph, indicating fantastic PPD performance of <1.0. Meanwhile in the wins, the points curve is not just ABOVE the drives graph, but it is substantially above it.
Adjusting for Strength of Competition
While we recall that the Steelers defense performed far better than the NFL average (blue line below red), we see also that the combined offenses of the Steelers' opponents was also below NFL Average. This means that (on average), the Steelers' opponents had below-average offenses.
Here's a list of the opponents that the Steelers faced, total points scored, and the percent difference compared to the NFL Average. Recall that the NFL average for total points scored is 343.
Baltimore (x2), 368 pts, +7%
Seattle, 300pts, -13%
Indy, 234, -32%
Houston, 386, +13%
Tennessee, 305, -11%
Jacksonville, 222, -35%
Arizona, 203, -12%
New England, 498, +45%
Cincinnati (x2), 330, -4%
KC, 210, -39%
Cleveland (x2), -34%
SF, 380, +11%
St Louis, 196, -43%
Aggregating and modeling the opposing offenses gives the above graph and an average total points scored of 244 pts (-29%) for the year.
What this says is thatthe offenses that the Steelers faced had, as a whole, offensive production that was 29% below average. Combining that with The Steelers' quality defense and the result is as you might expect : dominance.
The Steelers' defense was dominant on the field, When the offense and the defense were functioning together and the team was winning, the defense was frighteningly good, averaging <11 points per game.
But the Steelers' defense took advantage of a very favorable schedule that found 7 games where they faced offenses that were producing less than 30% of the NFL average points. Both the AFC South and the NFC West were in rather down years and the KC matchup was nice.
Even at that, though, the Steelers did what they needed to do on the field, holding these low producing offenses even below their own production level. But when we adjust to an NFL average, there are areas where the defense does appear mortal. Notably, right around the 20 yard line, the Steelers are performing near the NFL average.
By simply looking at raw values and performance, the Steelers are intimading and impressive. Once adjusting for the Strength of their competition, the mystique fades somewhat, bringing the Defense back to earth.