49ers vs. Vikings: Arif's Notes on the Game (Offense)

We've soaked in our win against what many people thought was the best team in the NFC, if not the NFL. Whether or not that assessment is true is irrelevant—the Vikings beat a great team and did it on their own terms. The offense was a big part of making sure any victory wasn't a fluke, and contributed by putting up 24 points against a dominating defense that last year only allowed 14.3 points a game—second in the league.

It's an offense that has been more efficient than many expected, both in relatively well known statistics (they ranked 18th in offensive yards per play, and 11th in yards per pass attempt) and in more obscure, but market-tested predictive statistics (they rank 15th in offensive efficiency, 18th in offensive simple rating system, 9th in net yards per attempt, 11th in defense-adjusted value over average, 13th in offensive success rate, 6th in pass success rate, and 18th in run success rate). The offense has also been leading in a number of familiar and unfamiliar explanatory statistics (4th in passer rating, 14th in offensive expected points added and win probability added, and 6th in adjusted net yards per attempt). If you are curious, there are a number of statistical measures that would put the New Orleans Saints* below the Cleveland Browns (although not below Tennessee or Kansas City).

People are happy with this offense and rightly so. It just so happens that there are multiple elements that allow this improving offense to "go," not the least of which is an adaptive game staff. A safe, but efficient quarterback lead the way, while the workmen of the offensive line do what they can to give him time. An additional threat at receiver and running back has helped a lot, and an emerging tight end is proving key.

Playcalling

The Swing Pass

I'll keep beating my Bill Musgrave drum. His playcalling has been superb through three games, and much of it has to do with an evolving gameplan. Musgrave and the other coordinators have not only adapted their gameplans to the weapons they have on the field, they have changed it according to current opponent and a season-long evolution. To demonstrate that concept, I'll do a short diagram of a play below, which is simply a swing pass to Percy Harvin.

First, I'll start with the play that helped set it up: a play action end around intended for Percy Harvin. The end around is a fairly common play for the Vikings, who used it quite well in 2011 for a few big gains, and deployed it at least once in 2012. Below are two snapshots of the Vikings deploying the end around against the Chiefs in Week 4 of the 2011 season.

Donovan McNabb is currently at the mesh point with Peterson and selling the run as Harvin runs behind the handoff for the play. McNabb then pitches it to Harvin who will attempt to outrun the player circled in red below.

The circling is there to indicate that in this game against a 3-4 system, the plan was to leave outside linebacker Tamba Hali unblocked in order to sell the run up the gut. That is, the left tackle would crash down to the defensive end so that the guard could move up and make a second level block. Hali actually does a great job not biting on the play action and waiting for the play to develop. Still, he can't prevent Harvin from breaking an 8-yard gain.

One of the keys to the play is that it starts on one hash mark, and Percy Harvin runs to the far hashes—a play design that gives Harvin room.

At this point, Hali is the only Chiefs player free and close enough to Harvin to make a play. Inside linebackers Derrick Johnson and Jovan Belcher are too far away from the play to make a difference.

The Vikings run a slightly different play against the 49ers. They block Aldon Smith to sell the pass off the play action (and keep him occupied), which might alert the 49ers to expect the pass and back the inside linebackers up. At this point, it looks like the Vikings have left two runners dead in the water and out of the play (until AP is presumably open later as the play action develops).

The play once again starts with Harvin lined up on the near hashes and pulling back and out as if reversing.

Carlos Rogers is trying to follow Harvin, but will naturally lag behind as he's attempting to read the play as well. Inside linebackers Navarro Bowman and Patrick Willis are grouped pretty closely to the near hash, giving Harvin space as he moves to the open side of the field. This play is more effective against man coverage—something the 49ers do a lot—than against zone coverage. This is for no other reason than the fact that there would be a defender patrolling the flat or curl/hook zone that Ponder enters to begin his run.

As Ponder hitches forward, Bowman crashes upfield, then back into coverage while Willis stays still. Both of those decisions take them out of the play. It no longer looks like there will be a runner, because Ponder has turned away from Harvin. This keys Willis, who opened his hips in order to stop a reverse or end around, to stay inside. Tarrell Brown is also wary of a run by Harvin, but releases to cover Jenkins as Ponder turns away from Peterson and Harvin.

On this play, Harvin received the pass 10 yards behind the line of scrimmage, but there were no defenders near him.

It's only because Willis has excellent reaction time and good speed that he ends up making a tackle at all, but by then, Harvin already moved the chains for a first down.

There are a few reasons this play exemplifies what I've enjoyed from the playcalling group at the Vikings. The first is that it is a natural evolution of the plays that the Vikings are known to run, particularly the devastating reverses and end arounds that Harvin ran much of last year and a little bit this year. It takes advantage of a well-known tendency, and tweaks it enough so that anyone taking advantage of the tendency will get exploited (because movement will shift coverage enough for the quarterback to move to another receiver, probably the running back).

The second reason I liked the play design is because it also took advantage of the defense's scheme, which includes largely man coverage. This means that there is empty space on the swing pass given Rogers' natural lag on finding Harvin. Harvin knows where he's going to be, doesn't have to make any reads and runs at a good speed, all luxuries Rogers doesn't have. It looks like Harvin's taken out of the play at that point, because Ponder turns downfield.

Finally, the play was entirely consistent with the gameplan Minnesota was gunning for, which was designed to limit turnovers and find creative ways to put playmakers in space. This was also part of a strategy that combined a very heavy running load with play action passes designed to freeze the defense. Eric expands on the run/pass balance the Vikings have maintained here. Right now, the Vikings rank 8th overall in the percentage of plays that are runs, with 48.2% of all offensive snaps dedicated to rushing the ball. In the Niners game, the Vikings called for 42 passes (7 of which became scrambles for a gain) and 34 runs, which means 44.7% of their plays were runs.

What was screencaptured above was the second play of the second drive of the third quarter, but I enjoyed it on the second play of the first drive of the game overall. I also enjoyed Rudolph's reception on the first drive that converted the third down with 9:45 left in first quarter. My favorite call of the night was the play action for the first touchdown on 4th and 1. Some people know me as a big advocate for going for it on 4th down, but this was another type of good decision.

Getting a touchdown instead of a field goal on the first drive puts the Vikings in the driver's seat and forces Alex Smith to pass instead of forcing Frank Gore to run. It's hard to plan for "get points early and make them play from behind," but when it's possible it should be done against the 49ers. Shifting them from a running attack to a passing attack gave leeway to the defense.

Further, the fact that it was a play action run after three consecutive runs up the middle was excellent. The decision to fake a run may not seem exactly inspired, but here it was a good way to take advantage of the Vikings' identity as a team as well as its tendency on 4th and 1. Every single player save a blitzing Dashon Goldson bit on the run, allowing Rudolph to find himself free.

I thought the call was great, but the play design was a little off, simply because Jerome Felton and Kyle Rudolph were in the same area at the same time. This is a little hard to avoid, particularly given that it's a play action rollout, which forces all routes to the direction of the rollout (to avoid difficult throws across the chest). Felton and Rudolph were the only receivers, but a backside option for a tight end (for a very difficult throw, but an additional read nevertheless) would have been nice.

Overall, I appreciated the design of most routes, which not only took advantage of outside man coverage, but found ways to open up normally tight windows, including levels concepts from the same side of the play (players streaking across the middle going in the same direction but with one of the players 5-8 yards further upfield). This made throws into bracketed coverage easier because it would force a covering player to break momentum and shift in the opposite direction in order to affect the flight path of the ball.

Strategy

Generally the philosophy was sound, and encouraged more plays than most teams will run (with 39 offensive snaps by the end of the first half, not including kicks or punts), which made sure that a) the Vikings offense to tired out the 49ers defense and b) the 49ers offense was off the field. Taking up time and reducing possessions is a time-honored way to beat teams marked as favorites. I wrote about my preferred philosophy against the 49ers here:

In this case, high-risk, high-reward does not necessarily mean risky individual plays, like gadgets or deep bombs. It means making the difference in skill less and less important. Against a heavily favored squad, the best opportunity to win games is making sure there are as few scoring opportunities as possible.

This may not seem risky, and is certainly not aggressive, but gives more of an opportunity for randomness to insert itself into the outcome. Statisticians would say that one is "reducing iterations" by doing this, and encouraging a boom/bust strategy.

...

[T]he law of large numbers ... is true for football and possessions. The more often the underdog allows the favorite to prove that they're the superior team, the more likely it is that the favorite will win.

Without the fumbles or kneeldowns, the Vikings had 8 offensive possessions. That's on the extreme lower end of the curve. A large part of this was in the game plan, and it shockingly has very little to do with wasting first and second downs. Given that the Vikings only had 14 third down scenarios (and only 5 in the first half, when they weren't trying to run out the clock), but 25 first downs overall, they were very efficient in this regard.

Despite my overall happiness with the planning, there were a few plays that I thought were poorly designed. On the third drive, the call from 2nd-and-13 the Vikings called for a screen to Harvin, with tight ends Rhett Ellison and Kyle Rudolph bunched near him and designated as blockers. There were no pulling linemen, and the bunched receivers to start the play had already attracted attention from the linebackers and safeties. Without additional blockers, Harvin wasn't going to go far. He only gained two yards on the play.

I also disliked the first play of the first Vikings drive of the second half, which was an Adrian Peterson run up the middle. The issue I have with the play is not that Adrian ran up the middle, but that the play called for Aldon Smith to remain unblocked as Kalil took on Justin Smith on the left side of the field. Brandon Fusco and Phil Loadholt crash down on Ahmad Brooks on the right side of the field, but Fusco is left blocking nobody by the time Peterson is three yards behind the line of scrimmage. Even when he slides free, there is no one on the second level to block, because they are occupied by Rudolph and Sullivan.

Aldon Smith ends up in perfect position to tackle, although game scorers only gave it to Navarro Bowman. Blocking Bowman is a different issue; there was someone who was on that assignment, but letting a potential playside tackler go free is puzzling. I didn't mind the call so much as the design. At any rate, it sounds like Ponder was audibling to the second play they had called anyway, given that he interrupted his cadence by yelling "Kill! Kill!" under center.

These objections are few and far between, however, and I appreciated seeing a gameplan that took advantage of what the 49ers were doing, as well as one that stayed within the constraints of the philosophy and talent.

Many of the players executed this gameplan well, too.

Formations

The Vikings remain committed to the "11" personnel package as the one that makes their offense run, and rely on it heavily. They've somewhat reduced their reliance on it, and didn't use it on their two third-and-short plays before switching to clock-grinding mode. They used this package 31 times of their 76 functional offensive snaps (not the three kneeldowns), which is less than before.

One entire (no-huddle) drive used only this package, which speaks to how much the Vikings value it. More curious to me, however, was that they only used a fullback in 29 of their plays. Outside of the last 12 plays (three or so possessions), they used a fullback 21 of the 65 plays from scrimmage that were offensive tries. That's only a third of the offensive plays, compared to near 50 percent of the time in the previous two weeks.

Players

Christian Ponder

This was Christian Ponder's best game yet, from a grading perspective. While it's easy to see that his statistical measures (60% completion rate, 5.7 yards per attempt, 198 yards in the air) may imply otherwise, the number of positive or "good" plays that he made was more than in any other game. It was also the first time his first half grading was net positive. Again, my definition of a "good" play may vary with others. Here, it means a play that exceeds the ability, intelligence or decisionmaking of an average quarterback in the same situation. A "bad" play, conversely, is one whose process is subpar in some way compared to the average quarterback in the same situation.

The vast majority of plays would go ungraded. Ponder ended up with twelve good plays to seven bad plays, and had his most consistent performance all around. This time around, he clearly went through reads, and performed a top-down look at plays, making sure to see if deep receivers were open at least five times by my count. On none of those plays did he see something he liked, so he went with an intermediate option twice, a short option twice, and scrambled once.

He did a much better job sensing pressure in the pocket, although his first scramble of the game was a bit curious. He's improved in terms of ball placement, but still has a long way to go. I didn't keep track of every pass, but I still noticed at least three plays that were hurt by poor placement for receivers on the run, two of whom missed the pass as a result.

Most impressive, however, was his play under pressure. With pressure in his face, he consistently converted, whether it was a first down or touchdown. The poise and accuracy under pressure was unprecedented for him, and it was the first time Pro Football Focus gave him a definitively positive grade for performance under pressure, despite low accuracy when facing a threatening pass rusher (3/11 passes completed). Part of this is because two of these passes were relatively wise throwaways, one was a drop, and another was a play where he was hit as he was thrown.

He took no sacks, a testament to his scrambling ability, having made the right decision to scramble 4 times and having made a questionable decision on three other attempts. It's easy to see that his 23 yard touchdown scramble was the right call, but his two earlier attempts were disconcerting, particularly given that they were his first two scrambles at all. On the first play, a somewhat difficult pass to Aromashodu exists for a first down, and Ponder could have thrown Aromashodu open by leading the ball into the middle of the field. No 49er would have been able to intercept.

The second scramble is more understandable but no better a decision, given that Jerome Felton is free on the backside of the play in the flats with no defender within 7 or 8 yards of him.

Naturally, it would be negligent not to mention Christian Ponder's well-thrown pass to Kyle Rudolph for the second touchdown. I didn't give Ponder a positive grade, simply because Peterson was not only open, but had no players anywhere near him for the entire play while in the end zone. Ponder made it too difficult for himself and still has a lot of work to do to make sure he goes through his reads.

Overall, he still is largely only reading half the field and will need to progressing to reading the full field. Sometimes this presents itself as staring down one receiver, and other times it presents itself as only having two reads before scrambling. Luckily, Ponder's first read still tends to be open on most plays. Naturally, of your first read is open, throwing to him is a good decision.

He threaded tight windows and found ways to move the chains. He stayed within the limits of what the offense gave him, and only missed a good decision for one deep pass. Incidentally, I gave him a negative grade for an overthrow relatively deep to Michael Jenkins that could have been easily intercepted by many defensive players, but Chris Culliver was not up for the task.

I did not negatively grade Ponder for the near interception by Donte Whitner late in the game, because when Christian Ponder is halfway through the throw, Jenkins is still upright and relatively open. Jenkins would have had to move up to high-point the ball to best make himself open, but the throw was good and relatively safe. Michael Jenkins fell down on the play when turning around for the comeback, and Whitner dropped a gift. It was a fine throw and an OK decision, but sometimes players don't always do what you expect them to do.

Wide Receivers

Percy Harvin once again put together a stellar performance. He was targeted 11 times and pulled the ball in 9 times. He still dropped the ball on one play, but was otherwise reliable. He is probably at fault for the other pass, too, as he was running out and missed the well-placed ball by a hair. His 24 yard gain could have gone for much, much longer if he had not stepped out of bounds, but he made a fantastic play regardless. Out of all of his games this year, this is his first where he didn't force 4 missed tackles (he only forced 2), but that's to be expected from a strong tackling defense. His utility role was important, even though he only rushed from scrimmage once (he bounced outside for 9 yards, a good decision). His strong route running and rare agility was on display once again.

Incidentally, yards per route run is my favorite receiver statistic, because it incorporates the ability to get open, chemistry with the quarterback, the ability to haul in a catch and the ability to gain yards. Harvin ranks second of all receivers, with only Steve Smith ahead of him. He has just over 3 yards per route run, and this would have made him second overall last year (when he was fifth overall).

Michael Jenkins, on the other hand, was a bit of a disappointment. There were at least two plays where he was asked to run block and functionally did not do it at all. Both times, the defender he was blocking made the play and tackled the runner. It was frankly a bit aggravating to see a player take this kind of play off, because it had a direct effect on the yardage that the Vikings gained. Both of these were near the end of the game, but that doesn't excuse his clear lack of effort. He also did poorly on at least one pass play. He was targeted five times, and only had one reception, although not all of the targets were catchable balls. He did a good job clearing defenders out of the way with passing routes, but that does not lend him too much credit.

Only Stephen Burton and Devin Aromashodu were given receiving targets otherwise, and Burton only had one. Burton's target did not result in a reception and this is in no way Burton's fault—an extremely poorly placed and timed pass sailed over his head as he was still fighting for the release on his slant route. His release in general has not been the greatest, but that wasn't the issue on this play. Overall, I appreciated his route running, but he couldn't seem to separate from the San Francisco corners. He missed a run block early in the game, but did not appreciably make many mistakes in the blocking game from then on.

Aromashodu, however, seems to have turned a corner. While he'll never be a feature receiver in the offense, nor a reliable starter, he no longer consistently draws the ire of fans in the stands or from those watching at home. He's still the best route runner on the team, and may be the fastest receiver as well. He doesn't and hasn't reliably played his defender very well and sometimes has trouble getting open. The bigger problem fans have had with him is not what he does to get open, but his reliability once the ball is thrown to him. Last year, Aromashodu caught 33.3% of passes thrown his way. This year, it has been 75% and he caught 100% of the passes thrown to him (2 of 2) on Sunday. One of his catches was a crucial third down conversion on what ended up being a touchdown drive (which ended in a Ponder scramble). My only issue with his play relative to his role was that he missed a run block late in the game. I don't know about you, but I really dislike when players turn off at the end of the game when play is still live.

His play during the preseason as well as his current run of play in the regular season may have helped him avoid being cut in favor of Allen Reisner when Simpson's roster exemption ended.

Offensive Line

The offensive line gave up no sacks, but that hides what was actually a subpar performance. Pro Football Focus faults them for having given up 10 hurries and one quarterback hit, but I counted at least one more hurry than that, if not two. The run blocking in the game was a huge disappointment more than anything else, however. Out of all of the pass blocking performances in the NFL on Week 3, I would rank the Vikings in the bottom third, having watched all of the games. Out of the run blocking performances, I would rank them in the bottom five. This is the greatest area for concern from the game for me.

The only player who did not concern me was Matt Kalil. I still feel like there are some scheme concerns, but he hasn't shown me anything all season or preseason that would make me concerned about his technical ability or strength. Twice I saw him miss out on blocking Aldon Smith, starting with the first drive of the second quarter, specifically on the play starting with 9:49 left in the half. He just gets turned too easily and may not realize how far back Ponder is. I was worried about him in pass blocking three or four more times over the course of play, but only one more serious error, which ended up being irrelevant because of Dashon Goldson's unnecessary roughness penalty downfield on Kyle Rudolph. Ponder got the ball out quickly anyway. Still, it was a turnstile play.

His run blocking wasn't phenomenal, but neither was it bad. His second level block on Willis on the third play of the second drive was fairly perfunctory and could have caused problems if Peterson had moved up. Other than that, I didn't note a poor run blocking assignment after that. Extremely encouraging.

Next to him was the mercurial Charlie Johnson, who I've been high and low on as a guard. This week, I'm low. Usually when I write in my notebook about an offensive lineman, it's bad. I wrote more under the CJ column than anyone else, period. I saw consistent pass protection problems, particularly when they twisted Aldon Smith inside of Justin Smith, but also did not have the strength and/or technique to deal with Justin on the inside. It was the Johnson we were familiar with in 2011, and the tight spaces did not help him out in the guard position.

Only two real hurries and a hit, but he was saved by quick releases more often than not. Justin Smith didn't quite toy with him, but it was a clear mismatch. There were some communication problems it seemed between him and Sullivan, but I blame the center for most of those. I did not see a single Johnson block in the run game that made me happy, and he did a poor job sliding off to get to the second level. On the occasions he did, it didn't seem like a winning effort. I do not know why that is.

His worst play was one where,with 9:10 remaining in the first quarter, slides to move up to the second level, makes contact with Navarro Bowman before inexplicably releasing him, then turning around to have Adrian Peterson run into him. It's not that Peterson would have had a touchdown if he hadn't run into Johnson, but it IS Bowman who secures Peterson's ankle before Goldson can finish it off. Ponder and Rudolph bailed Johnson out on the 4th down try.

Underrated, but extremely professional center John Sullivan had some issues as well, although my issues with him are a bit more nitpicky, like his timing sliding off of defenders to move to the second level. He did not play as well in the run game as I'm used to, despite displaying some incredible strength for his size. His run blocking was largely OK until the last few drives, where he had 5 poor blocks in 10 plays. It might be easy to say he shouldn't be faulted for it, but anyone who remembers that last drive would also remember the anxiety about giving up the goat, even in situations where it should be mathematically impossible, especially with all those fumbles.

On the last 10 runs, the Vikings ran for gains of 2, 7, 9, 3, 2, 0, 0, 2, -1, 3. That's 2.7 yards per carry, and it's not just because the 49ers knew the run was coming. There weren't too many unaccounted for defenders on these runs. Still, if you took those out, the running backs averaged 3.3 yards per carry. The running backs didn't play poorly, and the run blocking was very poor throughout. It just so happens that Sullivan's poor run blocking came near the end of the game, when it mattered less (although one could argue that better protection of Gerhart may have helped prevent those fumbles).

There weren't as many problems in pass protection for Sullivan, but he did fail to pick up on a blitz that could have hit Christian Ponder fairly hard. I might be expressing too much disappointment here, Sullivan still probably had a better game than the average center. Just not as obviously much as many Vikings insiders have hyped him to be. He gave up significant run blocks with 9:10 remaining in the first (the Peterson run where he ran into Charlie Johnson)

The biggest disappointment of the line on Sunday, however, was Brandon Fusco. It was a shockingly bad performance. Not only did he allow at least 4 hurries that day, he was consistently missing his assignment on run blocks, refusing to move to the second level when the crashdown block is secured and having the honor of being "that guy" who keeps allowing a missed block to abort a run.

The best example of this is with 4:22 left in the first quarter, where Fusco pulls in order to lead a block and just missed Patrick Willis by a mile, which is curious, given that Willis is about the size of a bus. Had Fusco taken Willis straight on, Peterson would have been able to grab at least 4 yards instead of none. Look at the replay if you can, it's a nearly incompetent play. I noticed missed run blocking at least once on nearly every drive, but nothing more egregious than that. For a good "Ole" block, check out his "block" on Justin Smith when he pulls to block for Peterson. Peterson's speed around the edge is what gives him that yardage.

Phil Loadholt was not notable to me, which might be good for him or inattentiveness on my part. I noticed one occasion of poor run blocking and only one hurry, which was in the third drive. It came relatively easy for Ahmad Brooks, who got an outside release and put pressure on Ponder relatively early. It was an incompletion. His pass blocking was quite good otherwise.

For what it's worth, others have indicated that he had a good pass blocking game (something I agree with) but a very poor run blocking game. I'm not going to go so far as to say that the run game Loadholt had was as bad as Kalil's, but it wasn't great. Unsurprisingly, Loadholt has been inconsistent this year, following extremely poor pass blocking (against Indianapolis) with good pass blocking against a better pass rush.

Tight Ends

This one is a bit hard. Kyle Rudolph had a great game as a receiver. That's not really difficult to say. Two touchdowns, five catches, 36 yards. He was our second most effective receiver and had a highlight catch that Vikings fans hadn't seen the like of in some time. I'll talk about that in a second, but first I want to mention his blocking, which was notably bad.

We know that Rudolph is strong, but he's also had some technique problems in the run game and still blocks too high. He also doesn't lock in on his blockers, who can shed his blocks easily. On the third play from scrimmage, Ahmad Brook does just that when Rudolph doesn't or can't drive him forward and out of the play, and Brooks stops Peterson for no gain. There are a few other examples, like the first play of the fourth quarter, a frankly embarrassing effort that prevents a Peterson touchdown. Aldon Smith swats him aside and tackles Peterson just short of the goal line. Yes, Rudolph has a very impressive catch the very next play, but the third down try should have never happened.

He was also somewhat of a liability in pass protection, although this is less of an issue. He gives up one hurry when trying to pick up Aldon Smith on a pass. It's also not for the highlight reels. Suffice to say, Ponder has an incompletion with a linebacker nearly on top of him. Aldon Smith is a very good pass rusher and all around player, but almost any player could have done what he did on those two plays.

He was an effective receiving option, yes, but also consider that 1) he dropped a pass and 2) 15 other tight ends had more yards than him in Week 3. He's a great red zone target; a big body with a good catch radius, he has some fantastic hands to go along with his range and can certainly make plays. He's faster than ever before, but he still needs to work on his route running, as he seems to have trouble getting burst on his breaks. I like Rudolph, and he was a bit of a hero, but he was not a complete player by any means. He could be one of the worst run blocking tight ends in the NFL right now.

By the way, if you want a fantastically good breakdown of the second Rudolph touchdown, one of my favorite football educators, Matt Bowen, has you covered:

Chalk Talk: Ponder vs. the 49ers' Cover 0 (via MattBowen41)

John Carlson was once again absent to most eyes, and was targeted once for an incompletion, the result of well played coverage. Fans have become impatient with him, and are expecting him to contribute on the field in a way comensurate with his $25 million dollar, 5-year contract. While his salary for the year is "only" $2.9 million, his cap space is $4.25 million. Including defensive players, that's the 5th highest on the team. On offense, only Adrian Peterson consumes more cap space.

He hasn't caught a pass all year, and he's not really known as much of a run blocking end. This might be the worst front office decision this year, although it's not as if the Vikings don't have cap space to spare at the moment, anyway.

He was given significantly more snaps than in the past (both weeks earlier, he was in on 18 snaps), in on 29 of the team's 79 offensive snaps.

He was neither impressive while run blocking nor pass blocking, but he wasn't extraordinarily disappointing, either. While I've cautioned patience in the past for Carlson to adapt to this system and develop chemistry with Ponder. We're approaching a point where this chemistry should manifest onto the field. If we continue to see this lack of production into Week 5, it's very likely that Carlson is a bust.

Rhett Ellison was only on the field on 18 snaps, 3 of which were 3TE sets. He only made one mistake as far as I can tell, which was a missed block on the last functional play of the offense (before the three kneels). He was useful and did not make a mistake I noted otherwise.

Running Backs

I thought the running back performance exceeded their yard-per-carry averages. Adrian Peterson didn't grab the statistics that would make anybody admire his performance, but he still forced a few missed tackles by my judgment and outperformed his terrible run blocking. His shiftiness is good, but I still found at least two plays where he didn't follow his blocking, potentially preventing himself from making big gains.

His performance does not warrant the typical accolades, but neither does it deserve criticism. His blocking broke down far too often, but the star running back couldn't get things done to exceed his blocking. His 1.6 yards after contact is certainly nothing special, and was lower than average for Week 3 running backs; he also could not develop momentum. Overall, his performance was OK but many could reasonably expect him to do better.

Toby Gerhart, on the other hand, found himself to be much more impactful on fewer snaps. Unfortunately, that impact was clearly negative. Still, I would like to defend his first fumble, at least from a technical perspective. It shouldn't take too much to figure out why the refs were wrong to reverse the call for a few reasons.

The first reason is obvious: Harbaugh did not have that challenge, as per NFL rules. That does not need to be discussed. Second, the ball was not eligible to be recovered by San Francisco by the time Patrick Willis picked it up. Per the NFL rulebook:

Article 1: Dead Ball Declared. An official shall declare the ball dead and the down ended:

(a) when a runner is contacted by a defensive player and touches the ground with any part of his body other than his hands or feet. The ball is dead the instant the runner touches the ground. A runner touching the ground with his hands or feet while in the grasp of an opponent may continue to advance; or

(b) when a runner is held or otherwise restrained so that his forward progress ends; or

...

(n) when an official sounds his whistle while the ball is still in play, the ball becomes dead immediately;

(i) If the ball is in player possession, the team in possession may elect to put the ball in play where it has been declared dead or to replay the down.

(ii) If the ball is a loose ball resulting from a fumble, backward pass, or illegal forward pass, the team last in possession may elect to put the ball in play at the spot where possession was lost or to replay the down.

A.R. 7.1 Second-and-goal on B2. Runner A1 goes to the line of scrimmage where he is tackled and fumbles. The ball rolls into the end zone when the Referee inadvertently blows his whistle as the ball is loose in the end zone. Defense then falls on the ball.
Ruling: A's ball second-and-goal on B2 (inadvertent whistle).

At the moment of the whistle, a few things are clear. First, Toby is clearly not on the ground at the time of the whistle. The runner is not down by contact, but down by forward progress. The play and the ball, are dead. That should be it. Gerhart is entitled to let up on the ball at that point.

Second, Ahmad Brooks is trying to rip the ball loose for a fumble, but clearly has not succeeded by the first tweet of a whistle. Gerhart was in possession and the ball wasn't starting to come out. This means, even if the ball wasn't declared dead by the referees, that the down finished with the runner fully in possession of the ball.

Third, when the ball does come out, the original possessor of the ball (the Minnesota Vikings) has the right to the ball after it is declared dead. Finally, even if the ball was live and no whistles had blown until right before Patrick Willis picked the ball up, even an inadvertent whistle would have given the ball to Minnesota, as per subsection (n) of Rule 7 (Ball in Play, Dead Ball, Scrimmage), Section 2, Article 1. Approved ruling 7.1 clarifies this further and indicates that even if a legal fumble and clear recovery occurs, even an inadvertent whistle would have ruled the play dead and the ball in the offense's control.

Naturally, even if it was not true that the ball was clearly in Toby's possession as the whistle was blowing, there is no clear video evidence that the ball was ripped loose before or during the course of play, and therefore no basis for considering the ball "fumbled."

With that in mind, it still makes sense for the Vikings, who wouldn't think that Gerhart had butterfingers at that point and therefore should not be credited with a single fumble, to put him back in. After the first (real) fumble, there's also nothing to further indicate this is a chronic problem, because nobody had consistently ripped the ball out of his hands while he was trying to maintain possession.

After the second real fumble, but third credited fumble, there is no reason for him to be on the field, but it was merely kneeldown territory at that point.

Defending Gerhart does not mean I think he did a good job in the game. His blocking was worse than Peterson's, simply because it looks like the run blockers all gave up near the end of the game, but that is no excuse for two fumbles. I don't much mind Toby's yardage or average carry, nor do I mind his yards after contact—too many 49ers reached him too quickly for this metric to make sense in this specific case.

You simply cannot fumble that often with only 8 carries. Disappointing without too much to redeem it, but it is once again not his fault he was held to 2.3 yards a carry. He gathered two receptions for 20 yards, but he also dropped a pass that hit his hands. He'll want to forget this performance fast.

Jerome Felton did not match his opening performance as a fullback, and was not as effective as a lead blocker. Even if Peterson had followed him more often, Felton's impact on the game was far less than I had hoped for, having been a liability on three of his plays. The first was the quarterback hit he allowed in pass protection on the first drive, when Ahmad Brooks rushed almost right through him to get to Ponder. He didn't avail himself well in this case.

The second bad block was on the final touchdown drive, with 4:23 left in the third quarter. It was marked irrelevant because of a holding call (on apparently no one, as he did not announce the number). He hits Patrick Willis without getting his hands on the linebacker, and Willis ends up tackling Peterson. Felton blocks at the wrong angle, too, giving Willis easier access to the running lane than if Felton had shaded inside by six inches when hitting him.

The final poor performance as a blocker was with 8:05 left in the fourth quarter. Once again he hits Willis without getting his hands on the player. Once again, Willis sheds the block, ready to make the tackle. Had others on the play not also failed on their blocks, Willis would have aborted the play.

Like Fusco, Felton now is subject to questions about who the real player is. Should Vikings fans expect the Week 1 Felton or the Week 3 Felton? Given Ellison's rawness as a blocker right now, the Vikings' run games can ill afford an inconsistent lead blocker.

Offensive MVP: Christian Ponder
Unseen Offensive Player of the Game: Matt Kalil (again)
Offensive Rookie of the Game: N/A (again)
Honorable Mention: Percy Harvin