Jimmystats: Making Charts from UFRs Comment Count

Seth March 15th, 2016 at 10:07 AM

image

No, Upon Further Review series is not comprehensive. Most years are absent Ohio State and bowl games (including last year), and 2014 checked out after Indiana. That said, I challenge you to find a greater cache of free data than Brian's masterful charting of Michigan plays going back to the DeBord Throws Rock age.

Every so often I pull all that into a massive Excel file and try to learn things like how spread the offense was, favorite plays, etc. Let's dive in shall we?

What're those pie charts at top? Shows the relative efficiency (by yards per play on standard downs) and the mixes of Michigan's backfield formation choices. For "standard downs" I mean 1st and 2nd downs when the offense wasn't trying to do a clock thing or go a super-long or super-short distance. So no garbage time, no two-minute drills, no goal line, and no going off on Bowling Green and Delaware State. The idea is to show which offense did they get in when they had the full gamut to choose from, and how many yards did it get when the goal presumably was to get as many yards as possible.

Nothing very surprising there. Rodriguez ran his shotgun offense, Borges inherited Denard and Devin and still managed to jam them half-way into an under-center offense in three years. Then Nussmeier ran his zone melange single-back thing. Harbaugh did what Hoke always dreamed of doing, and the offense climbed back to about where Hoke's offense was with a senior (but oft injured) Denard.

---------------------------------

[Hit THE JUMP for each year's most charted play, visualized Hennecharts, how many TEs Harbaugh used, how many rushers defenses sent, and LOOOOOTS of charts.]

Favorite plays & Formations

This is going to favor runs because of how Brian charts—sacks get charted as "sacks" and scrambles as scrambles even though that's (usually) not the play call. Passing gets charted as the route that gets thrown to, not the route combination.

Season First Play Second Play Third Play
2008 Zone read stretch (19%) Zone read dive (7%) Bubble Screen (7%)
2009 Zone read stretch (15%) Zone read keeper (4%) Hitch (4%)
2010 Inside zone (13%) QB lead draw (8%) Zone stretch (7%)
2011 Zone read dive (13%) Power off tackle (9%) QB power (6%)
2012 Inside zone (5%) Iso (4%) Inverted veer keeper (3%)
2013 Zone stretch (9%) Power O (7%) Iso (7%)
2014 Inside zone (16%) Zone stretch (6%) Scramble (6%)
2015 Power O (13%) Inside zone (7%) Scramble (4%)

Rodriguez kept it run-first and Borges couldn't help it so long as he had Denard, but Michigan became somewhat more of a passing offense the last few years.

image

Spreadiness:

image

Yes Bo, Jim has fullbacks. And he used them quite a bit more than in the past. Note that the 2014 numbers above counted Joe Kerridge as a tight end mostly, so, yeah, Jim used more tight ends too. I can break this into their own annual lines. Average receivers in a formation is a good stand-in for "how spread are they."

image

Yes there were more tight ends than any team since at least Lloyd Carr was coaching,

image

but the big change from Nussmeier was an average of just two receivers on the field. Rodriguez had a standard three, but even Borges and Nussmeier went 3-wide as often as not. Half of Harbaugh's standard-down I-forms had just one receiver out there, with the other four backfield spots taken up by TEs and backs of all varieties.

Given the play of the guys past Chesson and Darboh, limiting receiver snaps might have been as much about the roster personnel as Harbaugh's personal preferences. When A.J. Williams was on the field most plays to do nothing but help Lewan(!) while Drew Dileo sat on the bench, that was frustrating. A.J. Williams as he was used this year—not to mention Sione Houma and Joe Kerridge—seemed eminently more dangerous to defenses than Grant Perry or the other depth receivers.

(If your eye is really good you'll note 2013 only adds up to about 4.8 guys on the field. That was Hoke pulling a TE for an extra OL all the time.)

Formations: Rather than showing you basically the same line chart again I thought it would be cool to show the efficiency of those formations over the years. Size of the bubble is the % of snaps in that formation, and the Y axis is yards per play:

image

Those bubbles jive with what we know about the various offensive minds Michigan's had since 2008. Rich Rod was almost totally shotgun; Borges kept trying to run I-forms more often despite quarterbacks more effective in the gun; Nussmeier's offense went single-back for the most part and was mediocre at everything. And then there's Harbaugh, who true to form did better with his Ace (usually with multiple TEs) and I-forms.

The shotgun performance was a bit surprising. True this doesn't include Rudock's magnificent performance against Florida, but it's also missing Ohio State.

As for that 2011 "Ace" outlier, that's all that Denard Jet stuff, including a 59-yard Touissant run against Purdue that I'm sure you remember, a 40-yard TE wheel to Koger v. Illinois that I'm sure you don't, and a lot of throwback screens.

I can cap gains at 20 and losses at 5 and run the numbers again but it won't change conclusions except…

image

…to make you even more angrier about the I-form and tackle-over crap they ran with a fourth-year Devin Gardner.

---------------------------------

Quarterbacks: Do they matter?

This chart says "uh huh." For this I re-added those long situations since that's kind of the QB's job. I'm not just showing passing plays here—this is how well the offense moved when these guys were taking the snap:

image

It also says Michigan's coaches made the right calls with their starters, for the most part. Denard is still the gold standard, but the season we got out of Rudock (not including Ohio State and the bowl game, which probably would help him in total) is up there around that of a senior Denard and third-year Gardner in 2012. Starting Morris continues to not make sense except in a "getting the people involved with that decision fired was the best thing for Michigan" sense.

Why don't you show the…

The chart by year. Hover over the headers if you need an explanation. ddd

Quarterback DO CA MA IN BR TA BA PR SCR DSR
Threet 2008 16 95 (4) 7 47 21 24 11 11   52%
Sheridan 2008 5 56 (2) 6 18 (1) 14 6   1   62%
Forcier 2009 7 76 (4) 7 17 6 15 2 13   67%
Robinson 2009 1 5 3 2 2 2       50%
Robinson 2010 7 73 (3) 5 20 (1) 3 3 8 (1) 5 1 70%
Forcier 2010 1 21 (1) 1 3 6 2   1 3 69%
Gardner 2010   2     1         67%
Robinson 2011 22 104 (18) 15 (1) 49 (1) 24 11 6 (1) 19 15 (1) 61%
Gardner 2011   7 (1)   4   1   2 3 67%
Robinson 2012 22 71 (11) 11 (1) 20 (1) 10 4 12 11 6 68%
Gardner 2012 17 55 (11) 10 16 (2) 12 9 (1) 2 11 16 69%
Bellomy 2012 2 5 (1) 2 6 1   1 4 1 50%
Gardner 2013 39 133 (26) 17 (1) 37 (1) 25 (1) 23 6 (1) 43 32 69%
Gardner 2014 16 86 (27) 11 (3) 24 (1) 15 (2) 15 4 (1) 13 19 68%
Morris 2014 1 11 (2)   10 2 3 1   1 45%
Rudock 2015 30 143 (41) 22 (1) 31 (2) 27 25 12 26 18 67%
Speight 2015 1 3     2 1     1 63%

I found it odd that Gardner was more or less the same quarterback over his whole career. I guess we lost some of his worst parts of 2014 to Henri the Otter of Ennui. Rudock was about in that same distribution as Denard and Devin: 57% accurate, willing to scramble, prone to mistakes from system unfamiliarity (I'd really like to see how his Florida game charts).

I tried myriad ways of visualizing this and struck on donut charts (screens excised):

image

Rudock also faced Gardner-level pressure apparently.

Protection

The charting also showed significant upticks in pass rushers the offensive line had to deal with on passing plays as the offense went away from using the QB's legs. In fact since Michigan went mostly pro-style as of 2013 the OL has to fend off more pass rushers than they used to on 3rd and 10 when Denard was a sophomore.

image

note: passing downs include 2-minute drill plays.

Protection has been, well, up and down:

image

Denard was apparently Michigan's best pass blocker, i.e. the threat of him escaping kept DEs spying on the edge rather than attempting to close in for sacks. It's hard to prove anything but I think coaching has a lot to do with the above. Rodriguez was scheming to prevent pass rushing and once his offense matured it was excellent at it. Hoke maybe invited more attacking in '11 and '12 but still, you can only rush Denard with so many guys before you're asking to let him escape the pocket.

The 2013 team was legendarily bad at this. As for the high mark last year, it does leave out 7 sacks that were never charted versus Maryland and Ohio State at the end of the year, but Nussmeier's offense, schematically, was supposed to help with that. There's certainly room for improvement next year.

Comments

FanNamedOzzy

March 15th, 2016 at 10:46 AM ^

This must have taken a LOT of time and effort, and it really shows.

Most interesting thing I see is that the protection really wasn't THAT much better in 2015 compared to 2014, which is surprising to me. It just seemed like Gardner had no time at all to get anything done. In reality, it's likely that the QB coaching Rudock had allowed him to make decent reads despite the pressure and not get into the happy-feet mode that Gardner often found himself in.

Seth

March 15th, 2016 at 11:46 AM ^

I think there's a different story. Actually two.

First, schedule strength. The games I have charted are:

2014: App State, ND, Miami(NTM), Utah, Minn, PSU, MSU, Indiana
2015: Utah, Ore St, UNLV BYU, Maryland, NWern, MSU, Minn, Rutgers, Indiana, PSU

For comparison's sake let's say the matching teams were about the same. Then ASU=UNLV. So it's ND and Miami vs Oregon State, BYU, Maryland, and Northwestern. Maryland had Ngakoue, BYU was one of the better pass rushing teams due to their great DE.

Not appearing for last year: the M00N game, and a really bad protection day against Maryland (Ohio State is excised both years but 2014 OSU had 5 sacks).

Not charted.

Second, the 2014 OL got better as the season progressed.

Despite none of those teams being at all blitzy, Michigan's OL hovered between 71 and 74 percent protection in the first four games. But they were at near 87% against Penn State's ridiculous DL, and 100% against Indiana. So even if they fell off at the end of the season against very good DL in Ohio State and Maryland, we talked a lot about how Cole progressed as a true freshman from lost-ish to solid, and Braden was even okay by the end of the year.

As for 2015, the line was up and down every week. They struggled against Utah, BYU and Penn State, and got mauled by Michigan State, but hit 85% or higher the rest of the way. The charts must be consulted but if I had to guess I'd say Drevno was focusing more on installing his power running game for one, and two there are physical limitations of Mags and Cole as the edge blockers against elite rushers.

Keep in mind Taylor Lewan and Mike Schofield were on a lot of the lines we're comparing those guys against.

Anyway the point is 2014's OL was bad at the beginning of the year but by mid-season they were fine, but that lacking a true tackle (Braden, Mags and Cole all are best inside) they couldn't really get better than that.

DonAZ

March 15th, 2016 at 11:19 AM ^

This chart ...

Confused me at first ... but when I read the legend and started inspecting the rings it made sense.  It's a chart full of interesting stuff, but the eye has to adjust to it.

What it seems to say is (a) Gardner was pretty good, and (b) Forcier was as well.

(Both Robinson and Gardner have the X-factor -- Dileo -- working for them ... that little rascal would end up getting open and catching things no mere mortal should have.)

Oh what might have been -- Gardner under Harbaugh's tutelage with a decent OL.

Blue Ballin'

March 15th, 2016 at 11:25 AM ^

Great work, Seth! This is the type of detail I don't see anywhere else!

No surpise that Gardner had to scramble so much, but found it interesting that Tate was tossing more catchable balls than anyone else charted.  

Chris S

March 15th, 2016 at 11:30 AM ^

For real, great post Seth.

This is post alone would stand alone on any other site, but just another one of many great ones here. Nobody is messing with MGoBlog.

I think my favorite chart was the Average #pass rushers. It shows what other teams as a whole thought of Michigan.

Space Coyote

March 15th, 2016 at 11:43 AM ^

I disagree a little on some of the conclusions to take from the charts, but the information itself is very informative and does tell quite the story. Very good stuff. This post, as a thorough breakdown of Michigan football, stands with essentially any post this site has created, which is saying something.

Question on the "pass rushers sent" chart though: Is there something wrong with the y-axis or is it some sort of additional rushers (to the typical 4) or am I just not getting it?

gbdub

March 15th, 2016 at 11:48 AM ^

So will Brian ever go back and do some of those UFRs? With as essential as they are to MGoBlog in general, it's disappointing that we almost never get them for the biggest games. And Seth has just demonstrated what sort of cool stuff you can do with them.

I get Henri, but I'd still rather read a depressing UFR than a lot of the other offseason content.

Sent from MGoBlog HD for iPhone & iPad

Space Coyote

March 15th, 2016 at 12:08 PM ^

I'm known to voice my difference of opinion on some of what Brian says in some of his UFRs, but it is vastly more information and closer to the truth than we typically get without them (and as I've noted, I agree with the majority of it).

And to not chart some of the most important games from year-to-year leaves a huge gap in what we really know. The eye test lies, and at times it lies in major ways, so to not have bowl game information (final game of the season) or OSU game information (the last regular season game and a team we measure ourselves against) is, when weighted for importance, probably leaving a third of the value of the UFR on the table, if not more.

Similarly, I still think there is a lot of value in seeing what happened in a loss. What actually broke down, what actually went right, because I'll tell you, there is still plenty of both in a loss. And there is still plenty of both in bowl game wins. I haven't gone back and closely watched the OSU game, but I have a feeling going back that there is much more we'd be happy about with the offense. On defense, how much was pre-snap erros vs post-snap failing to execute? How did someone like Wormley perform against Decker? What about the safeties in general in those games, vs OSU and then after several weeks of practice without Peppers in the bowl game? That shows their potential improvement, if it's clicking, etc. It's extremely valuable information.

I look at MSU/Bama for instance, which I have rewatched once. Conklin actually had a really good game against them, particularly in protection. This would be important information for them if he was coming back. The DL and LBs actually held up well against the Bama OL for the most part, until they got worn down by the offense playing terrible. Again, there is value there for them. The same is true for Michigan, despite being blown out by OSU. Missing that information carries a lot of weight, for this sort of analysis (and it'd be interesting to see the difference between season trends and then "big game" trends), for actually understanding Michigan's performance/improvement, and for completeness.

gbdub

March 15th, 2016 at 1:34 PM ^

I guess I don't know what percentage of the effort goes into each section, but I'd be totally fine with skipping the "narrative" for each play, and just getting the scores. Maybe a sentence or two about anything interesting that stuck out, but no need for most of the editorializing comments at the end (which are great in the week of but probably don't mean much now).

Sent from MGoBlog HD for iPhone & iPad

Space Coyote

March 15th, 2016 at 11:57 AM ^

Figure 1 (formation by year). Was surprised to see I-Form stay fairly constant throughout Rich Rod's tenure, and Borges's tenure. I actually would have thought they went up with both. I was also surprised just how little ace formation Nuss used, though there is certainly some pistol mixed in there which serves as the same thing essentially. Would have been nice to see a pistol and shotgun breakdown, as I figure a strong portion of the gun/pistol mix in 2013 was pistol.

Table 1 (top 3 plays). Sack wasn't a top 3 play in 2013? Unfortunately impressive.

Figure 2 (pass/run on 1st down). This matched what I believed, that many people disagreed with based on the eye test. But as Borges's tenure progressed, he actually passed more on first down, until he was about 50/50 (which is probably more passing than he'd have preferred). Think people got caught up in the lack of run success that they felt he also ran on first down, but he was actually very balanced.

Figure 4. You pointed out that both Nuss and Borges ran 3-wide as much as anything else. I'm actually quite surprised how close they ever got to the number of backs that Harbaugh used per play. Also really surprised how much Rich Rod's TE usage dropped in 2010. Thought I remembered him using the TE more on standard downs that year, but it really dipped.

Figure 6 (backfield formation). I think Ace and I-Form went way up when DG was playing relative to when Denard was playing, which somewhat skews the stats there. I think in 2013, a take away between Figure 7 and Figure 6 is that Michigan got sacked a lot in gun, and also had a lot of very negative plays (bad snaps, poor reads by DG, etc). It also shows just how potent Michigan was with PA from Ace. The vast majority of big plays I remember were PA from Ace formation, which was a major part of Borges's strengths as a guy that could design routes to get guys open.

Figure ? (rushers sent): No surprise that teams sent fewer rushers at Denard. Biggest threat with his legs (although much of that was perceived threat because he really didn't scramble that much) and also the idea to make him beat you with his arm (teams rarely sent a true rush, focusing more on staying in their lanes, and forcing Denard to read the defense and beat it by throwing into smaller, zone windows). I think in 2012 teams also didn't know what to expect from a DG offense, so they played a lot more base. Obviously, in 2013, you see all hell breaking loose on the OL and teams just sent the house because Michigan couldn't pick up anything.

Anyway, lots to take away from this data. Like I said, this is really good information.

Seth

March 15th, 2016 at 1:41 PM ^

1. Formations: Why did you think Rich Rod used more I-form as he went along? In 2008 he used the West Virginia spread in its most basic form. In 2009 he played around more with under center stuff because he was starting Forcier. But in 2010 he went with Denard--even the goal line plays were shotgun.

Pistol was 10% of the 2013 offense, and 5% of the 2014 offense on standard downs. Those are the only years it made regular appearances.

O-Form 2008 2009 2010 2011 2012 2013 2014 2015
Shotgun 91% 87% 95% 72% 63% 32% 61% 26%
Pistol  -   -   -   -  >1% 9.5% 4.7% 1.3%
Ace 3% 6% 1% 10% 13% 27% 26% 27%
I-Form 6% 7% 4% 16% 21% 24% 7% 43%
Weirdness 0% 0% 0% 2% 3% 7% 0% 2%

Like most Hoke offenses, they started out with more I-forms and Ace formations, and became more shotgun the more the reality of their roster intruded on that fantasy. 

Except when kicking Indiana's butt allowed them to do otherwise, after the Minnesota game Michigan was a base shotgun because you'd have to be an idiot not to.

I'll get to the rest later.

Space Coyote

March 15th, 2016 at 1:57 PM ^

As he went along because I know for a fact that he wanted to be able to run some under center plays. The first drive against Iowa comes to mind. I know he was never going to run it even the vast majority of the time, but he certainly wanted it in the playbook. I figured in his first year he was so focused on installing his base offense, and by year 3 he would be adding the extra things he wanted in his offense, which included an under center component.

Also, I would argue that the ND game, where they were vastly higher from gun and pistol, forced them to adjust their path going forward in 2014. While I agree that they started out doing what they assumed was optimal and adjusted more to gun as the season went along, I think getting shutout by ND forced them to reevaluate in a big way what they were doing on offense, which caused them to try to be more balanced with their formations. Obviously, though, that wasn't really to root of the issues.

Seth

March 15th, 2016 at 4:34 PM ^

2. Passing on 1st down

Season 1st Dn Run% HC/OC
2008 66% RR/Magee
2009 67% RR/Magee
2010 67% RR/Magee
2011 72% Hoke/Borges
2012 67% Hoke/Borges
2013 63% Hoke/Borges
2014 51% Hoke/Nussmeier
2015 59% Harbaugh/Drevno

Borges ran A TON on 1st down--in 2011 even more than any of RR's teams--but that dropped progressively over his tenure. But that drop entirely a function of Denard vs Devin:

QB Season 1st dn Run%
Denard 2011 71%
Denard 2012 78%
Gardner 2012 48%
Gardner 2013 63%

Borges didn't change. He just thought Denard was a running back and Gardner was Tom Brady. He never got to 50/50. He did become slightly less than the 2/3rds of the time 1st down run rate of Rich "The Spread 'n Shred Offense guy" Rodriguez. Nussmeier was 50/50.

3. On TE usage by RR and FB usage by Nuss and Borges

I think you're mixing up Borges trying to I-form with Denard against Iowa in 2011 and Rodriguez in 2010, who used an I-form 3-wide formation they used ONCE before Forcier came in. He did use it a bunch for a Forcier drive because Forcier sucked at zone read anyway, so they built him a zone stretch/IZ/ISO/Power/Waggle package out of that 3-wide I-form (no TEs) with Kevin Grady at fullback. That was the "drive of RAGE" that ended with a Minor TD. The next drive Iowa adjusted, and after that it was comeback time.

For both 2009 and 2010 RR learned to use TEs more often. He inherited Koger and Webb, but didn't have any guys behind those two. At West Virginia RR typically didn't use TEs; he liked having a "Superback" instead. When they did start using TEs however, they only rarely put both on the field. And those 3TE sets that shift the average were non-existant.

So why did 2009 have more TE usage than 2010? Look at the WR sets. 2009 was the year a lot of the WRs were injured or sick or unavailable for other reasons, so Rich Rod rarely went more than 3-wide. Odoms, Hemingway, Savoy, and T-Rob all missed time. Roundtree, a skinny ass RS freshamn, was the starting slot and leading receiver. Mathews was hobbled. Stonum needed glasses. In 2010 they could use a lot more 4-wide sets because RR could play any of 'tree, Stonum, Hemingway, Odoms, Grady, or Gallon.

As for Fullbacks by Hoke's OCs, Nussmeier rarely had more than one, and Borges used one about 1 in 3 plays. Harbaugh had a fullback in there over half the time.

4. Backfield formation. Your memory is good on the sack rates, but bad on the context:

Sack rate Shotgun Pistol Ace I-Form
Standard downs 3.3% 0.0% 4.5% 0.7%
All plays 7.2% 0.0% 4.0% 0.6%

You remember all those sacks against shotgun passes because Michigan would burn its first two downs running despite not knowing how to, then put Gardner in a shotgun, burn a tight end to help Lewan's side, and watch him get chased by the inevitable up-the-middle rush.

On standard downs however the Ace formation led to a higher sack rate than the gun.

5. No, the Shotgun was still the most effective formation in 2013, though the Ace was kinda close:

Again this is standard downs. At least from the guy they gained yards 2/3rds of the time. If you look at the line of 0 or negative plays (including interceptions), the only formation that really stood out was the Pistol. 

 

kstevens26

March 15th, 2016 at 12:20 PM ^

This just makes me wonder what Harbaugh could've done with Gardner. The kid had a giant heart and great work ethic. He could've been something special.

Space Coyote

March 15th, 2016 at 1:11 PM ^

However, I do think there would have some struggles with the mental processing part of the game, particularly his field vision.

DG was at his best in 2012 when the offense was simplified and the defense was simplified because defenses didn't know what to expect. Yes, there were points were you saw issues with his vision, but that could be chalked up to him just moving back to the position.

In 2013, the issues resurfaced almost immediately, and confirmed some of the concern. While he was essentially Superman against ND that year, he kind of had to be because he allowed his eyes to drop down to the pass rush very easily. His OL certainly didn't help matters, and in fact, likely made it worse, so that prevented a lot of improvement. Borges's reaction to this was to run longer, more complex routes that resulted in significantly easier reads. The issue, again, was the fact that these longer results left DG vulnerable to the pass rush with a really bad OL. It wasn't a great combo.

Nuss took over, and his answer was to drastically simplify the route structures and reads. Unfortunately, with athletes recruited for an offense that relied on more complex route structures and rub routes, they weren't great at getting separation in this simplified offense. To my impression, that made the windows smaller, and forced DG to be better with his timing (two things he struggled with), though no one else on the roster proved better. Part of the flawed QB recruiting strategy was taking QBs with high ceiling but equally low floors. When these guys struggled to reach their ceiling or get off the floor, there wasn't really any QB with a high floor, which is why I believe Nuss ultimately recruited Malzone, who had the highest floor for a Michigan QB recruit since probably Henne (Tate would have been higher from a pure QB point of view, but his off the field issues made him a concern).

Now to Harbaugh. Harbaugh would have adjusted his offense to be more similar to his offense in San Fran. The issue would have been that Harbaugh's offense relies heavily on QB's reads, and reading the defense on time. Yes, it is simplified relative to his offense with, say, Andrew Luck, but it still has an emphasis on it. If anyone could improve that part of DG's game, it would be Harbaugh, but Nuss had a good record with that and struggled with DG as well. I think Harbaugh would have struggled a bit running his offense with DG still, although physically, DG would have been perfect.

As much as I hate to say it, I think Meyer's offense with Cardale would be optimal for DG. It was essentially a pro-style run offense that allowed the QB about 10 carries a game but really simplified the pass reads for the QB. Of course, this requires great OL play as well, something Michigan was never going to have in 2013, so it's hard to say either way. I honestly think with a stronger OL, DG fit Borges's scheme quite well. Sling the ball down the field, easier reads, hit the far sideline, use your legs a bit to scramble and a few designed runs, that kind of explains both Meyer's offense with Jones and Borges's ideal offense with DG. Again, the OL situation though didn't allow for it.

imafreak1

March 15th, 2016 at 12:49 PM ^

I always look at this type of data with an eye to confirm my disdain for Nussmeier especially in relation to Borges.

It is interesting to see in the "Spreadiness" figure how much variability from game to game there was under Borges. Particularly in 2013 when he was really changing his approach week to week either out of desperation or in response to the opponent. But then in the one year under Nuss the game to game numbers are pretty flat and similar. Nuss seems to have used the same approach with regards to personnel regardless of opponent. I always wondered when Nuss talked about watching video and gameplanning from week to week what exactly he was doing. To my eyes, it looked like nothing ever changed.

If you look at the bubble figures (yards per play by formation), you can see what we already knew about the 2014 offense. Nothing worked particularly well. Not passing. Not running. They all sucked equally. Nothing was particularly explosive.

With the Harbz data, you can see what your eye balls already told you. Harbz swung between modest gains and chunk plays.I'm sure that was by design and someone with more time and knowledge could explain it quite simply.

Unlike the Nuss offense which appears to be based on small incremental gains on most plays rather than explosive plays (regardless of how much he talked about explosive plays), the Harbz offense is about purposefully creating situations and matchups for exploitation and explosive plays. Borges believed more in this approach but was clearly not as good at it as Harbz.

I think it is pretty clear which approach I prefer.

1985sec4row23

March 15th, 2016 at 1:08 PM ^

Ace- how about the standard deviation for yards gained over time? I suspect that with the spread under Rich Rod, there was much more variability in what you get on any given play. Thus, it can be hard to sustain drives, even if you are getting 7 yards per play on average.