The Blitz is a Lie

“We need to blitz more.”

This is something I hear a lot, especially after a USC loss.  Doubly so after a loss that can be attributed back to a poor defensive performance like the one we saw against Stanford on Sept. 19th.  However, in my experience, this is often a false statement.  Many fans cannot accurately diagnose a blitz in real time but instead rely on after-the-fact analysis (ex: the QB got sacked and in watching the replay it is easily recognized as a blitz which caused the pressure).  However, when a team blitzes a lot but is still unable to put consistent pressure, this often leads to many fans incorrectly clamoring for “more blitzes.”

First, let us look at some defensive numbers regarding the blitz during the USC vs Stanford game then we’ll take a look at some further analysis that may shine a better light on what happened on that long Saturday night.

The Blitz

A blitz is a high risk-high reward play.  Typically, it is defined as any play that you send one or more more non-defensive linemen in to pressure on the offense’s side of the line of scrimmage.  In the terms of a 3-4 defense (which USC employs), rushing the three defensive linemen and a single linebacker is often called a “Dog” scheme and is not considered a blitz.  Here, we’ll simplify the definition of the blitz and count it as any time USC sent five or more defenders in as rushers (many other pressure concepts such as the zone blitz, twists, and stunts are ignored here for simplicity).

I rewatched the game and tracked 72 Stanford offensive plays (I didn’t count things like spiking the ball or kneels as plays).  First, I want you to formulate a guess on how many of those plays USC sent in a blitz.  Really think about the game and your game plan observations.  Have a number ready?  Initially, my gut feeling was about 20% blitzes, so 14 or 15 plays with a blitz.

In fact, the real number that I tracked was 43 blitz plays, or 59.7% of defensive plays sending five or more defenders in for the rush.  Surprisingly high, I thought.  Let’s take a closer look.

Of the 72 Stanford offensive plays, I noted 31 passes and 41 runs (note that I counted QB scrambles, sacks, and screen plays as passing plays whereas most official stats often track these as runs for college).  Let’s look at some passing numbers first.

Blitzing the Pass

On the 31 passing plays, 12 were blitzes (38.7%).  Against the blitz, Stanford gained 10.0 yards per pass against the blitz and 9.8 yards per pass against a non-blitz.  Completion percentage wise, we’re looking at a 66.7% completion rate against the blitz (77.8% if you count the pass that hit the tight end in the hands as a “should have been completed”) versus an 85.7% completion rate against a non-blitz.  Lower completion rate against the blitz but slightly higher yards, signifying they are making more yards when completing passes against the blitz.

Now let’s look at pressure against the pass.  USC got three sacks but only one sack came from a blitz (9% of blitzing plays against the pass) and two from a non-blitz (11.8% of non-blitzing plays).  Sacks aren’t the only end goal of the pass rush though.  54.5% of blitz plays resulted in the QB being pressured versus 35.3% of non-blitz plays.  You can see this reflected in the completion percentages.  Interestingly enough, “big blitzes” (sending six or more rushers in) resulted in no plays with pressure and no sacks.

Blitzing the Run

Now let’s look at the run.  Against 41 running plays, 31 of those plays were blitzes (75.6%).  Tracking blitzes on the run is trickier since upon reading a “run”, defenders go into their run fits and start filling gaps.  Who is filling a gap and who was blitzing?  I keyed in on the first step and initial attack movements of the defender to count the blitzes.  There will be some inconsistencies with this method, namely delayed blitzes and zone blitzes may be miscounted for one reason or another, but we’ll just go with best guesses.

Against the blitz, Stanford gained 4.2 yards per carry versus 4.1 against a non-blitz.  Sending the pressure did not result in less yards by Stanford.  My initial thoughts are that USC’s linebackers were not able to penetrate the bigger Stanford offensive front on the run, regardless if they had forward momentum of the blitz or not.  More on this later though.

Successful Plays

A play is considered “successful” for the defense if they hold the offense under 50% of the yards-to-go on 1st down, 70% on 2nd down, and 100% on 3rd or 4th downs.  I tracked whether a defensive play was successful or not in relation to the blitz.

Over the 72 plays, USC was successful in defending 33 of them (45.8%).  When blitzing, USC was successful on 48.8% of defensive plays compared to 41.4% of non-blitzing plays.  Not a huge difference between the two margins (about a 2 play difference).  The interesting thing is how unsuccessful big blitzes were.  Big blitzes only had a success rate of 15.8%.

So What Did Happen Against Stanford?

So USC blitzed on over half of their plays.  Why were we so unsuccessful overall in stopping Stanford?  My thoughts after re-watching the game are three-fold.

First, Stanford expected us to blitz.  USC blitzed heavily in the week prior against Idaho and Stanford expected us to do the same against them.  They utilized a number of techniques to combat the blitz, most notable being the QB rollout and screen passes.  Additionally, Stanford would often times keep extra blockers (RBs and TEs) in for protection.  On average, Stanford had 5.7 blockers in on each play, keeping at least the running back for protection on 66.7% of their plays.  They kept two additional blockers in on 22.2% of their plays.  On average, Stanford had 1.3 more pass blockers than we had pass rushers.  On only 4 plays did USC have equal numbers of rushers to blockers.  Additionally, the only time we had more rushers to blockers were screen plays, which is exactly what Stanford wanted.  Against these screen plays, our rushers did not adjust properly towards the screen and we gave up big yards on key plays (10 yards on a 2nd and 15 as well as 19 yards on a 3rd and 7, both which lead to 2nd half Stanford scores).  However, keeping extra blockers in means you have less passing options as you only have three or four guys out on routes.  Which leads me to my second observation.

Stanford utilized size mismatches effectively.  I noted at least two game-pivotal plays which size mismatches played a large role in our defensive breakdown on a blitzing play.  The first on a 3rd and 9 on the USC 16 early in the 4th quarter.  Here, Stanford expected our blitz (we sent in six) and kept the RB in for protection.  They easily found the size mismatch they wanted with Safety Marvel Tell III (6′ 2″, 190 lbs) attempting to cover TE Austin Hooper (6′ 4″ 248 lbs).  Hooper just used his body to easily box out Tell.  In another notable play, USC sent in a big blitz of seven rushers, leaving one-on-one coverage.  Stanford again predicted our blitz and negated this by calling a QB rollout, causing none of our rushers to pose any sort of threat to Hogan.  Safety Chris Hawkins (5′ 11″, 190 lbs) attempted to cover Devon Cajuste (6′ 4″, 227 lbs) in the endzone.  In this case, the wingspan and size advantage got us.  By utilizing size mismatches, Stanford was able to mitigate some of the effects of having fewer receivers out on routes.

Finally, an observation defending the run.  USC’s defense was often times in the correct position and gap responsibilities for the play, but would fall short on leverage and execution.  What this means is that the USC defenders were getting to the right spots but losing their one-on-one battles consistently.  We can quantify this by looking at yards after contact.  Stanford gained an average of 4.1 yards per carry.  Not great numbers, but it was four yards like clockwork.  When you have that type of consistent four yards every rush, it leads to first downs and extended drives.  It gets more interesting when you look at yards after contact.  Stanford’s rusher made first contact with a defender at 1.6 yards on average.  That means that after contact, Stanford would gain an additional 2.5 yards per carry.  Much of this was simply the runner always falling forward.  Often times, our defenders did not have the size, the leverage, or have their blockers beat enough to stop the runner in his tracks or drive him back and could essentially only trip the rusher.  In my opinion, this was the greatest shortcoming of the defense.

Conclusions

Joe Paterno often said: “You’re never as good as you think you are when you win; and you’re never as bad as you feel when you lose.”  I think those words ring true here.  I’ve rewatched and analyzed a fair share of USC losses over the past few seasons.  This one hurt a lot live but on film study it definitely wasn’t one of our worst.  The issues I saw seemed fixable.

Furthermore, I think our defense is much more suited towards stopping speed rather than size.  Most of the Pac-12 has gone speed, so we won’t need another “size” gameplan of this scale this season unless we have a rematch with Stanford again in the conference championship game, something that I think is completely achievable.  There’s a reason why both ESPN’s Football Power Index and the S&P+ Rankings both still have USC in the top 10 and highest rated in the Pac-12.  Let’s move forward and prove them right.

And remember, we didn’t need to blitz more…

USC vs Stanford ’13 – A Look at Some Statistics

This post will look at some key statistics from the USC vs Stanford game.  We’ll look at the run vs pass play calling from both teams in a variety of angles.  Then we will look at some other interesting statistics for 3rd down conversions, rushing histograms, and yards per drive.

Play Calling

USC came into this game with a very strong running game and a good but somewhat inconsistent passing game.  Overall, USC called running plays 25 times (39%) and passing plays 39 times (61%) [Please note that I consider sack plays as passes, unlike the NCAA which considers it a run].  In the five games prior to Stanford that Coach Helton has been calling the plays, there were 169 runs (53%) and 148 passes (47%).  We passed the ball a lot more against Stanford than we typically have done.

First, let’s break down the play calling by USC based on down.

2013_USC_vs_Stanford_RunVsPass_DownUSC ran the ball on 59% of their first downs.  This is a slight tick down from Helton’s previous games where he ran the ball on 70% of first downs.  From there, USC became very pass heavy.  Second down had nearly twice as many passes and runs and third down was almost exclusively passing plays.  In fact, our only “running” play on 3rd down was on the 3rd and 19 just before our game winning field goal.  This  was a designed Cody Kessler run which was called primary to center the ball.  For all intents and purposes, USC passed the ball exclusively on 3rd down.  This included passing on a 3rd and 1 and a 3rd and 3 (both incomplete).

For comparison’s sake, Helton has been exactly 50/50 on runs and passes on 2nd down prior to this game.  On third down, he has run the ball 22% of the time and passed 78% of the time.  For any third down that has been longer than a single yard, Helton has come out typically throwing (generally 70% or more passing plays).  This means that overall, Helton threw the ball more on every down compared to his usual tendencies.

Now let’s look at USC’s play calling by field position.  Please note that in the chart below, all “Goal to Go” plays are also counted in “Red Zone.”

2013_USC_vs_Stanford_RunVsPass_FieldPosUSC had the greatest discrepancy between runs and passes behind the 50 yard line.  With lots of room to work with and early in drives, USC would have more than double the amount of passes compared to runs.  Once we crossed midfield, USC began to run the ball more.  This is actually a bit opposite from what was typical of Helton.  Typically, Helton has a slight edge towards running behind midfield (56% running, 44% passing) and opens up the passing between midfield and the redzone (44% running, 56% passing).  The redzone and goal to go categories lined up generally with Helton’s tendencies with around 71% running in the redzone and 78% with goal to go.

Finally, let us look at the play calling based on quarter.

2013_USC_vs_Stanford_RunVsPass_QuarterAs can be seen here, USC started and ended the game with a pretty good amount of balance.  Even the second quarter was pretty balanced as well.  The third quarter was when USC started flinging the ball around with nearly three times as many passes as runs.  For comparison’s sake, Helton has typically been very balanced in every quarter except with a lean towards the run in the 4th (likely to close out a game that we are winning).

Conversions

USC had a 3rd down conversion rate of 29%.    Remove the 3rd down dive that Kesser took to center the ball for the game winning FG and we’re looking at 31%.  This is slightly under, but statistically in line with, Stanford’s average 3rd down defense this season which is 33%.

This becomes more interesting when we break up the 3rd downs based on distance.  USC was 0 for 2 on 3rd and short (three yards or less) and also 0 for 2 on 3rd and medium (four to six yards).  All of USC’s 3rd down conversions came on plays with seven or more yards to go.  In this category, USC was 4 for 10.  Remember that USC exclusively threw the ball on 3rd downs.  This hurt the ability to convert the short yardage situations consistently.  I don’t blame Helton for making this call as USC averaged only 0.9 yards per carry  in this game (remove the sacks and it increases to 1.6 yards per carry, which still isn’t great).  It is also worthwhile to note that 40% of USC’s runs were stopped for either no yards gained or negative yards.

On the flip side, Stanford had a 33% 3rd down conversion rate.  This included converting only 1 out of 3 on 3rd and short situations.  You would expect that Stanford, with their power run game, would dominate the 3rd and short conversions.  This is a testament to the strength of our defensive front as well as the aggressive pressure that DC Pendergast put (USC was often loading the line of scrimmage with up to six guys who were backed up by two or three linebackers.  This left single coverage on the receivers and only a single high safety up top).  More on this later.

Stanford was the most successful on 3rd and medium, with 2 out of 3 successful conversions.  This can be a funny distance to defend as there is the threat of both the run and the pass.  However, when USC was able to back Stanford up to 3rd and long, Stanford was 1 for 6 (17%).  USC was easily able to defend the clear passing situations by Stanford.

Other Notes

First let’s look at USC’s average distance to go versus average gain based on down.  [For this chart, I only included the 4th down attempt for 4th downs.  If you consider all 4th downs, the average distance to go was 7.78 yards.]

2013_USC_vs_Stanford_SCAvgDistVsGainThe interesting thing to note here is how much lower 2nd down is in terms of average gain.  I don’t really know why this is the case, but it was interesting to see.

For comparison, here is Stanford’s chart.

2013_USC_vs_Stanford_StanAvgDistVsGainStanford had a great 2nd down push, but had a poor 3rd down push.  In fact, Stanford converted more 2nd downs (35%) than they did 3rd downs (33%).  If I were to guess, this is likely due to how USC played on defense.  So much emphasis was put on bottling up the power run game that runs would either be stopped for very little gain or would break through to the second level and gain chunks of yards (enough to convert).  In the former case, the stop would potentially leave Stanford in a 3rd and long situation that USC would excel at defending.  To illustrate this, let us look at Stanford’s rushing histogram.

2013_USC_vs_Stanford_StanfordRushingHistogram

As you can see here, Stanford had a lot of runs clumped in the 0-4 yards gained range.  However, there is a definite dip  in the 5-8 yard range before another grouping between 9-12 yards.  Finally, there were five runs which broke for 15+ yards.  I believe USC’s gameplan was built around a front line defense which would limit Stanford’s runs to around three yards with not much second line to help until they gained quite a bit.

There was one final statistic that I found quite interesting when breaking down this game.  7 out of 12 (58%) drives for Stanford failed to gain a single first down. (6 out of 11, or 55%, if you don’t count the final desperation play of the game).  This includes 5 out of 7 (71%) of their 2nd half drives (67% if you don’t count the last play drive).  Other than the initial TD drive to start the 2nd half, Pendergast adjusted well to the Stanford attack.  Here is a line chart showing the yards gained on each Stanford drive.

2013_USC_vs_Stanford_StanfordDrivesNotice the downward trend line as the game progressed.  It is also worthwhile to mention that drive #9, the final spike for Stanford’s offense, ended in the interception by Bailey.

Conclusions

This game was definitely an interesting one to look at statistically.  You could see how OC Helton adjusted his gameplan to try to mitigate Stanford’s defensive strengths by going to the air (Stanford is ranked #3 in rushing defense, but #98 in passing defense).  You can also see how USC’s aggressive gameplan to stop the run at the line caused a decent amount of disruption, leading to 3rd and longs.  Both of these gameplans paid off in the end as they helped limit Stanford and ultimately win the game.

ASU – A look back

This post will be more of a stream of consciousness and notes that I took in a quick rewatching of the USC at ASU game this past weekend.  I won’t have in detailed analysis of specific plays, but look at some overall trends and strategy of the game.

As all of you know by now, USC got beat up bad enough by ASU to cause AD Pat Haden to fire HC Lane Kiffin.  I felt that Kiffin called a good enough game on offense (41 points, 542 yards total offense), but a lot of things went wrong to cause what happened last Saturday.  We’ll take a quick look at the offensive side of the ball, but most of this post will look at the defensive side.

Offense

Some quick notes on the offensive side of the ball.  Again, USC was able to run the ball effectively.  37 rush attempts for 247 yards (6.7 yards per carry).  Take out the sacks, and we’re looking at 33 rush attempts for 265 yards (8.0 yards per carry).

Unfortunately, I believe our entire gameplan was predicated on the running game.  Given USC’s success running the ball early in the season as well as ASU’s porous run defense, it made sense as a gameplan.  However, once the game started swinging in ASU’s direction, the running game was not going to be enough to jolt life back into our offense and get us in a position to really win the game.  Add on top of this the fact that we are not deep, especially in the wide receiver position, and you start to see that our offense is not built to come back from any significant margin.

As the game went on (and the score margin grew), ASU was less occupied with defending the run even though we were continuing to run the ball.  Play action, which was effective early in the game, began to dwindle.

Pass protection again posed a problem.  Generally, our pass protection holes came on the edges at both tackle spots.  This lead to Kessler being rushed often and also sustaining four sacks.  I saw significant missed blocks from the Right Guard, Right Tackle, an edge rush that the Running Back did not properly pick up, a play where both the tackles missed their blocks causing Kessler to be pinched, the Right Tackle, and one play where both the Left Guard and Tackle botched.

Defense

ASU’s offense was able to effectively change up their point of attack quickly throughout the game.  Early in the game, they attacked with their tight ends in the seams.

Click here to see three plays that attacked the TEs

The first two plays happened on the same drive.  The last play happened early in the 3rd quarter.

On the second and third plays, ASU was able to catch USC in a cover-2 shell.  The last play in particular happened almost exactly like what USC did against Utah State the week before (which I analyzed here).  In both these situations, our safeties are covering half of the field and are preoccupied with other wide receivers on the edges.  Once the TE breaks past the middle short zone, it is wide open right up the middle seam.

Once they established attacking the middle of the field, they attacked the sidelines pretty hard on fades, fade-stops, and wheel routes.

Click here to see the sideline plays

Notice that the plays weren’t poorly defended, as in there were corners right there with the receivers on most plays.  The majority of these plays came against man-to-man coverage.  Also, the majority of these plays came with our cornerbacks playing press coverage on the receiver.  There was no safety support on these plays and the defenders played with inside leverage over the top (makes sense when you have the sideline to help as an “extra defender”).  Overall, good throws and catches by their QB and receivers, as these sideline passes can be somewhat difficult to execute because of reduced margin of error due to the sidelines.

As the game progressed, USC’s defense shifted to drop more defenders into pass coverage.  We started the game often rushing five, as is normal in a Pendergast defense.  However, late in the game we started only rushing four and dropping seven into pass coverage.  ASU responded to this by shifting their point of attack to perimeter runs.  There were many fly sweeps and QB keepers to the edge in the 3rd and 4th quarters.  By this point, our defensive front was pretty gassed and these runs were able to get to the edges easily.  This caused our pass rush to diminish significantly as our front line was more preoccupied with establishing contain rather than pass rushing.  ASU also utilized QB rollouts, swing passes, and TEs releasing into the flats effectively to keep the QB pressure down in the 2nd half, which really helped them open things up.

One final note on the defense is that they uncharacteristically were making arm tackles often throughout the game.  This lead to a lot of missed tackles early in the game which helped ASU break through for larger gains.  Up to this point, USC’s defense had fairly strong fundamentals in tackling and would avoid the arm tackles.  I had often credited this to increased tackling in practice.  However, old habits die hard and for whatever reason, our defense was not wrapping up properly against ASU all game long.

A Look Back at Hawaii

The first game (and win) is in the books for the 2013 USC football season.  It sure wasn’t pretty, but it wasn’t ugly either.  The team was plagued with inconsistencies, but they didn’t leave too many big negatives and had a couple big positives.  Overall, there’s a lot to improve on, but this team definitely has potential.

This post will look at a few aspects of the game.  First, we will look at the offensive side of the ball.  We will start with the unfortunate safety which occurred early in the 2nd quarter then we will take a brief look at the running game.  Then we will switch over to the defensive side of the ball by looking at the blitzing numbers and the sacks.

The Safety

Early in the 2nd quarter, Hawaii punted the ball and it rolled to the 1 yard line.  Kiffin dialed up a passing play for the first play of the drive, figuring Hawaii would send heavy run pressure up the middle (they did with six rushers and a spy).  Note that by this point in the game, USC hadn’t gotten almost any running game going.  Hawaii was winning in the trenches and getting a strong push up front, leading to USC only having 24 yards on 10 runs (2.4 yards per carry).  Not exactly conducive to a high success rate running the ball with your back to the endzone, so I can understand why Kiffin chose to go with a pass.  On the flip side, the pass wasn’t exactly firing on all cylinders either.  At this point in the game, Kessler was 4/8 for 39 yards (4.9 yards per pass attempt).

View the play here

So let’s take a closer look at the play and how it was designed.

The play design

The play design

As you can see here, the play is pretty simple.  Agholor is on the left (bottom of the screen) and is running a slant route.  The slot receiver (who I believe is actually a TE spread out wide) is also running a slant, although at a greater depth than Agholor.  Marquise Lee is running a go route down the sidelines and appears to be a decoy.  The TE on the left quick blocks the outside linebacker and then transitions into the flats as the checkdown.

The primary receiver in this play is Agholor.  His slant route should be a 3-step drop and get rid of the ball.  Here we are at three steps.

The play after the 3-step drop.

The play after the 3-step drop

Kessler is looking at his primary read, Agholor, who has just broken into his slant pattern.  Kessler would first notice the depth at which the corner is playing off from Agholor, which is a perfect for the slant.  However, Kessler also sees the linebacker spy, who is moving to defend the checkdown TE in the flats.  This linebacker poses a problem as he is moving into the passing lane and could easily pick off the slant route.  First positive here is that Kessler doesn’t force the throw after he identifies that his primary read is not safe.  First mistake, rather than moving through his progression, Kessler hesitates.  His hesitation wasn’t long in real time for how much processing must be happening in his head during his first game ever (and on the road).  I timed it at just under a second too long.  However, a full second is an eternity for a quarterback in football.

After a split second, Kessler moves on to his second read

After a second, Kessler moves on to his second read

Kessler waits for that second, possibly waiting for Agholor to get past the linebacker or just thinking too much about what comes next.  Finally, Kessler turns to his second read.  The second read is the slot receiver, also on a slant.  With the assumption that the defense will be playing the run, you can expect the middle of the field to vacate as the linebackers blitz in to plug the running lanes.  This usually means that one of the two slant routes should be open.  In this case, the slot receiver is defended in press coverage, but was one-on-one with pretty good body position on the defender.  This is probably a safe enough throw if you throw directly at the receiver’s numbers or even slightly behind him (throw ahead of him and the defender might be able to get a handle on it).  With the amount of contact being made, you might even get a pass interference call.

However, none of this matters.  With that second of hesitation before moving to his second progression, the pressure was already in his face.  Hawaii got a great push up front right at the snap of the ball and Kessler hesitated just a hair too long, which combined creates a sack for a safety.

Running Lanes

USC’s run game was bottled up early in the game, but started to open up in the second half.  Personally, I think the run game opened up as the Hawaii defense got tired.  I believe that it was one of those games where the underdog team in week one is amped up when they are going stride for stride, but fails to pace themselves properly.

Another major impact on our running game (and run game play calling) was the fact that all our right guards were banged up, and thus playing injured.  This leads the coaching staff to attempt to run away from this position, which makes our running game a bit more predictable.  Let’s take a look at the histogram for what gaps we ran the ball to.

The number of times we called running plays to each gap

The number of times we called running plays to each gap

As you can see here, 24 runs to the left side and 18 runs to the right, or 33% more left runs than right.  Overall, both sides produced nearly equally.  The left side of the line averaged 4.9 yards per carry while the right side averaged 5.6 yards per carry.  Let’s break it down further.

There were more runs to the left side going to the interior (to the A or B gaps).  14 runs went to the left interior gaps compared to 9 on the right.  The left interior runs also averaged 5.4 yards per carry compared to only 1.7 yards per carry on the right.  Looking at only the A gaps, there were 12 runs to the left A gap for 3.75 yards per carry compared to 7 runs to the right A gap for 0.9 yards per carry.  This seems to suggest that the right guards playing thru injury had a decent impact on our interior running game and play calling.  Expect this to improve as our right guards heal up.

The number of perimeter runs (to the C or D gaps) was about equal between the left and right sides.  However, the perimeter runs to the right side were more effective, averaging 9.4 yards per carry on the right compared to 3.5 yards per carry on the left.  Chalk that up to having a Senior in Kevin Graf manning the right tackle spot compared to Redshirt Freshman Chad Wheeler on the left.

Defensive Blitzing

Overall the defense played really well.  We held Hawaii to 231 total yards and only 23 rushing yards, which is 0.7 yards per carry  (although if you remove sacks from rushing, you’re looking at 2.2 yards per carry).  On the passing game side, you’re looking at seven sacks and four interceptions.  Wow.  The defensive front dominated the game.  The secondary had some weak spots and were out of position on a couple plays.  Luckily, Hawaii was not able to capitalize.

Anyways, let’s take a look at some numbers.  Pendergast was his normal aggressive self.  I tracked that Pendergast blitzed at least one extra rusher on 61% of plays.  Typically, Pendergast would send in five rushers for a +1 blitz.

2013USCatHawaiiBlitzDistribution

It generally worked as well.  When not blitzing, USC’s defense allowed an average of 4.1 yards per play.  When USC blitzed, that dropped down to 2.4 yards allowed per play.  Hawaii’s offense definitely could not handle the extra pressure.

The blitzes came from different directions as well, adding to the confusion.

2013USCatHawaiiBlitzOrigin

Pendergast would typically blitz the linebackers.  However, he wasn’t against sending in the cornerbacks or safeties on blitzes as well.

One interesting thing to note is that all of our sacks came on plays with a blitz.  I have done quick diagrams of all the sack plays and the blitz pressure that we brought here.  The yellow arrows indicated that the players that broke free from blocks and played a key role in getting the sack.

Sack 1
Sack 2
Sack 3
Sack 4
Sack 5
Sack 6
Sack 7

Most of the sacks were +1 blitzes with the additional rushers coming from the linebacker spots.

One interesting note is that three of the four interceptions came on plays in which USC did not blitz.  I would have expected the interceptions to come off the pressure of a blitz, but this was not the case.

Conclusions

On the offensive side of the ball, I expect our run game to improve a good amount when we get healthy.  This includes getting Silas Redd back as well as healing up at the right guard spot.  Once the run game picks up, I expect the passing game to improve (like it did in the second half against Hawaii).

On the defensive side of the ball, expect us to continue to play aggressively and to have aggressive play calling.  I don’t expect quite as much blitzing in our upcoming game against Washington State due to their offensive play style,  but I am hoping for a number of interceptions.

What happened to the O-Line against Stanford?

As you all know, USC lost to Stanford 21-14.  It was a frustrating game for all USC fans as we watched the Trojans get pushed around the entire second half.  So what happened?  What went wrong?  The main thing that went wrong stemmed from the offensive line and that spread through all portions of USC’s game.

The simple out is to put the blame on Cyrus Hobbi as he is the replacement for injured starter Khaled Holmes, but upon further review this is not the case.  I went ahead and rewatched the game (not something I suggest unless you’re truly in the mood to be upset) so that I could take a very close look at what happened on the offensive line.

Strategy in the Trenches

First let us talk about what USC did to protect the quarterback compared to what Stanford did to get pressure.  USC averaged keeping 6.1 blockers in, usually the five linemen as well as keeping the running back in for pass protection.  USC went as high as keeping eight men in for pass protection leaving only two receivers to run routes.  Stanford, on the other hand, averaged 4.6 rushers.  Stanford would most often bring 5 rushers in, meaning two of the front seven would back off into pass coverage.

USC tried at times to protect the quarterback by lining up in the Shotgun.  Of the 76 offensive formations that I tracked (some of them stopped by timeouts or penalties so won’t show up on the official play counts), 20 of them had USC in the shotgun formation.  This is 26% of USC’s offensive formations.  USC had success in gaining yardage in the Shotgun with 7.06 yards per play.  However, while in the Shotgun, Barkley was sacked once, hit after the throw three times, and hurried five times.  This means Barkley was hit or hurried on 45% of our Shotgun plays compared to 23% of the plays when we were under center.

USC also attempted spreading wide as well as playing tight.  While USC definitely preferred to have the extra potential blockers with TE(s) and/or a FB, they did go with four or five WRs 13 times in the game. [Updated Note: As a forum poster pointed out, when I say 4-5 WR, I actually mean we lined up with 4-5 wide, meaning they line up wide to stretch the defense horizontally.  I do not mean that we sent out 4-5 wide receivers.  Often, I assume it would be 3 WR, TE, RB with maybe switching out a WR for another TE, RB, or maybe a FB.  The key here is that USC attempted to stretch Stanford’s defense out horizontally].  Again, USC had greater success gaining yards by spreading the field out with an average gain of 8.9 yards when using 4+ wide receivers compared to 3.3 yard average when not.  However, they also had much more pressure on the quarterback when spreading the field out.  Barkley was hit or hurried on 46% of plays in which USC was spread out wide compared to 27% of plays when we were not.

USC also attempted to spread the ball out quickly by running a couple bubble screens.  Unfortunately, since Stanford was able to get so much pressure in on Barkley without committing too many rushers, the defensive backs could simply play close in on the receivers and easily defend any quick route (such as the slant, bubble screen, etc).  This resulted in two of USC’s four bubble screens being incomplete and another being stopped for a loss of two yards.  The last bubble screen was successful in gaining 23 yards.

Many fans complain that USC did not run any running back screens, play action, or rollouts/bootlegs to try to mitigate the pressure.  Unfortunately, none of these tactics would have likely worked.  When your offensive line is being dominated across the board even when they are blocking, a running back screen becomes very obvious to the defensive linemen.  All the defense needs is for one defender to pick up on this screen and defend it and the whole play falls apart very quickly (high percentage of interceptions or sacks if the screen is defended).  USC did attempt some play action, but it was widely ineffective due to our running game never getting off the ground.  Rollouts and bootlegs, or getting Barkley out of the pocket and away from the pressure, would not work because there was not a single point of pressure to roll away from.  Furthermore, Stanford began spreading their outside LBs very wide (almost to the point that they were at our receivers) to mitigate any type of QB rollout.  With their LBs that wide, there is no way to wall off the pressure on a rollout to allow Barkley to get away.  They are also so wide that they would not even think about defending the run on a bootleg and that would just get Barkley killed.

Individual Mistakes

Now let us turn our attention to the offensive line and how they played.  I tracked who lost their individual battles or missed assignments.  This could include things such as being driven back on a block, missing an assignment, or not picking up the blitz.  Below you’ll see a chart with mistakes I found which shows the total number of mistakes (in blue) and then broken down by pass play (red) or run play (green).  Please keep in mind that these “mistakes” are simply my opinion since it is somewhat subjective without knowing the actual assignments and play call.  Furthermore, some of these mistakes could be marked as one player missing an assignment but in actuality it might be the center’s mistake for not switching the line protection call in response to a blitz look.

By my count, the Right Tackle position missed the most assignments or blocks followed by the center and the left tackle.  However, across the board there were mistakes made by linemen in both the passing and run game.  Some of these mistakes were fundamentals or execution errors (such as not getting leverage enough to hold the block), but a good portion of these mistakes were communication or mental errors (not picking up the blitz).

The offensive line played so poorly in this game that Barkley even got nailed very hard on a toss running play (the second touchdown that USC scored).  When your quarterback is facing pressure on running plays, that is never a good sign.

Stanford’s Gameplan

The first half of the game saw Stanford running a number of ways to create confusion on our offensive line based on who was coming in.  Since Stanford runs a 3-4 defensive scheme, the standard 4th pass rusher can come from any number of the linebacker spots which makes it more complex to block.  Sometimes they would bring in an extra edge rusher while others they put send someone through the gut.  See examples of various 4-man pressures that the 3-4 allows below.

Zone blocking concepts should take care of this as you block what defenders come into your assigned zone rather than trying to account for every potential rusher.  However, our offensive line did not always seem to pick up their assignments.  Some tricky fake blitzes were also run in the first half to throw off our zone blocking.  Stanford would begin the play rushing six, but two of them would initiate contact with offensive linemen then break contact and drop into pass coverage.  This caused linemen to initiate blocks against non-rushers, while other areas had noone left to block an actual rusher.  These fake blitzes greatly confused USC’s offensive line and left a lot of free rushers in against Barkley.  In the diagram below, the green arrows represent fake blitzers.  See how they disrupt the offensive line’s assignments by forcing blockers to engage, then overload one side once the blockers have already engaged in their blocks.  The result is two players who have noone left to block and one free rusher shooting through a gap.

Moving into the second half, Stanford adjusted their attack.  Rather than using numbers and fake blitzes to mask where the pressure would come from, they started using twists.  See the example of a defensive line twist below in the green arrows.

Zone blocking schemes are supposed to mitigate twists and stunts by the defense, as you are responsible for an area rather than a man so you simply block whomever comes into your area regardless if they twist or not.  However, USC’s line again looked lost and did not pick up these moves well.  In fact, on USC’s final drive, I counted four twist moves done by Stanford, none of which were picked up properly by the offensive line.  These four plays resulted in two plays in which Barkley was sacked, one play Barkley was hit after the throw, and one hurry.

In total, I counted 10 plays in which Stanford ran either a fake blitz or a twist.  Only two of them were picked up properly by USC’s front and one was non-effective for Stanford due to USC running a WR screen play.  The rest resulted in two sacks, two hits after the throw, and two plays in which he was hurried.

How it affected USC’s Defense

I felt like USC’s defense actually played pretty well.  In the first half, there was the one botched play where we allowed Taylor to go for 59 yards, but that was pretty much it.  This single play accounted for 33% of Stanford’s production in the first half.

The second half was not so kind as the defense allowed 53 more yards in the second half than they did the first half.  Much of this was due to the defense getting tired.  This is what teams such as Stanford and Oregon do to you (although these teams get there in different manners).

Look at the time of possession discrepancy between the halves.  In the first half, USC controlled the clock with 17:08 time of possession against Stanford’s 12:52.  This comes to an average drive length of 2:08 for USC and 1:36 for Stanford.  Even if you ignore the turnover drives, USC held Stanford to under one minute drives twice and almost a third time (drive was 1:07).  USC’s shortest non-turnover drive was 1:06 followed by the next shortest drive which was 2:07.  Move into the second half and these numbers flip.  Stanford had the ball a total of 19:19 in the second half compared to USC’s 10:41, or almost a two to one ratio in favor of Stanford.  This is time that our defense is out there getting tired and not having enough rest when the offense is out on the field.  The average drive for Stanford’s 2nd half lasted 3:13 compared to USC’s 1:46.  Ignoring Stanford’s endgame where they only took a knee, the shortest drive Stanford had was 2:05.  Comparatively, USC had three drives under a minute long.

Now let us look at the number of plays and yards between the two halves.  In the first half, USC averaged 4.6 plays per drive for 26.25 yards.  In the second half, USC had 5.2 plays per drive but tanked to 13.0 yards per drive.  This included four out of six drives being 3 and outs.  For Stanford, the first half had only 3.75 plays per drive for 22.1 yards on average.  This included four out of their eight drives being 3 and outs.  However, in the second half, Stanford averaged 6.5 plays per drive for 39.8 yards.  If you ignore their endgame kneel, this becomes 7.6 plays per drive for 48.2 yards on average.

Other Notable Mistakes

I felt like one issue that is definitely a consideration (but don’t have the means to test) is the speed at which Hobbi hikes the ball compared to Holmes.  If a hike is slower, it takes the quarterback that much longer to get into his dropback and get ready for his throw.  For shotgun, the velocity and placement of the ball can greatly affect how long the quarterback spends prior to actually looking downfield.  There were a couple of notable times that you can see on the replays that the hike would come out offset, even when under center.  Barkley would have to adjust, greatly slowing him down.  One example is the touchdown run where Barkley pretty much took the hike between his hand and knee before being able to hand it off.

Another issue I noticed were uncalled false starts.  There were at least two occasions where I saw a false start by LT Walker that were not called.  Add one that was called on him late in the game as well which shows how antsy he was in his stance against the rush he knew was coming.  To be fair, the RT also false started on this same play that Walker got called on, but the official chart will show this on Walker.

Finally, it was fairly evident that Stanford had our number in regards to our snap count.  Because of this, Stanford often got the jump on our line and knew exactly how to time their blitzes to get the maximum momentum and leverage.

Conclusions

Football is a game that is won and lost in the trenches.  When a team gets absolutely dominated up front, there is not much you can do to stop it.  Your deep routes cannot develop.  Your receivers start getting pressed and bumped off the line since the corners know they only need to prevent getting burned for a couple seconds before the quarterback has to get rid of the ball.  Running plays won’t work when the offensive line can’t get a push which negates all possibility of play action.  The list goes on and on.

Fortunately, Stanford’s front seven will most likely be the toughest we will face all season long.  I do not expect our offensive line to potentially be dominated in this manner until we either have a rematch with Stanford in the conference championship game or our bowl game (depending on the opponent).  Sure, our offensive line may be tested against other opponents that have strong fronts, but I believe none are to the level that Stanford is at.  There should also be time to fix this issue and hopefully get our injured players back quickly.  Either way, we should know a lot more about where this team is after this weekend’s game against Cal.

USC vs Hawaii Offensive Play Calling

USC did not establish offensive balance in their season opener against Hawaii.  USC called 42 passing plays against only 22 rushing plays (I counted Wittek’s sack as a pass play).  This means we were passing 66% of the time.  If you only count the times the starters were in, this becomes 38 passes compared to 14 runs, or 73% passing.

Why did USC pass so much?  You will hear a number of answers if you ask around.  For example, USC didn’t want to show their hand in an opener that was pretty much wrapped up after the 1st quarter.  That is a fair enough assessment and the lack of typical USC running plays such as the Power might be shown as evidence of this.  Lane Kiffin said after the game that we didn’t run the ball because of the aggressive fronts that Hawaii was showing.  Fans often translate this to: Hawaii put “eight men in the box.”  This post will look more into that phrase as well as Kiffin’s statement and delve into why USC may not have run the ball.

“Eight Men in the Box”

The phrase “eight men in the box” works on a conceptual level because in a typical defense you will have some combination of seven front-line defenders (defensive linemen and linebackers), be it a 4-3, 3-4, 5-2, or any other type of base defense you might see.  This implies seven men “in the box” is standard.  Thus, adding an extra run defender, usually by bringing down a safety, to make it eight in the box has to mean that the defense is expecting a run, right?

The answer to that is that it depends on the formation that the offense is in.  The defensive alignment will depend on how the offense is aligned.  For example,  if the offense breaks the huddle and goes four wide Shotgun, it is doubtful that the defense will put the standard seven in the box, nor the eight if they are expecting a run.  They will likely have five in the box, six if they are expecting a run.  Conversely, if the offense comes out in a power I formation, the defense is likely to align with eight in the box, nine if they are expecting the run.

The reason for this is because football is a numbers game, especially when running the ball.  Assuming all things equal, if the offense has an equal amount or more blockers than defenders, they will win the battle more often than not, and vice-versa.  This means that the defense should always have at least one more defender to stop the run than there are blockers with the unblocked defender making the tackle.

Another way to look at it is that there should be a defender responsible for every running gap.  Here is an example of how gaps are designated for a typical I-Formation.  Coaches typically letter the gaps from the inside out.

There will be a gap between every blocker and then gaps on the edges.  Running backs will always run through a gap.  Lead blockers, such as a fullback as shown above, create what I like to think of as a mobile gap.  The fullback will run through one of the existing gaps, thus splitting that gap and creating two, one on either side of him.  The defense must account for this.  To cover every possible gap, you will need one more defender than there are offensive blockers to successfully defend the run.

Of course, not all run defenders need to line up in the box.  You can have a corner cover the edge gap whose responsibility it is to force the runner back inside.  A safety could come down in run support after reading the run play and fill one of the gaps.  A defense could also try to win their leverage battles and beat the blocks rather than overload the blockers with numbers.  However, Hawaii likely wasn’t expecting to win the leverage battle against USC as they were physically outmatched pretty much across the board.  If they were to stop the run, they would have to do it with numbers.

In a way, USC forced Hawaii’s hand defensively in regards to how many players that would align in the box.  This happened because of the offensive alignment.  USC often lined up in running formations, such as the I-formation with a fullback and one tight end or the Ace formation with two tight ends, as shown below.

Both these formations have seven potential blockers (five linemen plus the tight end(s) and/or fullback).  If our theory that Hawaii was likely not expecting to be able to to win their leverage battles is correct, they should have brought in an extra run defender, meaning eight in the box.

I decided to run with this idea and rewatch the first half of the game (time restrictions limited me to the first half only) and see how many defenders Hawaii put in the box compared to the formations that USC lined up in.  Let us look at what I found.

Hawaii’s Gameplan

First let us talk about USC’s formations and the statistics around them.  There were 42 offensive plays that I tracked.  USC lined up in a standard I-formation (fullback and one tight end) 10 times.  USC lined up in Ace formation with two tight ends 10 times.  USC lined up in a power run formation (fullback and two tight ends or single back with three tight ends) four times.  See the screen captures below for examples.  Combine these together and it means that out of 42 offensive plays in the first half, 24 of them (57%) were in a fairly strong running formation with at least seven potential run blockers giving a strong run threat.

On average, USC lined up with 6.7 potential blockers.  Hawaii responded to this by lining up with in-kind numbers.  On average, I saw 6.9 players in the box.  This brings into question why USC did not run the ball more.  Hawaii lined up with in-kind numbers and USC has the physique advantage, it should have been a field day on the ground.  Instead, USC ran a lot of isolation on the edges with Marquise Lee and Robert Woods in single coverage through the air.

One likely explanation is safety alignment and play.  While the safeties weren’t coming down into the box at the snap, it is likely that they were crashing down on the run after the snap and would play run support as priority over passing support.  This is conjecture on my part as I did not have the time to look for this specifically.  I do think it is a likely case though, given the amount of single coverage and lack of safety support we saw on our wide receivers, especially on deep routes.

However, one thing I did observe in rewatching the first half was the sheer amount of blitzing that Hawaii was doing against USC.  By my count, Hawaii blitzed at least one extra player on 30 out of the 40 plays I tracked in the first half.  That is 75% of plays with a blitz.  11 plays (28% of the plays I tracked) blitzed one extra player, 14 plays (35%) blitzed two extra players, and five plays (13%) blitzed three or more extra players.  My guess is that Hawaii was hoping not only to stop the run by always sending players into the backfield, but to also attempt to overwhelm our line (i.e. prevent us from providing double team help for our new left tackle Aundrey Walker).  This also puts pressure onto Silas Redd and Soma Vainuku, both on their first games at USC, in their pass blocking.  With this in mind, it is a strong positive to see that Barkley was not sacked during this game.  Although he was hit after the throw a couple of times and had to step up to evade a sack at least once, Barkley was not using hot reads and seemed to have time to sit in the pocket for the deep passes to develop.  This means that the protection held despite the heavy handed blitzing Hawaii was doing.

It was this aggressive blitzing that I believe Kiffin was talking about when he said USC didn’t run because of Hawaii’s aggressive schemes.  With pass protection being adequate enough to give Barkley time to execute, it only made sense in Week 1 to stick with what works and not do anything too fancy (aka, not give too much away).  Quite simply, there was no need to run.

Conclusions

I am still a little worried about the running game.  However, it was Week 1, and you should expect to see problem areas in this early in the season.  I expect to see a much stronger ground game against Syracuse, regardless if they stack the box or blitz. I doubt many other teams will play so aggressively against USC’s run because of the playmakers we have on the outside.  Playing this aggressively also leaves you susceptible to play action (which I did not feel USC did a ton of in Week 1, but could be inaccurate upon further review).  I want to see a stronger run game because I want to know that our offense can execute the plays before we hit the bulk of our season.  There were mistakes and inconsistencies against Hawaii, but they say that teams improve the most between Week 1 and Week 2.  I expect that to be true this year as well with marked improvements in not only the running game, but also in the passing game as well.

USC vs Washington Defensive Recap

This post will look at the defensive side of the ball for USC.  The post will first look at the play calling by Washington throughout the game.  Then we will look at the conversion rates allowed by USC, the field position battle, and the yards allowed per quarter.

Play Calling

Overall, Washington relied on the pass over the run.  Washington ran the ball only 17 times this game (30%) and passed 40 times (70%).  This is odd since I would have assumed Sark would keep a fairly balanced offense.  I would have especially expected them to rely on the rushing attack of Chris Polk to establish the run game.  However, this does not seem to be the case.  First let us look at the play calling by quarter to see how the play calling evolved as USC grew their lead.

As can be seen here, Washington never really established the run game, even early in the game when it was close.  While USC was still within a touchdown score to tie, Washington passed the ball over twice as often as they ran with only six runs (29%) and 15 passes (71%).  Washington kept this percentage fairly consistent regardless of the score margin.

Now let us look at the play calling by down.

As can be seen here, the largest threat of the run was on first down.  There was a semblance of balance in this down for Washington.  However, once Washington was on its later downs, they relied heavily on the pass.  Washington’s only run on 3rd down came on a 3rd and 10.

Conversion Rates

USC had a phenomenal 3rd down conversion rate defense this game.  In fact, Washington did not have a 3rd down conversion until the 4th quarter.  Part of this was due to very strong defense on early downs by USC.  Washington was in 3rd and long for every single 3rd down until the 4th quarter, when they finally had a trio of 3rd and mediums.  Even still, Washington was only able to convert one out of three 3rd and mediums (33%) and one out of ten 3rd and longs (10%).

Let us look at the average distance to go compared to the average gain per down.

USC held fairly strong on early downs, but the key number here is that USC allowed an average of only 1.3 yards on 3rd down.  All this while Washington had an average distance to go of 10.1 yards on 3rd down.  USC was able to achieve this low average gain on 3rd down due to the sheer number of sacks USC had, especially on 3rd down.  Of the seven (yes, seven) sacks USC had in the game, four of them came on 3rd down.

Field Position

This is a real quick look at the field position battle, which encompasses all phases of the game.  Washington had an average starting field position of their own 27 yard line.  They only took 25% of their offensive snaps on the USC side of the field.  Compare this to USC who averaged starting on their own 31 yard line, but took 39% of their offensive snaps on the Washington side of the field.  USC also forced six out of 12 Washington drives to end without converting a single first down.

Yards per Quarter

Let us take a real quick look at the yards per quarter allowed by the USC defense.

As can be seen here, USC’s best quarter was the second, where it allowed only 32 yards of offense.  Part of this is due to the negative yardage drive that USC’s defense was able to force which resulted in a safety.

USC’s defense allowed the most yardage in the 4th quarter where Washington scored a very late touchdown.  However, this may have been due to USC bringing in the backups for mop up duty.

Conclusions

USC’s defense played well in this game.  They were able to prevent Washington from converting first downs, which kept their offense off the field.  This is seen with a nearly 10 minute advantage for USC in time of possession.  They were able to apply a lot of pressure against a team that wanted to establish a passing game and resulted in seven sacks.  This defense played well overall.  We’ll just have to see if that will translate into next week’s road game against a very potent Oregon offense.