Eight weeks ago, the Vikings were undefeated and ranked No. 1 in DVOA. Now they're 6-6 and probably done. What went wrong?
24 Apr 2013
by Matt Waldman
When it comes to evaluating football talent, I believe analytics can be a part of the conversation. The Playmaker Score is a good example.
I think the Playmaker Score is a nice attempt at data mining where Vince Verhei has reverse-engineered a formula with the hope of developing a successful model to predict future results. The score does enough to have value in discussions about specific wide receiver prospects, but Playmaker is not the entire conversation.
I believe player evaluation can be likened to a three-legged, wooden table:
If one of these legs is missing or not factored into an evaluation enough, the table doesn’t function as it should. The fourth factor is character, but I think it’s best considered as liquid -– either it enhances, like a wood stain, or degrades, like water spot.
Scouting is such a hit-or-miss process because no one has figured out a way to consistently make all three legs of the table functional for every player evaluation. Nor can they predict if the player’s character will enhance or ruin the "table."
The Playmaker Score has a grim outlook for this group of wide receivers. In contrast, many scouts and analysts tout the position as deep and talented. Before going any further, here is Vince’s overview of Playmaker:
When we first devised our Playmaker Score projection system to predict NFL success for wide receivers, we looked at individual collegiate production only. A second version added team statistics to account for players who might have seen their numbers inflated by prolific passing offenses. This year, with Playmaker Score 3.0, we've added Combine data for the first time. Now we're measuring not just football skills, but raw physical talent.
We've made one other fundamental change to Playmaker Score, and it involves the way collegiate data is handled. In the past, using career totals considerably underrated those players who were so spectacular that they skipped their senior seasons and entered the draft early, while using per-game numbers depressed the ratings for guys who had lots of one- and zero-catch games as freshmen and sophomores before exploding as juniors and seniors. We tried to get around this by using a mix of career totals and per-game data, but the results were a little confusing and somewhat illogical. So we've gone back to the drawing board, and we're now using the numbers from each player's best season. This makes the most sense because it rewards the biggest stars at the expense of more mediocre players who pad their statistics with multiple starting seasons. Obviously, there's now a danger of overestimating one-year wonders, but we're working on some methods to correct for that in the future.
We checked the numbers for every receiver drafted from 2005 to 2009, a group of 149 players. That gave us five years' worth of recent history, while giving every player at least three years to break out in the NFL. For each player, we determined their NFL success by dividing their career receiving yardage by the number of seasons that had passed since, whether they were still in the league or not. We also compared their Combine numbers to their NFL statistics and checked which were most accurate when it came to predicting NFL success.
Playmaker Score addresses the basic question of how good the quarterback is when it comes to giving his receiver opportunities for production. However, I think zeroing in on passes per game as a factor makes Playmaker Score a slave to a specific kind of receiver production that doesn’t tell the entire story.
Playmaker does not fully address all the dimensions of production that can make a wide receiver a playmaker. Yards after catch is a significant example, although not a purposeful oversight. The staff here at Football Outsiders would love to have this data, but the college game does not supply it in an easily available format.
However, this is part of the reason why I think the revised 2011 Playmaker Scores missed on A.J. Green (227 points) and Randall Cobb (136 points), but it liked Jonathan Baldwin (464 points) and Torrey Smith (448 points). I think Baldwin and Smith played in offenses that were suited to what Playmaker Score rewards players for: red-zone production and longer plays from pitch-to-catch. It doesn't factor in ball carrying of any sort, and that's part of what it missed about Green and Cobb.
I liked Smith’s prospects, but I wasn’t high on Baldwin. One of the reasons reflects what is missing from Playmaker: it is a formula based on production and athleticism (as measured by certain sprints and jumps, at least) but it doesn’t account for technical skill.
However, I’m jumping the gun on technique. The athleticism factor also requires more discussion.
The way Playmaker uses Combine data is a problem endemic to the area of football evaluation. I believe the primary function of the Combine should first be to determine which players have the baseline speed, strength, quickness, and size to perform at an NFL level. Instead, there is too much emphasis placed on stronger-faster-higher-longer.
Playmaker’s inclusion of Combine measurements assumes that the better (certain) Combine numbers, the better the player. I think most people make this assumption, and it’s easy to do: Some people believe Calvin Johnson is a better player than Dez Bryant because he jumps higher, runs faster, and does it in a bigger body.
Athleticism is a baseline requirement, but there is a point where a player’s value can be inaccurately inflated or depressed because technical skill or capacity to learn the game isn't properly accounted for. To give examples, we have A.J. Green on one hand, and Robert Meachem on the other.
Green had actually scored high in the previous version of Playmaker (v2.0) but does not fare well in the current version because of the Combine metrics. His NFL Combine metrics were NFL-quality, but not stellar. Yet what the formula misses, game observation catches: aplayer who gets the most from his athleticism because he integrates his physical and technical skill sets at a high level to make plays.
Meachem, on the other hand, was a fast prospect, but he couldn’t catch the ball with his hands away from his body. I admired his efforts as a senior at Tennessee to extend his arms to the football, but he often had to revert to a trap technique.
One example I recall was a game in Tennessee against LSU. Meachem dropped three passes in the first quarter alone while trying to use proper hands technique. He then reverted to a trap technique for the rest of the game.
While I admired Meachem for trying, some technical skills are more difficult to learn at this stage of football than others. Receivers who don’t make receptions with their hands away from their frames are unlikely to acquire these skills when they transition to the pro game.
This issue is what I would term a fatal error. Difficulty securing the ball after contact from a defender during the act of the reception can also be a fatal error when it comes to evaluating the technical skill potential of a prospect.
Although Meachem eventually made some strides to improve this skill, he has never been comfortable enough to make a complete transition. He has not become the impact player commensurate with his first-round grade. These two flaws were why I didn’t like Meachem’s chances of doing that, and he proved me right.
As for other examples from 2011, I loved Green (my No. 1 rated receiver in 2011), Randall Cobb (No. 3) and Torrey Smith (No. 5). In contrast, Jonathan Baldwin (No.13) was an overrated commodity in my book. Green, Cobb and Smith demonstrated excellent skill versus contact and both Green and Cobb were strong players after the catch. Baldwin’s hands techniques were spotty and his technique and conceptual execution of routes were inconsistentat best.
I think a better way to view measures of athleticism is to, as I said, first assess if it meets the baseline requirements to play in the NFL. Then assess if that player also meets baselines for skill level/capacity to learn the skills.
Once an evaluator establishes that the prospect has the basic tools, I think the fast-stronger-longer-higher quality of the Combine measurements become a refining characteristic to the evaluation process. The same goes for productivity metrics.
The problem is that we want to throw these tools into the evaluation process at a premature stage. As a result, the data can often overestimate (or underestimate) the value of players.
Cordarrelle Patterson is a good example. Playmaker has a low score for the Tennessee wide receiver because it is production-driven and its value of athleticism is based on bigger-faster-better rather than using the metrics as a baseline to determine whether "he can hang in the NFL."
Playmaker bases Patterson’s production on one season at Tennessee (46 receptions for 778 yards and five touchdowns) -– and only his receiving stats. In contrast, Patterson’s two seasons as a junior college player at Hutchinson yielded 113 catches, 1832 yards, and 24 touchdowns. Playmaker -– right or wrong –- dismisses the value of Patterson’s JUCO numbers. I believe production is a reflection of how well a player executes his role in an offense. It is not how good of an overall receiver he is compared to other receivers with different roles in their respective offenses.
If you rely solely on Playmaker to determine a wide receiver’s future success, then you will wonder how Patterson can be so highly rated when he wasn’t that productive. Of course, this is a one-dimensional viewpoint (receiving yards) of a multi-dimensional player from scrimmage who had 71 touches, 1086 yards, and eight touchdowns.
I could also open another can of worms with Patterson’s 28 returns for 772 yards and two touchdowns, but let’s set aside the argument that one might have about the merits of earning yards as a return specialist versus carrying the ball at the line of scrimmage.
The reasons I see Patterson as a big-time prospect, despite the fact that he didn’t produce to that merit in the singular role of pass catcher, are the following:
These factors underscore a critical point about Patterson that is a reflection of Playmaker’s low score –- the Volunteers wideout is one of the biggest boom-or-bust prospects in this draft class. However, I also think there’s enough reward to place Patterson among the top prospects in this wide receiver class.
There is a belief that Patterson isn’t good because he didn’t produce against quality teams. However, the way one defines ‘quality’ or ‘production’ is an important consideration.
Here’s how the 2012 Tennessee passing game performed in terms of distribution percentages among its top targets.
It's worth noting that this was Patterson’s first year on the team and he accounted for the second-most receptions and yards, while tying for third among receivers in touchdowns. While this doesn’t account for missed opportunities, Patterson’s share of production isn’t bad for a new receiver. This is especially good for a player breaking into a lineup with a quarterback who has worked extensively with slot receiver Zach Rogers, split end Justin Hunter, and tight end Mychal Rivera for four years (as well as running backs Rajon Neal and Marlin Lane for at least two).
Rogers, Hunter, and Rivera are quality talents and strong red-zone players. Counting on receiving touchdowns as a vital metric for projecting Patterson’s NFL success might lead one to underestimate the receiver’s capability as a playmaker.
Hunter and Patterson also provide this team with two dangerous athletes outside, which means there will be plenty of situations where the coverage will dictate the quarterback must make the inside receivers the primary targets. When that team has a slot receiver and tight end the caliber of Rogers and Lane, the quarterback is encouraged to work inside.
And when it comes to the subject of Tennessee's quarterback, Tyler Bray has been unpredictable in a bad way. There are many examples of Tyler Bray ignoring good looks in the passing game for risky options because he was overconfident in his arm strength and accuracy.
Bray’s average production per game is a telling indicator on the surface of this deeper point: 22 completions, 301 yards, and 2.8 passing touchdowns. Bray had at least 250 yards in nine of twelve games in 2012, and 300-yard performances in six of those nine games, but his completion percentage was 58 percent.
Bray completed less than 55 percent of his passes against Florida, Georgia, Mississippi State, Alabama, and Vanderbilt. When he completed at least 60 percent of his passes, the opponents were Georgia State, Akron, Troy, Missouri, N.C. State, and South Carolina.
It isn't all on Bray. Patterson and Hunter had their share of drops -- some of the most untimely and damaging variety -- but Hunter and Patterson's carelessness and lapses of concentration were overshadowed by Bray's poor decisions. More often than not, it was Bray opting to target his receivers on high-risk plays due to poor reads and decisions. It’s the hallmark of Bray –- a physically and technically talented, but conceptually flawed quarterback. I encourage you to watch Bray’s appearance on Jon Gruden’s Quarterback Camp for examples.
Neither Rogers nor Hunter carried the football in this offense, but Patterson averaged two rushing attempts per game. He generated another 300 yards on 25 carries and three touchdowns to his receiving totals on the ground.
My colleague Nick Whalen studied Patterson’s elusiveness this season for a piece he wrote for the RSP blog. Whalen defines the applicable situations where the players in this study had a chance to make a defender miss.
Of Patterson’s 25 carries, Whalen counted 17 attempts where the receiver had an opportunity to make the defenders miss and found that Patterson eluded 23 players on those runs. As a point of comparison, Tavon Austin eluded 20 players on 29 eligible runs.
As with Cobb and Percy Harvin (another low scorer in the Playmaker scoring system), Patterson and Austin were versatile playmakers as runners or receivers after the catch, not just specialists in the exclusive act of making the deep reception like Baldwin or even Torrey Smith.
Here’s Patterson as an I-back on a pitch against Mississippi State.
This is a well-blocked play at the edge. The Volunteers get a push downfield. However, Patterson does his share of work when the Mississippi State safety crashes the flat and forces Patterson inside before the receiver crosses the line of scrimmage.
Freeze the video at the 4:00-mark. Most running backs in this situation would have difficulty avoiding this unblocked interior defender coming from backside pursuit. However, Patterson cuts inside the safety and maintains a course downhill without a strong cutback into the pursuit.
Patterson slides between No. 90 and the flat defender at the Bulldogs 29 for another 12 yards and the first down. He finishes by lowering his pads into the contact, which is another technically sound thing.
This is a mature piece of running for a player often over-characterized as a stop-start player who won’t get away with his flights of fancy in the NFL. I agree that Patterson may have to tame that stop-start style of movement he uses on reversals of field and certain cutbacks, but it’s far from the only tool in his bag.
Here’s a play against Georgia where Patterson demonstrates a strong downhill mentality as a runner on a reverse option pass. I like the football intelligence to wisely tucks the ball rather than throw it -- something he has done with success at Tennessee. Patterson turns this play into a 46-yard touchdown.
Once again, the Volunteers have a man-on-man blocking advantage at the edge. Bray actually gets into the act with a nice cut block on the pursuit.
Patterson cuts this play inside the block to reach the line of scrimmage and reads Alec Ogletree’s pursuit to the outside. He cuts behind it with enough quickness to leave the first-round linebacker prospect flat-footed.
Patterson then dips across the field and outruns the secondary to the opposite edge, picks up a block, and outruns the rest of the field for the score. His rate of elusiveness is even larger on pass plays. Patterson had 16 receptions with opportunities to elude a defender and made 27 of those defenders miss. Austin made 25 opponents miss on 44 receptions.
It’s easy to assume that Patterson makes a vast majority of these plays as a glorified running back on dump-offs, but just as it was with Cobb and Harvin, it’s an inaccurate assessment. Cobb, Harvin, and Patterson will make a lot of hay on dump-offs, but their ability to turn intermediate passes into longer plays is what separates them from former high school running backs placed in the slot in a college spread offense and fed short passes for a lot of production.
Here is an intermediate cross where Patterson catches the ball in stride 15 yards downfield for the first down, and then makes multiple Missouri defenders miss for additional yards.
It’s the combination of quickness and strength that makes Patterson so dangerous. The receiver does a good job squaring his body to the target on this 13-yard cross, then breaks a tackle and bursts past another defender for another six yards with economy.
Against teams like Troy and Georgia State, Patterson may attempt multiple stop-start moves and reversals of field, but he still demonstrates a strong amount of selectivity when it comes to his decision-making according to the situation.
Even this play is a good demonstration of a downhill mentality that will earn him a lot of yards after the catch in the NFL. There are no stop-start moves; only cuts, dips, and spin moves while working forward.
The reason Patterson is such a wild card in this draft is that he’s considered a raw receiver. He drops passes, his routes aren’t as polished as other prospects, and if you listen to sources who talk to The Sideline View’s Lance Zierlein, Patterson needs a lot of reps to learn his assignments.
However, Patterson is adept at the fundamental aspects of NFL receiving that makes or breaks other pass catchers with better route technique. Two things that I think "makes" Patterson a viable talent are the same things that "broke" Meachem in my estimation: Catching the ball away from his body and making plays after contact.
Here’s a solid turn and adjustment to the ball on a back-shoulder fade:
Patterson telegraphs his turn and this will need to be addressed, but he extends his arms and catches the ball with his hands while staying in-bounds. These are two things Patterson demonstrated consistently.
This reception on a slant was a nice adjustment on the run as well.
Patterson extends his arms with his back to the quarterback and secures the ball over his shoulder while the defender wraps him.
I like that he’s also willing to attack the ball in tight coverage.
Patterson does a good job maintaining position with his back to the defender, coming back to the throw to keep his back between the defender and the ball while high-pointing it, securing the catch, and maintaining possession as he’s wrapped.
Patterson’s catches on slants in tight coverage are close to his body, but he still makes these plays with his hands. Even on wide-open crossing routes, I rarely see Patterson trap the ball.
Just as Patterson is comfortable with physical play as a runner in space, I see signs of similar comfort when the ball is in the air.
This adjustment away from the Alabama defender’s reach is a demonstration of a strong catch radius from an awkward position in tight coverage.
Here is another strong play on a sideline route in tight coverage against Florida.
Patterson does a good job of snatching the ball while leaning over the boundary.
I think all of these plays are good examples that counter the drops you'll see in the low-light packages that will be shown for Patterson during the draft. Patterson’s drops tend to be issues of concentration rather than difficulty using his hands. Most of the receiver's egregious drops are wide-open passes where he turns his eyes upfield before he secures the ball.
In other words, Patterson’s flaws are non-fatal errors. But for a player with the athleticism and basic receiving skills to earn first-round consideration, there are enough non-fatal errors with technique that he presents more risk than some teams will feel comfortable taking. At least in the first round, anyway.
However, the only potential issue that he has that could become a fatal error is his capacity to learn the game, and that’s an issue I don’t know enough about. What I have seen on the field is a wide receiver that runs good enough routes to get open on slants, hitches, crosses, and fades, and has room to get better.
Combine Patterson's capacity to become a better route runner with his uncommon skill to integrate strength, quickness, and vision as a ball carrier, then add his ability to catch the ball with his hands, and I see a promising wide receiver. Go one step further and add the fact that Patterson can make receptions against physical coverage, and I see a potential playmaker of the highest order.
Matt Waldman authors the Rookie Scouting Portfolio, an online publication devoted to play-by-play analysis of offensive skill position prospects. Now in its eighth year of publication, the RSP provides pre- and post-draft rankings, analysis, and stylistic comparisons. Waldman also includes all of his play-by-play notes, grading reports, and a glossary that defines and explains his grading criteria. The RSP is available for download now and 10 percent of each sale is donated to Darkness to Light, a non-profit devoted to preventing sexual abuse in communities across the country.
7 comments, Last at 11 Mar 2016, 6:36am by footbluefox