Guest column by Tim Murray
Ed. note: This article was very popular last April, and we've had so many people asking about it that we decided to run it again. You'll find the original article and discussion thread here [1]. These are all the same charts as last year; for updated charts with 2006 starters, you can buy the ESPN Magazine Draft Special. That's one of the many pieces of FO-related content. Not to sound like a salesman, but it's a really awesome combination of scouting from Scouts Inc. and stat analysis from FO, written for the hardcore football fan rather than the casual fan who reads the usual biweekly ESPN magazine. I'm awful proud of it. Anyway, enjoy Tim's article again. -- Aaron Schatz
Last April, Paul Tagliabue walked to the podium at Madison Square Garden and let the cat out of the bag. “With the tenth pick in the 2005 NFL Draft, the Detroit Lions select Mike Williams, wide receiver, University of Southern California.� Tagliabue is a pretty straight-laced guy, and a certain businesslike demeanor is probably a prerequisite for the Commissioner position, but you have to wonder how much Tags had to fight the urge to preface the pick announcement with, “You’re not going to believe this.�
It wasn’t the “Mike Williams� part of the announcement that was so surprising – Williams was generally regarded as a top five talent in the draft and draftnik extraordinaire Mel Kiper, Jr. had spent the first few hours of the draft on ESPN talking him up as the top talent. Rather, it was the “Wide Receiver� part of the announcement that got everyone so hot and bothered. The Lions were, and by all counts still are, an awful team with gaping holes at just about every position that had used top ten picks in each of the previous two years on the wide receiver position.
Several hours later, the New York Jets provided another surprising announcement from the MSG podium by using their initial pick, a second-rounder, on kicker Mike Nugent. There was somewhat less initial surprise at this pick than that of Williams, but it did provide material for roughly half of the jokes in Pro Football Prospectus 2005. Again, the question is not Mike Nugent’s ability as a place kicker, but rather whether using an early draft pick at that position was an efficient use of team building resources.
The central point in both instances was the position of the player being drafted. There is an oft-repeated draft cliché that could be called the “Best Available Player Theory.� If I had a nickel for every time I heard Charley Casserly explain his draft decisions while Redskins General Manager as “we picked the best guy on the board,� I might be able to afford the ridiculous price of parking at FedEx Field next year. And he’s just one of many GMs who use the standby line somewhat compulsively. The theory basically says: always pick the best player available, regardless of position. Picking the best player regardless of position was exactly what the Lions did, and exactly not what the Jets did. Both were extreme interpretations of the value of the Best Available Player Theory, and neither proved to be a wise decision.
Clearly there is a point between the two extremes where optimal adherence to the Best Available Player Theory exists. Having a remarkable collection of wideouts does you no good if there is no competent quarterback to throw to them or an offensive line capable of protecting the quarterback long enough for them to get open. On the other hand, grabbing a place kicker with your first pick because you believe that is your biggest need exhibits a complete disregard of the need for roster depth in the injury-laden NFL.
So how should GMs navigate the grey area in between? In practice, most seem to rely on their gut. The magnitude of influence the head coach wields is often a factor as well, typically pushing away from best available and toward greatest areas of need. The bottom line is that it’s probably a more subjective than objective process. And for us numbers geeks, that just won’t do.
One way to improve decision making would be to acquire knowledge of the “opportunity cost� (to borrow a term from economics) incurred by any team that chooses to ignore a position of need in order to take the best available player. More simply put, if a team’s positions of need can be addressed later in the draft (or even with an undrafted free agent) then the team can afford to go with the best available talent. In an effort to provide a cursory understanding of opportunity cost at each position I have compiled histograms of the round in which each starter on all 32 NFL teams was drafted. For each position we can see if certain positions can be addressed later in the draft than others, and identify where the real drop-off in talent seems to occur.
First, let’s look at the distribution of starters at all positions, so we have an idea of how individual positions differ from the norm:
|
|
The blue bars represent the percentage of league wide starters drafted in that round. The red bars represent the percentage of league wide starters drafted in that round that are still with the team that drafted them.
Both bars show a steady decline in the probability of players drafted in each subsequent round becoming starters, but with a somewhat unexpected rise between the sixth and seventh rounds. This is likely skewed by the fact that the NFL awards a lot of compensatory picks in the seventh round, so there are a few more seventh-rounders floating around than sixth-rounders. However, that shouldn’t be enough to cause an increase, so we can probably presume that there is little to no talent drop off between the sixth and seventh rounds. One might also presume that this trend continues past those players selected in round seven, which would mean that there is very little difference between sixth and seventh round picks and “priority� undrafted free agents. In other words, don’t get excited when your teams lands a sixth round pick in a trade, it is only a small upgrade over signing a guy that went undrafted.
Another point of note is that the drop off between round one and round two is significantly larger than that in any other round. Does this mean that there is an added premium between rounds one and two? It probably does, but this effect may also be skewed by the enormous contracts handed out to first round picks which often force teams to give starting roles to their top picks whether they have earned them or not.
All of this is mildly interesting and useful, but the real value to draft strategy lies in the positional analysis. The individual position histograms can generally be grouped into four categories: top heavy, normal distribution, early round peak, and flat. Based on these categories we can make some general assumptions as to when it becomes too late in the draft to confidently address an area of need.
(Note: click here [2] to download a list of the players considered at each position, and when they were drafted.)
“Top Heavy� are those positions which are almost entirely addressed in the early portion of the draft, mostly via the first round. The inference is that these are positions that need to be addressed very early in the draft. This group includes quarterback, running back, #1 wide receiver, defensive tackle, and offensive tackle.
Essentially what we see encompassed here are high-profile skill position players (QB, RB, and WR) and guys with exceptional athletic ability for their size (DT, OT). These are the types of players that are hard to miss when evaluating talent. Even a small college skill position player who puts up eye-popping statistics will grab the attention of scouts, so exceptional skill position players will rarely get overlooked. Those that do slide will 1) have durability (or “character�) issues that have kept them off the field, 2) be somewhat raw in terms of technique, or 3) have been misfit to their college offensive system. Meanwhile, the “Planet Theory� guys that you’ll find excelling at offensive and defensive tackle are also going to make extra large blips on scouts’ radar screens. (The Planet Theory is Bill Parcells’ philosophy that there are so few men both large enough and athletic enough to be NFL linemen that they are intrinsically valuable.) They may not have impressive statistics to shed light on their talents, but their “measurables� will make them very hard to miss.
 |
| Percentage with the team that drafted them:53% Percentage of Undrafted Free Agents (or 8+): 16% |
The quarterback position is especially top heavy. Exactly half of the starting QBs in the NFL (when this data was collected) were drafted in the first round. For whatever reason, the sixth round has been fairly lucrative, but pretty much anything after round one is a long shot. Interestingly, some of the better signal-callers have not been selected in the first four rounds: Tom Brady, Mark Brunell, Matt Hasselbeck, Marc Bulger, Trent Green, and Jake Delhomme. But only Brady has been productive for the team he was originally drafted by. (If the Saints had recognized the potential in either Bulger or Delhomme, they would have been a much better team the last few years. At least the Packers got some compensation for Brunell and Hasselbeck – and seemed to have a pretty productive guy at QB anyway.)
The QB position is by far the toughest to fill, and most teams are (and should be) willing to fill it by any means necessary. First round QBs are no sure thing, but anyone picked after the first round appears to be a very long shot that will require several years of development. And even those late round guys that do work out will probably have to move to a new team to be successful. The bottom line here is that until you’ve found your guy you should jump at any opportunity to acquire a quality starter, be it early in the draft, late in the draft, via trade, or through free agency. Don’t pass on a QB early in the draft because you also like a guy that could be had later. If you are really in need of a QB you should probably draft them both.
It’s less imperative to grab a runner in round one, but the well appears to dry up after round four. It doesn’t appear that you have to use a top pick to get one of the better runners, but it certainly improves your chances of success. Alternatively, you could just wait for the Broncos to get tired of the one they happen to have at the moment.
|
|
Wide receiver is the only position the NFL lists on depth charts with a corresponding number one and number two. All the other positions with multiple starters are designated via left or right. This allows us to differentiate between the premium players and the secondary one on each team (more or less, that is. There are a lot of number twos that would be number one on another team. And occasionally the designations seem a bit off – Terrell Owens was listed as number two by the Eagles in deference to Greg Lewis.) The histogram for WR1 is nearly identical to that of running backs. There’s not much available after round four, first-rounders have a significantly greater success rate, and stars can be had in all of the first four rounds. As is the case with running backs, if you have a need here you’d better plan on addressing it in round one or with multiple picks in the first four rounds.
|
|
The defensive tackle position is heavily weighted toward the first round, but fairly flat thereafter. This seems consistent with the Planet Theory explanation. The DTs with truly freakish abilities aren’t plentiful enough to occupy most of the 57 starting DT positions in the NFL (there are not two per team because of the many 3-4 alignments in use), so once the elite guys are taken at the top of the draft the talent pool evens out quite a bit. The late round picks and undrafted players are almost entirely one-dimensional players that can occupy multiple blockers but aren’t much of a threat to rush the passer. The lesson here appears to be that you’re not going to have many opportunities to find a two dimensional DT, so don’t pass on a guy like that lightly. Alternatively, if you are just looking for a big guy to occupy blockers on first and second down, that need can wait.
The offensive tackle histogram actually does not look particularly top-heavy, but a closer look reveals that most teams fill at least one of their tackle slots with an early round pick. Generally the top guy mans the left side, but that is not always the case. The key is that you need at least one guy on the edge who can be left in one-on-one situations and not need help. Those guys are rare, and are pretty much always picked early. Below are histograms which break down the position into “T1� and “T2�. This is not a reference to Schwarzenegger films or high speed internet connections. T1 is the starter for each team who was drafted earliest. T2 is the starter drafted later.
|
|
That’s quite a difference. So we see a dichotomy similar to that of the defensive tackle position. You’ll need at least one elite guy, and you are going to have to find him early. Another point of note is the high percentage of T1s that are with the team that originally drafted them. This is probably a function of the particularly high success rate of early picks used on OTs and the tendency for teams to hold on to them. So the opportunities to find a top OT outside of the early part of the draft are extremely limited. If you don’t already have a guy you can trust in one on one pass blocking situations, you’d better have a good reason to pass on one, no matter how early in the draft.
“Normal Distribution� describes those positions where the histogram follows the pattern similar to that of all positions in aggregate. Normal Distribution includes defensive ends, cornerbacks, linebackers, tight ends, and number two wide receivers. These are positions where it’s best to look for value. There are starting caliber players to be had at these positions throughout the draft, but there is definitely a drop off in success rate with each successive round.
|
|
At defensive end and cornerback you pretty much get what you pay for. The elite players are usually gone by the end of round two, but there are plenty useful players available after that. These positions are somewhat intertwined in that a good pass rush can help mask poor cover guys and vice versa. As a result, most teams tend to spend early picks on either one spot or the other according to their defensive philosophy. Teams that neglect both positions are those that either use a 3-4 alignment where linebackers and safeties provide most of the pressure (like Pittsburgh) or have a defense that routinely gets lit up like a Christmas tree (like Arizona and Green Bay).
(Note: R.Bartell and E.Hobbs have been fixed in cornerback graph.)
|
|
Linebacker is a position where there never seems to be a shortage of talent. It’s kind of the anti-Planet Theory position. I’m not saying there are 6-foot-2, 230 pound guys that can run a 4.5 on every street corner in America, but there seem to be 10 to 20 in every draft class. The elite guys are set apart by instinct, desire, and hard work. That’s not always easy to identify, so there will be some bargains late in the draft and even some that don’t get drafted. The gifted guys who were extremely productive at the big time programs are the ones that go early and seldom disappoint, so there’s no reason not use an early pick to fill a need at linebacker – just don’t do it at the expens of filling a need at a top heavy position. There will be good players available later if you know how to find them. OLB bargains are often undersized college defensive ends who make the transition to “playing in reverse� well (like Clark Haggans and Adalius Thomas). ILB generally requires more football intelligence than OLB, so the bargains are in the form of guys with relatively average athletic skills but great football minds (like Zach Tomas and Antonio Pierce).
|
|
Tight ends and number two wide receivers are usually the guys that keep defenses honest. Roll coverage toward Marvin Harrison and away from Reggie Wayne and Dallas Clark at your own risk. Top notch TEs and WR2s are luxury items. They’re great to have if you can afford them. Most teams can’t, because they’ve got bigger needs elsewhere and wisely choose to address this one later in the draft. Every so often a superstar TE like Tony Gonzalez or Jeremy Shockey comes along with the kind of ability that allows them to be focal points of an offense, so they’re worthy of an early to mid first round pick. Vernon Davis is probably one of these guys. But for the most part, TEs and prospective number two wideouts only slip into the back end of the first round when teams are looking to compliment an already strong roster.
“Early Round Peak� includes only safeties and centers. In both cases the number of starters is at a peak after the first round (2nd round for safeties, 3rd for centers) and then reverts to a fairly normal distribution. The most likely explanation for this is that the top athletes usually get pushed to other positions in college and high school, so there just aren’t many that are talented enough to warrant first round consideration.
|
|
Take a look at the former first round picks playing at safety. Other than Troy Vincent, who is a cornerback who moved to safety after slowing late in his career, you’ve got some pretty heavy hitters and exceptional ballhawks. But there have also been plenty of great players available deep into the draft. I think most of the guys with exceptional speed and the size to play safety get moved to wide receiver in high school and college, so the freakishly talented athletes are rare at this position. The type of guys that succeed here are similar to inside linebackers: guys with great football instincts that don’t shy away from collisions. Those qualities can be tough to identify, especially with a lack of statistics to help identify the standouts. This is a position that can be addressed late in the draft if need be, but you’ll be hard pressed to find a good one after the fifth round.
Centers taken in the first round don’t seem to be consistently that much better than guys nabbed in rounds two or three. There’s also plenty of solid guys who were available deep into the draft pool. Teams looking to address their offensive line early in the draft would probably be better off looking at the tackle position.
“Flat� histograms are found at guard, fullback, and special teams. These are the positions where starters have been found with similar success in all rounds of the draft as well as via undrafted free agents. These are the positions where teams would clearly be better off filing needs with a training camp battle royale of sixth and seventh round picks and UFAs.
|
|
The guard position has produced a number of quality players in round one like Steve Hutchinson and Alan Faneca. Teams with power running games probably need at least one power blocking guard they can run behind, which they’ll probably need to find in the first three rounds. The successful zone blocking teams (Denver, Atlanta, Indianapolis), on the other hand, seem able to find solid performers amongst guys picked in round five or later.
Fullback is a position where there’s just no reason to use an early round pick. The histogram is actually somewhat misleading, as most of the early picks are guys who are not true fullbacks. Rather, they are either H-Backs (Chris Cooley), players used often as featured runners (Greg Jones, Mike Alstott), or players who may be featured at fullback on the official team depth chart because the head coach has been joining them for recreational activities (Ricky Williams).
|
|
Special teams are important. It is an area which teams cannot ignore and expect to be successful. But it should never be treated as more than an afterthought in the draft. Neither quality return men nor quality kicking specialists are more likely to be found earlier in the draft than late. In fact, most of the better special teamers do not find success with their original employer. If you need to improve your special teams, find a good special teams coach and pay close attention to the waiver wire. On draft day: fuggetaboutit.
So what does this all mean when April 29 rolls around? First, let me point out that this is nothing more than a snapshot of history. 832 data points represents a lot of information, but it's still relatively small as sample sizes go. As further disclaimer, the past is not a perfect predictor of the future, although it's usually the best information we've got. Then again, a few years ago there were no data points telling us to go look for the next great tight end on Kent State's basketball team.
With that said, I believe there are some draft strategy guidelines that can be gleaned from the data:
1) Address your needs at "Top-Heavy" positions first and foremost, particularly QB, DT, and T1.
2) Look for value at the "Normal Distribution" spots. Many of the best value picks come from guys at these positions who slide into rounds two through four.
3) Immensely talented safeties are rare, but you won't have trouble finding a quality safety in rounds two through five. A talented RB like DeAngelo Williams is probably a better use of resources than someone like Donte Whitner if you've got a need for both.
4) Wait until at least the end of round one to fill a need at center. Nick Mangold is awfully talented, but not worth a top pick if you've got needs elsewhere.
5) Spend late-round picks and undrafted free agent bonuses to collect prospects on offensive line, one dimensional defensive tackles, fullbacks, and special teams. A few of these will work out, and allow you to use your more valuable resources elsewhere.
At the end of the day, every team has pretty much the same arsenal of resources with which to build their roster (although some may have a little more cash for signing bonuses than others). The teams which can manage to use those resources most economically will ultimately field the most talented teams. It's not just about player evaluation; it's also about knowing how and when to address your specific needs.
Tim Murray works for an investment advisory firm in Bethesda, Maryland. He's been an ardent NFL draft observer for the past fifteen years. If you have an idea for a guest column, something that analyzes the NFL from a distinctive point of view, please email us at info@footballoutsiders.com.
Links:
[1] http://www.footballoutsiders.com/2006/04/24/ramblings/nfl-draft/3828/
[2] http://www.footballoutsiders.com/murrayappendix.doc
