Tuesday, August 14, 2007

Evolution of an Offense | by Michael

On Media Day, Notre Dame beat reporters wasted little time in asking Charlie Weis how he planned to replace some essential starters he was losing from the year before. What would he do without Brady Quinn, Jeff Samardzija, Rhema McKnight, Darius Walker, or Marcus Freeman?

Wait a second. Marcus Freeman was a starter?

While Freeman's inclusion in that list may surprise some readers, the backup tight end from St. Paul, Minnesota was technically a starter in three separate personnel groupings, or packages, in the offense. Yep, that second TE slot, while not as glamorous as a starting running back or receiver, is still an essential component of several key packages, and it will have to be filled.

During fall camp, as "starters" are announced and "back-up" players are given roles in each grouping, the offense will begin to discover its identity. Earlier this year, Charlie talked about the process of putting things together:

It all depends how the thing evolves - it really does. It depends how this evolves. You don't know what your run/pass ratio is going to be; therefore, you don't know what your lead personnel grouping is going to be.

I think a lot of times you have to figure out what your personnel groups and what your personnel can do, and then it might get heavy one way or another ... things that, right now, the jury is still out.
And it's not just as simple as picking a #1 running back, or the #1-#3 receivers. Packages, above all, are situational; Armando Allen might be your starting running back for a 4-Wides deployment, while Robert Hughes might find himself starting in the Goal Line package. Just this afternoon, coach Rob Ianello was asked about setting a "rotation" for receivers, and in his answer he rejected the premise. Instead of a simple pecking order, Ianello said, "I believe you find-- you know, Coach believes you find roles for guys in personnel groupings."

While the players practice, and the staff deliberates, and Charlie starts forming up his offense over the next couple of weeks, there's no better time to take a look at the different packages the Irish have used over the past two years. For each package we'll show you a common formation, break down some of the stats, and talk about its effectiveness. We'll also opine on how that package might be used in the '07 campaign.

To compile this post, we went back and charted each game of the '05 and '06 seasons, identifying the packages and recording the outcomes of each play. The terminology used in this post is taken from a publicly-available 2005 Notre Dame spring playbook, which Charlie had adapted from the 2004 Patriots. Actual terminology in use by the Irish may have shifted since then.

Let's start by looking at the most commonly-used package by the Irish, Half.


HALF
WR
TE
HB
FB
3
1
1
-

Stats
'05
'06
% used
37%
56%
% run
30%
31%
ypc
5.07
5.95
% pass 70%
69%
comp %
68%
63%
ypa
7.93
7.61

This picture above probably looks familiar, as Half was by far the most popular package used by the Irish last year. Half usually utilized this zero-out formation, although it did appear in other incarnations, such as this bunch set. One look at the data suggests that the Irish were an offensive powerhouse out of Half, one that actually improved from '05 to '06. After all, they averaged a Holtzian 6 yards per carry. But the raw numbers don't tell the whole story.

The Irish used this package more than half the time, whereas it was used only slightly more than a third of the time in '05. Why was it used so often? For starters, Notre Dame was forced to play catch-up against Michigan, Michigan State, LSU, and UCLA, and getting three receivers on the field was crucial. In another game, against Army, Notre Dame ran all 62 plays -- the entire game -- from Half. That was the first time in two years they've used just one package for an entire game. In fact, they kept the gameplan so simple that they essentially used just 8 formations, and nothing terribly exotic.

The finesse running game thrived when the defense was spread out; even the inside zone play, which averaged 3.26 yards all season, improved to 4.5 yards per rush out of Half. That was actually the worst average for any of Notre Dame's running plays from Half, and two other runs, Outside Zone and Jab Sub (a counter), averaged nearly 7 yards per carry. Even if you take out Stanford, Purdue, MSU, and the academies, the running game was still respectable out of Half, as the Irish averaged 5.1 yards per carry.

The passing offenses from '05 and '06 were strikingly similar. The third receiver (Shelton in '05, Grimes in '06) caught about 70-75% of his passes; the tight end (Carlson, Fasano) caught about 80-85% of his passes; the Z receiver (Samardzija) caught 63% of his passes; and the halfback (Walker) caught 75% of his passes. The only difference was at the X receiver position. In '05, Stovall caught 63% of his passes, while McKnight caught just 53% of his passes. Likewise, the yards per pass play were all within a yard of each other except at tight end, where Carlson's big play potential was evident, and at the X receiver position. In '05, Stovall averaged more than 12 yards per pass attempt; McKnight, on the other hand, averaged 7.5 yards per pass attempt. The disparity at the X receiver position shows up in the overall decline of yards per pass attempt and completion percentage for Half. McKnight was simply not as strong a fit for the X receiver as Stovall was.

"Half" in '07:
If Half is used more than 30% of the time in 2007, chances are that Notre Dame will have been playing catch-up in a lot of games, and that will bode ill for the season. If things go well, the percentage will likely hover between 20-25% of the time, based on the expectation of a more conservative power rushing attack. The Irish will still need Half as a situational grouping, but it shouldn't be the lead grouping as it's been the last two years. As for personnel, Grimes may have the #1 spot locked up, but where will he play? He played the F receiver last year, and in the open practice video he lined up as the Z. That should be a safe bet. Which two receivers will win the other two spots?


DETROIT
WR
TE
HB
FB
2
2
1
-

Stats
'05
'06
% used
20.1%
16.7%
% run
43%
39%
ypc
4.60
3.71
% pass 57%
61%
comp %
57%
58%
ypa
6.97
6.05

Expectations for the duo of Freeman and Carlson soared after last year's Blue-Gold game, but that production never materialized on the field. In Detroit, the Irish averaged nearly a yard less per carry and per pass attempt than in '05. Even when Weis attempted a little artistry with Detroit, like splitting the tight ends wide, the results were similar. The casual observer might initially guess that losing Anthony Fasano caused this under-performance, but that was hardly the case. Looking at the stats, Carlson's numbers are comparable to Fasano's, and Freeman's to Carlson's (from '05). Instead, it was the WR production that dramatically declined. First of all, Brady Quinn completed less than 50% of his passes to McKnight (45%) and Samardzija (47%). Contrast that with the '05 output, where Maurice Stovall (63%) and Samardzija (56%) caught a much higher percentage of passes. Stovall averaged a whopping 10.92 yards per pass attempt, whereas McKnight managed just 5.68 yards per attempt last year. Meanwhile, Samardzija's production out of Detroit declined from 8.26 to 4.95 yards per attempt.

Two primary factors explain the drop: the inefficiency of the play action passing game and the loss of Stovall's big play capability. Because the Irish couldn't run the ball as effectively as they ran it in '05, play action passes were not nearly as feared nor respected. Most quarterbacks might be content with Quinn's '06 numbers (13 of 24 for 194 yards, 54% completed, 8.08 yards per pass attempt) but they paled in comparison to his '05 performance (24 of 31 for 284 yards, 77% completed, 9.16 yards per pass attempt). Normally, a quarterback's completion percetnage for play action should be higher than his overall rate, but that was not the case for Quinn last year. Finally, this grouping lacked the ability to make a big play without the benefit of a play action fake: while Stovall had five plays of 20+ yards, McKnight had zero.

"Detroit" in '07: With the talent the Irish have stockpiled at offensive line during recent recruiting classes, expect a stronger rushing attack. Higher per play averages should result across the board, regardless of who wins the starting job opposite Carlson. Play action passes should be as effective a strategy as they were in '05, which should make quarterbacking an easier task for whoever wins the starting spot. Detroit should be a featured package this year.


REGULAR
WR
TE
HB
FB
2
1
1
1

Stats
'05
'06
% used
20%
12%
% run
46%
48%
ypc
4.54
3.72
% pass 54%
52%
comp %
68%
55%
ypa
10.54
5.10

Asaph Schwapp's season-ending injury in the Penn State game likely played a large role in the diminished usage of Regular last year. In fact, it was used 21 times against Georgia Tech, which was about 28% of all offensive plays in that game. However, Weis clearly didn't have the same confidence in backup Ashley McConnell, and Regular pretty much disappeared. By the end of the year, that one game against the Yellowjackets still accounted for one-fifth of all Regular usage.

Several stats should immediately jump out in the decline from '05 to '06: -0.82 yards per carry, -13% decline in completion percentage, and a precipitous -50% drop-off in yards per pass attempt. Further, Regular yielded the worst yards per play average among last year's most commonly-used groupings. With a fullback on the field, the Irish averaged a paltry 3.4 yards per play, as compared to 5.7 yards per play for the entire season.

The biggest issue the Irish had last year in Regular was pass protection. Nearly a quarter of the time Quinn dropped back to pass, he was either sacked or forced to scramble (as compared to just 11% of the time in '05). Then there was the screen game, or more to the point, the complete absence of the screen game. In 2005, the Irish threw seven screen passes out of Regular, completing six for 89 yards. And it was a diverse screen attack, as Rashon Powers-Neal, Fasano, and Walker all caught screens. Consider, too, that more would have been thrown in Powers-Neal's direction had he not been suspended for the season. Last year? No screens were completed using Regular personnel. Worse, both attempted screen plays resulted in a total of -27 yards, as the defense sacked Quinn both times.

Meanwhile, the rushing offense struggled mightily. The best run out of Regular was arguably the draw play, a finesse run that averaged 4.14 yards per attempt. By contrast, the same play averaged 4.4 yards per carry the year before, but was only the fifth best run play, as even FB handoffs (8 attempts, 5.13 yards per carry) gained more yardage. To further illustrate last year's impotence, the Irish didn't even attempt to run an outside zone play with Regular personnel, and they ran only 4 inside zone plays, all against weak fronts (Stanford, Navy, and North Carolina). By contrast, the Irish ran 3 inside zone plays alone against Tennessee's vaunted front in '05.

One last tidbit...although Fasano caught 7 balls for 107 yards in '05, only one pass was even attempted to Carlson last year out of Regular.

"Regular" in '07: Schwapp is back, although he must now contend with Luke Schmidt, who has had a year to soak up the playbook. Schwapp will likely never be the threat that Powers-Neal was in the four games he played in this offense, but his healthy return is still welcome news. Although #44 has struggled at times as a lead blocker, he had great success laying "wham" blocks on Tennessee's defensive line. (The wham block is when an offensive lineman allows one defender to penetrate so that he can get to a linebacker. It's a trade-off; instead of lineman blocking lineman and fullback blocking linebacker, the roles are reversed.) Running wham inside zone plays is what Schwapp does best, and with a bigger, stronger, interior offensive line, more yardage should be achieved on the ground. Pass protection in Regular must improve, but when there's a threat of a FB releasing to the flat, catching a flare, and picking up big yardage because linebackers are blitzing, defensive coordinators are more hesitant to dial up the blitz. That's why the Schwapp-Schmidt battle should be interesting to watch and why it's difficult to predict how often we'll use this package. It might seem easy to stuff blitzes with seven players in protection, but considering Weis's complicated protection schemes, if two's company and three's a crowd, sometimes seven's chaos. At least it was at times last year.


TWO TITES
WR
TE
HB
FB
1
2
1
1

Stats
'05
'06
% used
9.9%
4.6%
% run
70%
67%
ypc
3.35
1.92
% pass 30%
33%
comp %
59%
67%
ypa
11.32
9.50

Although most of the time it's used for running, Two Tites also launched a bunch of big-play passes. A play-action fake freezes the linebackers and causes the safeties to creep up, and during '05 and '06, Samardzija usually found himself open downfield. There were five pass plays of 20+ yards in just 21 attempts in '05, and last year there were 3 big plays out of 12 pass attempts. Even without much of a running game, the Irish hit a big play every 4 passes. Last year teams blanketed Samardzija more, which resulted in Carlson catching two balls for 27 and 25 yards against Michigan State and Purdue, respectively. If the Irish could develop a reliable running game, that big play rate could actually improve. Last year's rushing effort out of Two Tites was an embarrassing 1.92 yards per carry, and that number actually dropped to 0.94 (17 carries, 16 yards) yards per carry against BCS conference bowl teams.

"Two Tites" in '07:
The sky is the limit for this package if a legitimate rushing offense emerges. A stronger, bigger offensive line, power backs, a healthy fullback, and depth at tight end should improve the ground game out of Two Tites, and open up the field for play-action. Whoever replaces Samardzija in Two Tites will require big play capability, because the ball is definitely coming his way.


3 WIDES
WR
TE
HB
FB
3
-
1
1

Stats
'05
'06
% used
0.1%
2.8%
% run
100%
63%
ypc
5.00
6.27
% pass 0%
38%
comp %
-
43%
ypa
-
4.43

Although at first blush it looks like a passing package with the three receivers out there, the Irish ran the ball out of 3 Wides almost two-thirds of the time. The rushing stats look impressive, but most of that yardage came against Stanford's porous front seven. Notre Dame ran 10 times for 69 yards against the Cardinal and 4 times for 15 yards against Purdue, UCLA, and USC. The passing game was a mixed bag. On the one hand, Quinn completed 2 of 3 passes against Purdue and Stanford for 26 yards. On the other hand, he was sacked twice (once by Navy) and completed 1 of 4 passes for 5 yards against the stronger defenses of USC and UCLA. As with "Regular", pass protection problems cropped up with two backs on the field.

3 Wides pulls the tight ends out of the game, but they sometimes still appear in the guise of an H-Back, which is basically a tight end playing like a fullback. Carlson, for instance, llined up as a fullback 19 times during the middle run of last year's schedule. Fasano was used in the same H-back role once against Tennessee in '05 in 3 Wides.

"3 Wides" in '07: If other packages can't jump start the rushing offense, 3 Wides could be a viable option. The Seattle Seahawks have enjoyed a dominant running game with Shaun Alexander running behind a TE-less line; there's no reason why the Irish can't do the same. It all fits into Weis's master plan for running the same offensive plays from different groupings and formations.


NEW YORK
WR
TE
HB
FB
1
3
1
-

Stats
'05
'06
% used
7.1%
2.7%
% run
76%
78%
ypc
3.46
2.28
% pass 24%
22%
comp %
54%
60%
ypa
6.62
8.00

Anyone remember the first play of the Weis era? One receiver, three tight ends, and a halfback - that's New York style. Samardzija, who was the "starting" wide receiver, went in motion, and Quinn received the snap. Play action bootleg!

Unbeknownst to Weis, it was the ultimate irony: predecessors Ty Willingham and Bill Diedrick inherited a mobile quarterback in Carlyle Holiday, whom they forced into the West Coast offense like a square peg into a round hole, and they never, ever, ran a bootleg pass -- even though it appeared to be the simplest and most suitable idea to anyone who followed the Irish. Fast-forward to two years later, though, and only one other bootleg has been run (later in the first half of that Pitt game). In '06, t
he inability of Notre Dame's line to create any push for inside zone runs played a major factor in not utilizing New York more often, as well as the fact that two freshmen tight ends were still learning the ropes of the Weis offense (not to mention, the power back was spending the majority of his snaps at linebacker). Over the last two years, 17 of the 20 pass plays out of New York have been play action passes.

"New York" in '07: Similar to Two Tites, this could be the year we see a lot more of New York. What Weis was lacking last year he now possesses: bigger, stronger interior linemen, multiple power backs, three tight ends, and even extra offensive linemen anxious to play some snaps if that third tight end isn't pushing around defensive linemen like rag-dolls. Like in Two Tites, one question mark exists: who will be the lone wide receiver? Is a bigger receiver preferred in this package for blocking? If so, could this be an area where someone like Duval Kamara gets some early snaps as he learns the offense?


JAX
WR
TE
HB
FB
4
1
-
-

Stats
'05
'06
% used
3.4%
2.2%
% run
10%
5%
ypc
3.33
7.00
% pass 90%
95%
comp %
69%
69%
ypa
4.08
11.13

Just because BYU isn't on the schedule doesn't mean the Irish can't use Jax. Although the grouping is most familiar from its usage against the Cougars in '05, it's also been used when the offense is facing a crucial third down, and four wide receivers and a tight end trot onto the field. Remember the first scoring drive against Michigan in '05? First and goal, wide open slant to McKnight? That was Jax. By and large, the grouping has been highly successful, and the slant has been an integral part of its success. Although the yards per pass attempt average is low for '05, keep in mind that against the blitz-happy Cougars, it was used essentially as a horizontal running game to spread the field as blitzers attacked Quinn. Without anyone in the backfield to pick up blitzes, most passes are usually 3-step drops, and if it's third and short-to-middling, chances are at least one receiver is running a slant. All runs out of Jax are quarterback draws or sneaks.

"Jax" in '07: This will never be more than a situational grouping, and that point is all the more true considering the current talent and depth at running back.


5 WIDES
WR
TE
HB
FB
5
-
-
-

Stats
'05
'06
% used
0%
1.6%
% run
-
14%
ypc
-
4.00
% pass -
86%
comp %
-
73%
ypa
-
9.09

All fourteen plays with 5 Wides were run against USC, when the Irish were without John Carlson. The first time 5 Wides appeared, it was this second and long that resulted in a scramble by Quinn. The next thirteen were run after the Irish were down by 18+ points. It's safe to say that the only reason that Weis chose 5 Wides rather than Jax must have been that he considered Chase Anastasio more of a receiving threat than Marcus Freeman. As far as the runs go out of 5 Wides, there were two: a swing pass behind the line of scrimmage to Samardzija, and a quarterback sneak.

"5 Wides" in '07: Barring injury to a tight end and/or falling way behind in a game -- will this ever be seen again?


4 WIDES
WR
TE
HB
FB
4
-
1
-

Stats
'05
'06
% used
0.1%
0.8%
% run
0%
29%
ypc
-
0.00
% pass 100%
71%
comp %
100%
40%
ypa
33.00
8.00

4 Wides, much like 5 Wides, was used mostly against USC because of the injury to Carlson. Although Walker lined up in the backfield, he did not receive a carry; the two runs were a swing pass behind the line of scrimmage to Samardzija and a sneak by Quinn. The only other time we've seen it was way back in '05 against Michigan State, and it resulted in a 33-yard completion to Matt Shelton.

"4 Wides" in '07: Too many great tight ends currently live in or are on their way to South Bend for this to become anything but a stop-gap measure in times of injury.


GOAL LINE
WR
TE
HB
FB
-
3
1
1

Stats
'05
'06
% used
1.7%
0.2%
% run
87%
100%
ypc
0.92
-1.00
% pass 13%
0%
comp %
0%
-
ypa
0.00
-

There were only two snaps from a traditional Goal Line package last year. The first one was the last play of the Georgia Tech game, when Quinn snuck through the pile for a first down on 4th and 1. The second is shown above, and the result wasn't pretty. Upon the snap, Bob Morton was blown off the ball and Travis Thomas was stuffed for a loss of 3 yards. The previous year wasn't much better: 5 of Notre Dame's 13 rushing attempts yielded no gain or a loss of yardage.

"Goal Line" in '07: There is no place to go but up for Goal Line. Part of the answer lies in that picture. Two offensive linemen, Paul Duncan and Mike Turkovich, lined up as tight ends. Given the current depth, there is going to be at least one lineman who doesn't crack the offensive line rotation; whoever it is needs to play in the Goal Line package. Imagine a left side of Chris Stewart, Duncan, and Turkovich. Who wouldn't want to run behind that? Watch out for Robert Hughes, too. He is an ideal power back, and Goal Line would seem to be a natural fit for him.


OUT PEOPLE
WR
TE
HB
FB
2
1
2
-

Stats
'05
'06
% used
0.4%
0.1%
% run
0%
0%
ypc
-
-
% pass 100%
100%
comp %
50%
0%
ypa
4.25
0

The Irish used Out People once last year. In the red zone against UCLA, Darius Walker and Munir Prince lined up as shown above. Walker motioned to the wide side of the field, leaving Prince alone in the backfield. Upon the snap, Prince ran a wheel route in the area vacated by Samardzija, and it appeared he had a step on the linebacker. However, Quinn had no time to throw the pass because one of UCLA's quick defensive ends had beaten Sam Young to the inside. The grouping was used four times in '05 with Walker and Travis Thomas; twice double screen passes were thrown, including on the first play of the game against Tennessee.

"Out People" in '07:
There is a lot of potential here. If an offense sends out two wide receivers, a tight end, and then flares both backs, the defense has a lot to think about and even more ground to cover. If there's man coverage, it's even better, because with the speed that the Irish can now send out onto the field, there will undoubtedly be a match-up that favors the offense. On the downside, can Notre Dame run out of this? That is an important question because, right now, seeing this personnel grouping obviously indicates a pass play. One last thought: double screen to Armando Allen.


DENVER
WR
TE
HB
FB
3
2
-
-

Stats
'05
'06
% used
0%
0.1%
% run
-
100%
ypc
-
3.00
% pass -
0%
comp %
-
-
ypa
-
-

Like its brethren Jax and 5 Wides, Denver always results in an empty backfield. The Irish ran this once, and it appeared to be a pass play that Quinn audibled at the line of scrimmage into a draw. Why it was run from Denver as opposed to Jax is a great question. It couldn't have been an injury issue, because the Irish used Jax three times in the same game. Weis must have seen something in UCLA's goal line defense that prompted him to do this, but whatever that was, Quinn's audible denied fans the opportunity to witness it.

"Denver" in '07: The talent is certainly there to play Denver, but what about our five tailbacks? It's hard to imagine that Weis will want to take the ball out of their hands all that often in 2007.