Why Johnny Can't Pass Rush: And What We Should Do About It
by CPT Hoolie on Sep 7, 2010 4:25 PM CDT
Email
Print[Bumped. The Captain is back! -KJ]
There seems to be some angst among the Spartan fanbase about MSU's perceived lack of a pass rush, based on 2009 and on the first game of 2010.
I took a look at the stats and found that it may not be as bad as we fear, relatively speaking.
CFBStats' excellent database only goes back to 2006, so I had to use Rivals' stats database for the seasons 2003-2005.
Here are the sacks per game averages for that period:
| Year | Sacks / game |
|---|---|
| 2003 | 2.45 |
| 2004 | 2.50 |
| 2005 | 1.90 |
| 2006 | 2.09 |
| 2007 | 2.12 |
| 2008 | 1.98 |
| 2009 | 2.00 |
As you can see, the national average number of sacks per game decreased by 0.5 sacks / game from 2003 / 2004 to today.
MSU's sack stats during that time period:
| Year | Sacks / Game |
|---|---|
| 2003 | 3.75 |
| 2004 | 1.92 |
| 2005 | 1.18 |
| 2006 | 0.67 |
| 2007 | 3.42 |
| 2008 | 2.17 |
| 2009 | 2.75 |
MSU has been right at or slightly above the national average except during the terribly stinky years of 2005-2006.
I think there are two years that we have imprinted on as a standard for "successful" pass rush seasons. 1999 (not shown here) was an amazing year for Julian Peterson; he had 15 sacks (and Robaire Smith added another 8, for a combined total of 23 sacks), but honestly, he was a linebacker rushing as a defensive end (a la Pat Swilling -- remember him?). In 2007, the "Sackmaster", Jonal St.-Dic, had 10 sacks (combined with Ervin Baldwin's 8.5 sacks for a combined total of 18.5, and up from 2 sacks the year before). As you can see from the stats, these two years were, honestly, outliers. (2003 was a more combined but anonymous effort, when Greg Taplin (10 sacks) combined with Clifford Dukes (7) and Matthias Askew (6) for 23 of the team total 45 sacks.)
The bottom line is this: sacks are down across College Football. The reason for this, I believe, is the rise of the spread offense.
Primarily, most passing spread offenses rely on 3-step, quick release, short passes. Even if a defensive lineman had a running start and was untouched by the offensive line, there is little chance that they can get to the quarterback quickly enough to get a sack before the quarterback releases the ball (save for a "coverage sack").
In addition, wider splits by the offensive line increase the distance that an edge rusher has to go to reach the quarterback. (Many teams increase this distance even further by designed rollouts away from the blind side.)
Finally, the five-man offensive line has superiority of numbers against the four-man defensive line. The only way the defense can regain superiority is to bring blitzers -- but the LBs and DBs are needed in coverage since the number of eligible receivers has now increased under the spread.
Look at the leaders in sacks allowed (i.e., fewest sacks per game).
| Name | G | Sacks | Sacks/G | |
|---|---|---|---|---|
| 1 | Boise State | 14 | 5 | 0.36 |
| 2 | Stanford | 13 | 8 | 0.62 |
| 3 | Army | 12 | 10 | 0.83 |
| 4 | Oklahoma State | 13 | 11 | 0.85 |
| 4 | Air Force | 13 | 11 | 0.85 |
| 6 | Georgia Tech | 14 | 12 | 0.86 |
| 6 | East Carolina | 14 | 12 | 0.86 |
| 8 | TCU | 13 | 12 | 0.92 |
| 8 | Georgia | 13 | 12 | 0.92 |
| 10 | North Texas | 12 | 12 | 1 |
| 10 | Arizona | 13 | 13 | 1 |
| 10 | Nevada | 13 | 13 | 1 |
| 10 | Oregon | 13 | 13 | 1 |
Some of the spreadiest of the spread offenses there. The one that jumps out as not fitting is Army, but Army only attempted 161 passes out of 754 total plays (80%-20% run-pass balance) and so provided fewer opportunities to be sacked.
What should we do about it?
I think it is time to recalibrate our expectations. I believe the days of 20 sacks from the D-line are long gone, at least until we see the resurrection of the pro-style dropback passer that takes 5- and 7-step drops.
A more important defensive line metric against spread teams will be QB hurries (MSU had 5 hurries in the Western game). I also believe that the defensive line's role against spread teams changes from one of pass rush to "soak up as many blockers as you can, clogging gaps and preventing the quarterback from sprinting up the gaps in the middle".
Therefore, I have reset my expectations to 2 sacks/game with 2.5 being a great success.
MSU had 1 sack against Western Michigan, so they better get busy on Saturday regressing to the mean!
This is a FanPost, written by a member of the TOC community. It does not represent the official positions of The Only Colors, Inc.--largely because we have no official positions.
21 comments
|
0 recs |
Do you like this story?
Comments
Looking at it through the lens of pass attempts per sack
Another way to look at this might be the number of opposing pass attempts per sack. This would tell us how many times we were able to sack the quarterback relative to how often the opponent passed the ball. The general trend still pretty much nets out with your numbers.
Using cfbstats.com and the Rivals stats you posted (and adding the number of sacks to the number of pass attempts to compensate for the rule in college football where sacks count as a rushing play rather than a passing play), this is what I came up with:
Pass Attempts per Sack (2003 – 2009)
2003 – 10.6
2004 – 17.0
2005 – 30.1
2006 – 23.2
2007 – 11.9
2008 – 17.2
2009 – 15.6
This pretty closely mirrors your analysis with the exception that the 2005 pass rush may have been worse than 2006, simply by virtue of the fact that we had fewer sacks (13 vs. 16) with slightly more pass attempts against (391 vs. 371). Same goes for last season – we had more overall sacks (and, thus, more sacks per game) than in 2008, but it was basically a function of opponents passing significantly more on us (514 pass attempts in 2009 vs. 448 in 2008).
As for the Western Michigan game, we had 2 sacks on (essentially) 55 pass attempts, which equates to 27.5 pass attempts per sack. So, obviously there’s some work to do.
by Spartan-Football on Sep 7, 2010 8:39 PM CDT reply actions
How many of the Western pass attempts were "downfield" passes?
Base stats: Western passing 4.5 YPA.
Looking at the drive chart, I count 7 passes complete for 10 yards or more. Without watching the tape again, though, not sure how of those were from YAC and how many actual attempts were more than 10 yards downfield.
Averages
The problem with using averages here (good research, by the way) is that your number include 120 teams… so you are seeing a strong representation not only from what we would consider “bad” teams (like WMU), but also from “terrible” teams (like Louisiana-Lafayette). In this context, being “slightly above average” just doesn’t cut it for a team that is supposedly bowl bound from a BCS conference.
I think you are correct on the spread offense and the widespread adoption of the WR screen and plays like it. These ultimately make it difficult to get to the QB. I think the problem most of us spartan fans see is not really the “sacks” stat, but more of the “hurries” stat (is this even kept in college? can we look it up?). We seem to get very little pressure on the QB, especially on the downs where it would really help matters (3rd and long, for example) and where the QB is doing more than taking a 3 step drop and tossing it to a RB. Pressure on the QB doesn’t necessarily have to translate into a sack (although it would be nice if it did regularly)
"It's a trap!"
Bad/terrible teams are also bad/terrible at pass blocking
So I’m not sure using the full sample of FBS teams skews the findings.
MSU ranked 2nd in the Big Ten in sacks last season (conference games only). Also, the team ranked first in the league in QB hurries through 11 games (including non-conference games).
There’s certainly still room for improvement—and my goal is to track all these numbers more closely this year (see the most recent post)—but I think CPT Hoolie’s main point is spot on: we need to lower our expectations as to what an effective pass rush looks like statistically.
Fight for The Only Colors: Green and White!
by KJ@theonlycolors on Sep 8, 2010 7:23 AM CDT up reply actions
I don't know what the correlation is between win percentage and sacks.
I’ll have to run it.
But I looked at the national averages initially to determine what a reasonable number of sacks per game was, and noticed that the overall trend was that the number of sacks was decreasing.
Yes, but how much of that is number of opportunities?
Let’s face it, nobody bothered to run on us last year because the secondary was so prone to getting torched. 10% more sacks in 50% more attempts = bad.
(I don’t know if this is actually the case, but it’s something to look at.)
Per the scatterplot below
MSU ranked 3rd in sack% in Big Ten play.
Fight for The Only Colors: Green and White!
by KJ@theonlycolors on Sep 9, 2010 7:21 AM CDT up reply actions
interesting article
but, I feel like there is a problem with only looking at sacks as an indicator of the defense’s pass rush. Specifically, by including lb’s, cb’s, and def. linemen in the measurement group, you are ignoring the problem our 4 man front is having with getting pressure on the qb.
The sack numbers that you discuss seem to be primarily (at least last year) a result of bringing a linebacker or member of the secondary to get extra pressure on the qb. While this may result in an increase in sacks, having to sacrifice a lb or a cb to get that sack leaves us more exposed in the passing game.
What would be interesting would be to measure the amount of sacks and hurries caused by our front four over the past few years. This sort of goes more to KJ’s pressure stats that he talked about last year (at least, I think it was KJ, in his attempt to quantify every aspect of football – which i think is a good thing!). That would be more indicative of how our true pass rush is, and how this line stacks up to past defensive lines. It would also give us an idea of how free the rest of the defense is to stay in coverage in passing situations.
Good point
This gets at the crux of the issue. It is definitely possible to create pressure on the quarterback at the expense of pass coverage. The real question is whether that’s what MSU is doing or not. As KJ points out above, we were getting pressure last year. But our pass defense numbers suggest that it may have been at a cost. The jury’s still out on that question for this year.
Interesting question
Here’s my attempt to quantify it (conference-only data):
At first blush, it doesn’t look like the pressure MSU generated came at the expense of giving up big plays in the secondary, at least compared to teams with less impressive pass rushing numbers.
Yards/completion seems like the sensible number to look at here, although it doesn’t fully capture how bad the MSU pass defense was last year: last in yards/game, second-to-last in interceptions and completion percentage.
Could be that the linebackers blitzed a lot while the DBs gave a lot of cushion (both fit my intuitive sense), resulting in a lot of short completions and a lack of big plays by the DBs.
Fight for The Only Colors: Green and White!
by KJ@theonlycolors on Sep 8, 2010 8:12 PM CDT up reply actions
I think you nailed it
As a side note, the horror that was our pass defense last year defies many attempts to capture its awfulness. For example, MSU was 93rd in the country in pass plays yielded of 15 yards or more and 113th (out of 120 teams) in plays of 25 yards or more. So although our yards per completion is not horrible, I would say it mainly reflects a refusal/inability to adjust at all to stop the short pass. This resulted in SO MANY short completions (MSU faced more pass attempts – and completions – than all but 4 teams in the country) that it obscures the fact that, despite the cushion the DBs were giving they STILL couldn’t stop the deep ball. They gave up 32(!) pass plays of 25 yards or more in 13 games. . .OK, I’m better now. (Stats from cfbstats.com.)
It might be helpful to look at completion percentage as well
If you could combine yards per catch with completion percentagem that might give you the real picture – lots of high percentage short to midrange passes that we never stopped, enabling teams to march down the field throwing the ball. It was probably a function of having to drop everyone we could deep into the secondary to stop long passes. I guess that was our optimal strategy though – the alternative was giving up lots of long passes, which is worse than giving up short ones.
by TheCrestedHelm on Sep 9, 2010 8:54 AM CDT up reply actions
We were merely very bad in comp % allowed
Rather than historically awful: 60.9% allowed (80th in the country and ahead of only Illinois in the conference). The problem is that we did give up lots of long passes too, more than anyone in the conference. It’s just that we gave up even more short ones. Teams figured out immediately that the defense could not stop passes of any kind and did nothing but throw the ball. Only Oklahoma State and Troy played the same number of games or fewer and faced as many pass attempts as MSU. This suggests that 1) Offenses were able to stay on the field to throw more passes and 2) No adjustments were made to stop the bleeding.
I guess I figure
being merely bad at both preventing completions and yards per catch = really bad at defending passes. I agree though – the stats really don’t describe how bad we were against the pass last year.
by TheCrestedHelm on Sep 9, 2010 10:10 AM CDT up reply actions
Scatterplots woo!
Who are those 3 teams down in the bottom half of the graph?
Also, since we have been talking about “hurries” being generated by defenses vice sacks, I wonder what it would look like if you plotted sack % against yards per attempt (that is, including incompletion) and sack % against opponent completion percentage.
OSU/Iowa/PSU (from bottom to top)
The team way over at the right is Purdue.
Scatterplot vs. yards/attempt looks almost identical.
Scatterplot vs. completion percentage:
More of a neat linear trend, with the outliers being teams with a very good secondary (Iowa) or very bad ones (MSU, Michigan).
Fight for The Only Colors: Green and White!
by KJ@theonlycolors on Sep 11, 2010 11:11 PM CDT up reply actions
I think it's the 4.5 YPC stat that really stands out
We do need to get more consistent pressure on the QB. I’d have to go back and watch the game to determine whether WMU’s measley yards per catch number was due to throwing a lot of very short passes by design, or whether they were trying to go downfield but couldn’t because the coverage was there, so they checked down. If it’s the former, then you really can’t blame the pass rush because they were just getting rid of the ball early enough that we had little chance to get there. If the latter, it’s definitely an encouraging sign for our coverage but maybe not so much for our pass rush. Either way it appears we limited YAC which speaks to good tackling fundamentals, and that is a very good sign.
All in all I think we had a workman-like performance this past weekend, and have reason to be relatively pleased. There’s nothing other than the dropped passes (and a couple of dropped interceptions), to really be disappointed about. We performed at an acceptable level on D and in the passing game (other than the drops) and really excelled in the run game, even considering the nature of the competition. We seemed to sleepwalk through parts of the game once we’d built a solid lead, which is something we’ll have to watch going forward.
I think the point about the spread offense depressing sack totals is spot on.
The 4.5 was YPA -- yards per attempt.
That includes incompletions but not, as Spartan Football says above, sacks.
Western was actually at 7.8 YPC.
By comparison, MSU was at 8.5 YPA and 14.2 YPC.
It's still a pretty good number
Thanks for catching my typo. Our opponent YPA is still right in the wheelhouse compared to numbers put up by other Big 10 winners facing substandard competition this weekend. I’m not ready to call us a top shelf defense yet but early signs indicate we may at least be average rather than atrocious against the pass this year. But it’s too early to tell anything.
by TheCrestedHelm on Sep 8, 2010 11:01 AM CDT up reply actions
"slightly above average"...
…was a careless term on my part. Let’s see if I can use a more precise definition — maybe “1 standard deviation above (or below) average”.
In 2009, the range of sacks/game was 0.83 to 3.69. Mean 1.99, median 1.92, standard deviation 0.57. So in 2009, MSU’s sacks/game of 2.75 was 1.3 standard deviations above the mean.
There, hopefully that’s better!
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
Certain photos copyright © 2012 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
