USC 2006 Run Defense: Overall & Game by Game Comparison In a previous post I put up some overall trend information from 2001-2006 related to USC's defense in the Pete Carroll era. I thought it was worth taking some time to go back and look at USC's run defense this past season in greater detail and see how the unit performed overall, first half versus second half, and some snap shots of each game. As the run defense trend chart shows below there is no doubt that the run defense in 2006 was better than the 2005 unit however it did not quite reach the levels of the 2002, 2003, or 2004 squads. Remember with a shorter game in 2006 due the clock rule adjustment the 2006 figures also have about six to eight (10%) fewer plays to defend against than in recent years. In 2007 the unit might have a chance to reach those levels again but we will have to wait and see. Let's start with the high level and compare the first half versus the second half and see how things look in hindsight. Table 1  Season Comparison 2006 USC Defense Games 1-6 Games 7-13 '06 Season Run Yards Per Game Allowed (4Q's) 100.2 83.3 91.1 Run Yards Per Game Allowed (3Q's) 82.8 69.6 75.7 Opponent Rushes Per Game 27.0 32.1 29.8 Opponent Rush Offense Ranking 50.2 57.0 53.8 # of Big Runs Allowed (>10 yards) 16.0 15.0 31.0 Avg. Big Runs Allowed Per Game (>10 yards) 2.7 2.1 2.4 Yards Per Carry Allowed  3.7 2.6 3.0 Avg. Sacks Per Game 1.5 3.7 2.7 Opponent Rush Offense Ranking 50.2 57.0 53.8 Avg. Sacks Per Game 1.5 3.7 35 Sack Yards Per Game 11.0 22.7 225 Rush Yards with Sack Yards Adj. 111.2 106.0 108.4 Opponent Rushes - Sacks Adj. 25.5 28.4 27.1 Rush Defense Sack Adj. 4.4 3.7 4.0 Overall in this table you can see that a couple of things were evident during the season. The unit did improve in terms of rush defense as the season progressed in terms of both average yards allowed (whether measured at the end of the 3rd or 4th quarter), and yards per carry allowed. The strength of the opponent's rush offense was slightly weaker the second half of the season but this effect is marginal I suspect. The first half had a weak rush team in Arizona (110th) and the latter had Stanford (115th). The first half had good run teams such as Arkansas and Nebraska and the second half had Cal, Oregon, and Michigan. All in all it seemed a decent comparison. In terms of similarity both halves of the season allowed about the same number of big runs of over 10 yards. There were 16 in the first half and 15 in the second half. The effect of these 31 plays was huge however in terms of impact on the numbers which I will point out in the final summary. Only 2.6 yards per carry were allowed by the defense in the second half of the season while 3.7 were allowed in the first half using the NCAA statistics. Here the comparison is slightly flawed however in my opinion as Sedrick Ellis was out for USC for three games during the season in the first half of the year. I suspect the numbers would be closer if Sedrick had been healthy and available the entire time but injuries are just part of the game. Also one thing I was most curious about was how the inclusion of Sack data was affecting the run equation. USC did blitz more in the second half of the season and produced more sacks per game. Unlike the NFL the NCAA takes the sack yardage off the opponents run game which can give some false impressions at times. Statistically the USC run defenses in 2002-2004 for example were all greatly aided by the fact they also lead the country in sacks and sack yardage. This significantly affected the year end rush defense totals. In 2006 since my data is more complete I can measure the extent of this effect for this season more easily. As you can see from above adding the sack data back into the opponents rush totals inflates the totals somewhat. The run defense allowed 111.2 yards per game the first half and 4.4 yards per carry while the latter part of the season the unit allowed 106.0 yards and only 3.7 yards per carry. The difference between the two halves is about within normal variation so it is tough to conclude the unit was much better in the second half but based upon averages only it is a fair statement by a small margin. It is probably more interesting to look at the game by game trend and see the differences at that level. Here are some charts that summarize the game by game trend in 2006 with respect to rush defense. I'll show it both ways a) using the NCAA statistics and b) with the sack yards put back in for comparison. Here is the first chart that uses the NCAA method and summarizes the game by game totals. The average yards allowed by the rush defense in 2006 is indeed 91.1 yards however there is a fair amount of variation in the numbers game to game. In order to show this from another angle let's organize it into a histogram for comparison. Here you can see that "normal" games tended to allow rush yards in the area of 100 to 150 yards when compiled this way. The average is lower due to the excellent job done by the defense in the game in the Rose Bowl against Michigan, and early on versus Nebraska and the weaker run opponents Arizona and Stanford. Let's add back in the sack yards and see what the picture shows from a trend point of view and a histogram. Next below here is the game by game trend chart in 2006 with the sack yardage netted back in. Sacks occur on pass plays but the data is subtracted off the run totals for yardage in the NCAA. As you can see the average is higher and this may be a better look at how the run defense performed over all in 2006 game by game. The trend is fairly flat again overall and the low points are not as low since the sack yards are now back in. Here is the data organized into a histogram again for comparison. The data clumps up around the 125-150 range in terms of yards allowed. The average is lower of course due to the games to the left of the central distribution. Which view of the data is better sort of depends upon your opinion. In either case the 2006 USC run defense was good but not great. Also it did not vary too significantly in terms of performance overall during the season. Arizona was an excellent run defense game in the first half although the Wildcats were only the 110th ranked rush offense in the nation. The Michigan game was the best run defense game of the second half against a top team and hopefully a sign of things to come in 2007. Regardless of how the above data shakes out it only reflects what happened on the field. At the end of the day the problems and breakdowns that lead to big run plays are what the coaches work on and try to improve. Here are three examples of things that stood out in hindsight in 2006 with some examples that I am sure will be worked on by the staff. Example #1 Run defense is mainly a process of defending "gaps" in the Pete Carroll base defense .  USC plays very little 2 gap football and instead tries to put a man in each gap to defend the run. There are exceptions to this method however against certain alignments. Sometimes the nose tackle aligns directly over the center. In other cases USC will try to defend eight run gaps with only seven defenders. Here are some examples of that from 2006. 2006 ASU Game 2006 WSU Game 2006 WSU Game The first picture on the left shows an example where the nose tackle is attempting to control two gaps up front. From this position he can slant into either gap or play the center straight up and defend both A gaps. This is an example of a Bear or Double Eagle type front used at times by USC but not very often. The WSU pictures to the right are examples of attempting to control 8 gaps (four to the left and four to the right) with only 7 bodies up front. USC had some trouble against this 2TE Ace back formation when Washington State did run out of it. A couple of times the RB Tandy broke off gains of over 10 yards by finding the open hole in the line. When USC can't contain the run with seven bodies it is forced to cheat a safety forward for run support. That is fine in the event of a run play but WSU will send the 2TE's and 2 WR's out into the pattern on some of these formations on a 4 verticals pass route. Then any defense has to have 4 deep backs in the secondary just to defend the pass and prevent the big play. I expect we'll see more single back sets out of the 2TE alignments in the Pac 10 next year against USC as a result in certain situations. Example #2 ASU Draw Pic 1 ASU Draw Pic 2 ASU Draw Pic 3 ASU Draw Pic 4 Here is a sequence that gave USC trouble all year long. I will merely use the ASU game example since it affords a better viewing angle. As I looked at the data for the 31 big run plays of over 10 yards that USC gave up in 2006 the two biggest types were either busted containment plays and QB scrambles (example below) or the basic draw play on passing situations. Here is a good example of one of the latter. It starts with USC probably in a Cover 1 shell and #29 Taylor Mays in deep center coverage. #4 Kevin Ellison is closer to the line of scrimmage and covering a tight end. The play looks like a pass play in picture 2 as the defensive lineman are starting back in pass protection (hats up) and not down low for run blocking. The USC LB's #58 Maualuga & #55 Rivers bite on the fake and both start backwards. In the picture 3 however the play is actually revealed as a draw play up the middle. The center and guard double team Sedrick Ellis on a combination block and the USC lineman Jackson, Barrett, and Cushing are influenced blocked to the outside allowing them up field. In picture 4 you can see the TE gets out on #58 Maualuga and one of the lineman from the initial double team on Ellis splits off and blocks #55 Rivers. The back gets upfield before the LB's can shed their blockers and make the tackle. The tackle was often then made after a 7-12 yard gain. Expect to see teams continue to use this play in pass situations until USC can more effectively shut it down. When properly executed this play works very well against any team and not just USC. The other types of big plays came from containment breakdowns or in one case a trick play. Defensive coordinators hate to face a QB that is a threat to run as they are hard to scheme against. Even with a good game plan running QB's can still shred the defense at times (e.g. Ell Roberson of KSU in 2001-02, Bryan Randall first quarter in 2004 vs. Va Tech, and Vince Young vs. Texas in 2005 Rose Bowl). Even non-running QB's can cause problems. The ASU QB broke off a big run against the Trojan Defense in 2006 as did Brady Quinn of Notre Dame. Neither is a "running QB" however they can exploit the defense when it over pursues or gets out of position. The pictures here illustrate what can happened when a defense gets caught being out of position while being aggressive and attempting to pressure the quarterback. I believe this is the biggest run play surrendered by USC all year long. After the snap in picture 2 Sedrick Ellis stunts around behind the other DT of USC and attempts to rush Brady Quinn from behind. Unfortunately this aggressive play opens up a gap to the right that Ellis vacated. The responsibility appears to be left to a linebacker to protect but he too drifts over to the left it appears in this sequence. As a result a sharp veteran QB like Quinn tucks the ball down and scampers down field for a 60 yard gain. Aggressive plays like stunts and blitzes work and of course are a necessary part of the game but we tend to forget that they also have a risk reward curve associated with them. This one backfired in hindsight and unfortunately lead to a big gain. Pic 1 Pic 2 Pic 3 Pic 4 Pic 5 Pic 6 The other big play that stood out as a example in the data this past season was a reverse play that University of Washington ran against USC in the second half. Marlon Wood ran 46 yards on a reverse that caught USC off guard and lead to a big gain. Picture 1 Picture 2 Picture 3 Picture 4 Here are four pictures from the opening Husky drive in the third quarter. To open the second half they executed an on-side kick that caught USC off guard. On the first play of the ensuing drive Washington lined up with the strength of the formation and 2 WR's to the left. The 3rd WR Marlon Wood aligned to the right and came on a reverse that USC did not pick up right away. The LB's were all a step slow to respond and the WR's on the left ran downfield routes that carried their defenders with them. Wood ran for 46 yards down the sideline and was finally forced out inside the 10 yard line. From there the USC defense stiffened and held Washington to a field goal. Relatively few big trick plays were successful against USC of this type but they do work from time to time. Playing overly aggressive can play into the hands of the opponent at times as can playing too conservatively, or simply falling asleep at the wheel for a moment. For simplicity I'll summarize it this way in the end.  Using the NCAA statistics (rewarding the run defense for sack yards) USC allowed 1184 yards total on 399 carries for a 2.97 yard average in 2006.  They ranked #9 in the country tied with UCLA in this category.  USC gave up 582 of those yards however on just 31 big run carries that went for an average of 18.8 yards per pop.  In other words USC actually allowed a measely 1.64 yards per carry on 368 carries for 602 yards.  Then on 7.8% of the run plays (31 to be exact) they allowed 49.2% or 582 of the yards. Those 7.8% are what the coaches will work on fixing. The keys will be reading draw plays, not overly biting on play fakes, sniffing out trick plays and playing solid gap control defense. If USC can eliminate even half the 7.8% plays that lead to big gains then the run defense will be something special in 2007.