One thing that gets lost in the discussion of draft strategy is the landscape of the draft. Very often, teams are tasked by fans of either drafting the “best player available” regardless of need (recently, the Jets’ decision to draft Leonard Williams is held up as an example), or drafting players at positions of need with a looser definition of board value (basically, half the time a team drafts a quarterback in the first round).
The truth of what teams do is of course closer to the middle than what fans think (or what teams say)—conversations I’ve had with multiple agents and team personnel confirm that teams largely draft players at positions that need to be filled within the current three-year window (putting the Jets’ Williams pick in more context, given the contract situations for Muhammad Wilkerson and Sheldon Richardson). Even in the middle-to-late rounds of the draft, teams draft players at positions that need to be filled that year.
Most teams only have boards of 75-150 players and the Patriots last year drafted a long snapper in the fifth round because all of the players on their draft board had been drafted. Teams really can only do this if they are paying very specific attention to the specific traits for their team and to a big extent, what they need (this should also give you context for what it means when an anonymous scout says “he’s off our board”).
Josh Norris at Rotoworld mentioned that teams very often keep scouting players that are not going to be on their board simply so they have the lay of the land; teams care a lot about adding players they need.
None of this is to say that teams do not pay attention to which players are the “best” or if there is a big talent dropoff, but it’s good to keep in mind the next time we debate “BPA vs. Need”—teams try to draft the “best player at a position of relative need”.
But there is one thing to keep in mind that general managers are constantly aware of: scarcity. If there are very few good players at one position and a number of good players at another position, it may help a team more to select the inferior player at a position with fewer good players than a better player at a plentiful position.
Let’s expand Jacob’s example with a two-round “draft” of wide receivers and linebackers. There are ten receivers, with ratings of 90, 87, 86, 85, 84, 81, 79, 78, 75 and 72 and eight linebackers with ratings of 85, 83, 81, 76, 71, 69, 67 and 65.
If you have the option of picking any of those players with your mid-first round pick and you picked the best player, you’d get a great receiver—a 90. But you also know that 4-5 receivers will be drafted by the time you pick again (as has been the average in the last ten drafts) in the middle of the second round, and only four linebackers will be selected by that same moment.
If you selected the 90-rated receiver, the four top linebackers will have been selected and you’re left with picking between linebackers rated 71, 67 and 65. The team therefore added two players whose ratings add up to 161.
On the other hand, if you selected a linebacker who clearly was not quite as good in the first round (rated 85) and the top five receivers are off the board by the time you pick again, you can pick up the 81-rated receiver and get a total rating of 166.
If you throw in safeties, perhaps rated 79, 60, 58, 55, 51, 47 and a third round, you approach something you might recognize from grueling calc sessions: an optimization problem. Your choices look like this:
Round | WR | LB | S |
1 | 90 | 85 | 79 |
2 | 81 | 71 | 55 |
3 | 61 | 65 | 47 |
If you pick the best player available in each round (and only selected one position), you would end up with a 90-rated receiver, a 71-rated linebacker and a 47-rated safety for a total of 208.
On the other hand, if you take into account value in later rounds, you could draft in the order of safety-receiver-linebacker and get a total of 225 (79-rated safety, 81-rated receiver and 65-rated linebacker). You might improve their team by 17 points by selecting a player rated 11 points worse in the first round!
Now, there are many more moving parts than this (not just more positions and rounds, but the value of the players being replaced and the value of the positions they play) and summing Madden-style ratings don’t capture a lot of effects, like the fact that a 90 player and a 80 player together probably are better than two 85 players. But the point remains the same, only the values change.
We shouldn’t expect general managers to be able to do what ends up being really complex calculations to this effect, but there is one thing that gives us a shortcut to all of this: talent cliffs.
The reason it was so easy to construct the above scenario is because I created a cliff of talent for safeties in the first round and a small cliff in the third round for receivers. If there are talent cliffs in the draft, are there particular spots NFL teams should target specific positions?
On Tuesday, we looked at which positions groups in the draft were the deepest, which will have a lot of the same work we’ll cover below. Yesterday we looked at which position may be the most difficult to evaluate. Both of those pieces draw upon my work to collate 40 different analyst boards and create a consensus ranking of players.
One thing it can also do is identify where the talent cliffs are. We saw something like it when looking at the talent tiers from the depth piece, but those tiers are a little too nuanced for work designed to project 256 picks in the draft. It may be better to look at larger groups of talent to get a better handle of where to draft players at each position.
We’ll expand the groups so we still have ten tiers, and look at it by dividing the number of positional players in each tier by the total number of players in each tier (the final tier is 70 players, so having four players in the final tier is not the same as having four players in the first tier):
If you are a team with a quarterback who does not need a new one and you took the step to combine the two interior offensive line positions, you could conceivably pick one player from each tier. If it was done in the principle of avoiding cliffs, you could go:
- Running Back
- Off-Ball Linebacker
- Cornerback
- Wide Receiver
- Safety
- Interior Defender
- Tight End
- Guard/Center
- Edge Defender
- Tackle
More realistically, you could imagine a team like the Vikings, who have eight picks—one in the third tier, one in the the fourth tier, one in the seventh tier, two in the eighth tier, one in the ninth tier and two in the tenth tier.
Say the Vikings want to find starters at wide receiver, linebacker and safety, as well as long-term developmental players at tight end, offensive tackle, interior offensive linemen and interior defensive linemen as well as backups at quarterback and edge defender.
The Vikings would be in a tough spot because the linebacker talent never really recovers… unless we do what the Vikings did with Anthony Barr and consider Georgia edge defenders Leonard Floyd and Jordan Jenkins as off-ball linebackers. Then you would get the following mock draft:
Pick 23: Leonard Floyd, Linebacker, Georgia
https://www.youtube.com/watch?v=Lu4aqlerDgc
Leonard Floyd might be an unusual selection given that the Vikings aren’t looking for a pass rusher in the first round, but he would be an excellent selection for the Vikings for the same reason that Anthony Barr was two years ago—he’d play as a versatile linebacker, not an edge threat.
Not every undersized edge player can convert to off-ball linebacker like Anthony Barr or Von Miller did, but two of the critical components necessary for that transition are fluidity and instincts. Floyd demonstrates both, and has the kind of mental acuity that not only allows him to play quickly, but to be the kind of versatile chess piece Georgia needed him to be. I recently mocked Floyd to the Vikings for Inside the Pylon, and you can listen to the analysis below:
Pick 54: Leonte Carroo, Wide Receiver, Rutgers
No one describes Leonte Carroo better than Matt Waldman does in his excellent Rookie Scouting Portfolio:
If a mad scientist created a machine that performed mashups of humans and he kidnapped Treadwell, Doctson and Coleman to stick them in one end of the contraption, Leonte Carroo would come bounding out the other side. He’s my favorite receiver that everyone knows, but no one talks about as a headliner at this position.
While perhaps not quite as athletic as Doctson or Coleman, Carroo excels after the catch, before the catch and during the catch as a route-runner, physical receiver and bully with the ball. Carroo has underrated flexibility and a true catch radius larger than most receivers because of his ability to adjust to the ball and catch passes thrown behind him, low or honestly even to opposing defensive backs.
He shows quickness, toughness and the ability to run routes well. Many project him to be available in the third round—and that may be the case—but not where the Vikings pick, and it wouldn’t be a waste to use a second on him. His releases need quite a bit of work, but it may be a coaching error.
Pick 86: Sean Davis, Safety, Maryland
We mocked Sean Davis to the Vikings last time, and it was in part because of a crazy judgment by our availability model, Drafttek. This time we’re using the consensus board’s overall ranking and he projects as a third-round pick. We’ll quote the scouting report from the last mock, but one thing to note: it may be interesting to unite Sean Davis with Leonte Carroo after Carroo did this to him.
Davis has had a number of struggles in man coverage against quick receivers as a cornerback, but has played extremely well at safety for Maryland—manning up against tight ends just fine, and playing the short and deep zones very well. He’s aggressive and has some ball skills, while also showing pop in the run game—hitting harder than his teammates at linebacker.
He and Harrison Smith should have a time forcing fumbles; Davis forced five this year alone and led the team in solo tackles without too many of them coming in coverage as a corner.
He does have issues with run angles and wrapping up, but has also flashed good technique, too. He’ll likely have to tone it down in the NFL, but his attitude, athleticism, smarts and raw skills should be enticing for the Vikings.
Pick 121: Willie Henry, Defensive Tackle, Michigan
There’s a lot of variance for Henry, with some boards—like Corey Chavous’ at DraftNasty putting him at #27—while others—like NDT Scouting‘s Kyle Crabbs’ board—doesn’t rank him in a group of 300 players.
The thing I like about him in a Mike Zimmer defense is that he could gain weight to be impactful as a run stuffer or lose weight to be like a prototypical three-technique, like how Zimmer asked Sharrif Floyd to drop weight, or how he didn’t ask Geno Atkins to gain weight in Cincinnati.
He was a versatile player for the Wolverines, but in Minnesota will likely feature as a back three-technique and rotation nickel-down pass-rusher. He needs technical work and occasionally lacks awareness in the run game, but is generally pretty athletic and long while maintaining his fundamental assignments very well.
Pick 160: Isaac Seumalo, Offensive Guard, Oregon State
Isaac Seumalo really doesn’t get enough credit by most evaluators. Last season he played both tackle and guard and he has center capability as well, having played all five line positions for Oregon State throughout his career.
Per Pro Football Focus, he only gave up four pressures all season despite playing outside and inside as an offensive lineman, and ranks second in their pass-blocking efficiency metric (ninth in run-blocking success rate) among draft-eligible offensive linemen.
Seumalo accomplishes this with a combination of high-level awareness, active arms and excellent coordination. He’s smart, balanced and plays with good technique, and he shows good movement in space.
Unfortunately, size concerns have led to some analysts calling him “center-only,” and those size concerns are paired with some issues with power. He’s not the most athletic specimen at guard and he may even be an unlikely fit in Tony Sparano’s new scheme, but he’s a good player and good players play.
Pick 180: Dean Lowry, Defensive End, Northwestern
Lowry is one of the best athletes in this year’s draft and the fact that the consensus board lists him so low is a surprise in my eyes. He requires a lot of technical work, but there are times in his game that people assume he can’t get off the ball as a result of his assignment, rather than his natural ability. He’s asked to hold his lane despite his one-gap responsibility and could transition to a 3-4 defensive end very well but for the Vikings will be a hybrid defensive end/tackle.
There’s technical work to clean up for Lowry, but the instincts and athleticism are there. He had 20 passes deflected at the line of scrimmage throughout his career and has shown the mobility to drop into space on stunts. His length is a big concern, as are his limited array of pass-rush moves, but the potential is there for him to develop into a supersub along the line.
Pick 240: Brandon Doughty, Quarterback, Western Kentucky
Fans and teams often want to draft with the philosophy of only drafting players that can start or be developed into starters, and that’s a great philosophy in general, but it would mean passing on valuable backups like Brandon Doughty, who I wouldn’t expect to start but shows a lot of mental acumen as well as physical ability.
Most of the time, when we talk about quarterbacks who are students of the game, we think of game managers who avoid risk. Not Doughty, who moves through progressions faster than most quarterbacks in this class, but also loves to grab big yardage plays. With a big arm and quick decisionmaking, Doughty can fit tight windows.
Statisically, Doughty is one of the most accurate passers in the class (second-best accuracy rate in the class, per Pro Football Focus), one of the most consistent under pressure (again, second in accuracy under pressure, per PFF) and one who can help his offensive line (with the lowest sack rates under pressure in the class). He has the fifth-best accuracy on the deep ball as well.
Those statistics don’t quite translate to the film; he has footwork issues that impact his ball placement, and he needs his receivers to adjust to the ball. Though he isn’t fazed by pressure he needs to do a better job climbing the pocket in response to it. Doughty is one of the worst athletes in the class, if not the worst, and presents little running or play-extending upside. At 24, there’s not much additional development room for him either.
Still, he should be better than some backup quarterbacks and he and Taylor Heinicke should be able to compete for the second or third quarterback spot.
Pick 244: Beau Sandland, TE Montana State
Beau Sandland has drawn comparisons to Kyle Rudolph, so the idea that he might fit with the Vikings isn’t a particularly novel one. He’s an athletic tight end in a class without many of them, and his jump scores in particular stand out for someone of his height and weight.
He’s moved around a lot in his college career, from a JuCo player to Miami to Montana State, and has only one year of starting experience, but looked pretty good playing for the FCS underdogs. He will need technical work as a blocker and route-runner, but is a natural receiver with the athleticism to line up anywhere. He could push the Vikings tight end corps and perhaps prepare for a future without Kyle Rudolph and his hefty contract.