The NFL Draft is a special event. It provides teams with an avenue to add cheap talent to their roster. A good draft class can potentially provide enough talent to propel a team over the hump and get them one of those Super Bowls everybody keeps talking about. The best thing about draft classes is that they are cheap. In the salary cap era, you want as much cheap talent as possible so you can form the best team you can under the circumstances. Hence, the term positional value comes into play. I break down positional value from two different vantage points:
1) What positions end up performing the best in each round relative to each other.
2) What positions make the most money in the NFL in free agency.
So let's tackle each question. The first one is interesting. I'm going to be drawing from an article by Zach Whitman, who you may recognize as the creator of pSPARQ to measure athleticism, and as a writer for Field Gulls, SBNation's Seahawks site as well as Rotoworld.com. I found this article on his web site 3SigmaAthlete. It is a must visit site for those who want to advance the way they view the draft and the NFL through the prism of analytics. Tremendous stuff.
We will not be talking about Quarterbacks at all during this piece because they are their own market altogether.
Introduction: Metrics Definitions
So let's dig in to some terminology here. We're going to define very simply what Whitman used to come up with his research so we know exactly what's going on. A much more detailed explanation can be found on his site, which, again, I urge you to visit.
Anyway, Whitman uses a measurement device called "Approximate Value" to assign how well players performed per season. Approximate Value was designed by Doug Drinen of Pro-Football-Reference. It essentially is a formula that assigns players an integer that describes how much "value" the player had to his team that season. It's not perfect (nothing is) but he had this to say about the statistic:
AV is not meant to be a be-all end-all metric. Football stat lines just do not come close to capturing all the contributions of a player the way they do in baseball and basketball. If one player is a 16 and another is a 14, we can’t be very confident that the 16AV player actually had a better season than the 14AV player. But I am pretty confident that the collection of all players with 16AV played better, as an entire group, than the collection of all players with 14AV.
So with that said, Whitman sets his limits as all players from the 1999-2012 drafts (he didn't want to use 2013 and 2014 because he wanted to use at least three seasons worth of data). He specifically uses "AV3", which he defines as:
The sum of a given player’s three best Approximate Value seasons. This means we’re measuring peak performance.
With his data set, he was able to come up with an "average" AV3 of players by position in each round. He called this "Expected AV3" or, in simple terms, the performance that's "expected" from a 3rd round OL, or a 6th round TE, and so on.
To create a graphical representation of how each position compares to each other (for example, how much better a fifth round CB performs compared to a fifth round S), he uses a formula for "Marginal AV3." All that means is that he's relating the average of position with the average of the others for that round (otherwise known as the "Expected AV3"). Positions that do well will have a positive Marginal AV3, while those that are inefficient will have a negative Marginal AV3.
He also defined a metric called "Hit Rate." He defines it as:
I’ve arbitrarily defined a "hit" as a player who generates an AV in the top 30% of all drafted players from the same position group.
So basically, a player in the top 30 percent of his group is considered a hit. Easy enough.
So with all that said, let's take a look at what he found.
Positional Value: What Positions Do Well In Each Round?
Let's take a look at this. This measures how well each position does compared to the other (since it's Marginal AV3) in each round. You can see that in the first round, relative to other positions, tight ends and running backs perform far better. Why is that? Well, it has to take a special type of player at the tight end and running back position to be drafted early, so it's only natural that they might perform better.
Wide receiver starts off well, though one drafted in the top 16 will give you an average performance relative to other positions, from pick 17 to pick 64, that's the sweet spot for wide receivers it seems. Offensive linemen are kind of the opposite. OL taken in the first round do not perform well at all. That's compared to the later rounds where there are more hits with fliers.
My theory? I think offensive linemen, especially those taken in the top 16, are pressured to start right away. They aren't allowed to sit and develop. They've got the talent but throwing any rookie into the fire like that is bound to be full of mistakes as they grow and that causes a depression in their approximate value. You look at the stark jump from Round 1 to Round 2. There's a healthy mix of superior talent from the OL pool along with the fact that not all of these players start right away.
As for the OL in the late rounds, they beast out compared to other positions because I think it's easier to stick on to a 53- man roster as a late round OL since there's five OL positions. Much harder to stick around as a WR or TE. I'll bet a few of those OL developed into slam dunks as well.
Now, let's look at the percentage of players by position that hit (remember, Whitman defined "hit" as a player with an AV in the top 30 percent of their position) in each round:
Obviously, the further down you go by round, the hit rate plummets. That's to be expected. As we discussed before, tight ends with special traits are drafted early, and they almost always hit.
Offensive line and wide receiver follow similar curves. They go from around a 77 percent hit rate at the top half of the first round to around 50 percent by Round 2. After that, it's a pretty steep curve down. Round 3 presented a hit rate of around 30 percent and it went downhill from there. Nobody suffers more than running back. Goes from 60 percent in the latter half of Round 1 to 30 percent in Round 2.
Now for the defense, you can see that if you're taking corners, take them in the first round. Teams do a pretty good job scouting them and the best ones are taken early because they are an important position. No surprise that first-round corners do well, and there's also a positive hit rate in the seventh round, where teams take fliers on talented but flawed players that can play special teams. It's no surprise as well that corners stuck in the middle, like the third round, present one of the worst values in the NFL. The talent pool is likely drained and it's too early to take your sleepers/character concerns like you can in the seventh round.
One of the most interesting phenomenons is the EDGE players (and to a smaller extent DL). They present negative value in the first, but the production skyrockets in the fourth round. Whitman posits (and I agree with him) that it's due to the severe over-drafting of EDGE players every year. The position has become so valuable that teams will reach and reach early, so less talented players are taken earlier, depressing relative AVs of those players. That's why you see the biggest negatives coming in for EDGE players in the latter half in Round 1 and Round 2.
So then, why is there such a huge jump? Well let's take a look at the hit rate graph of the defense to explore further:
Let's continue our discussion of edge rushers. What's astonishing here is that from Round 2, you've got a hit rate of a little over 30 percent. It certainly shows a huge drop from Round 1, and strengthens our argument that there's over-drafting of EDGE players early on in the draft. But what's more astonishing is looking at the hit rate for EDGE players in Round 5. It's slightly under 30 percent. Think about that.
The difference in hit rate in EDGE players taken in Round 2 to Round 5 is less than 10 percent! Why is that? Well, Whitman asserts that it is because there's so much projection involved with pass rushers. They are often misused in college and their role in the pros will often be drastically different ... more so than with most other positions.
So all of this means that, well, GMs don't really know what's going on with edge guys from Round 2 onwards. That's tough to hear but it makes sense. It confirms our "MarginalAV3" chart from above. If there's such little difference from Round 2 edge guys and Round 5 edge guys, you'll have a negative "Marginal AV3" the earlier you are, while having a positive one later on relative to everybody else.
Much like with WRs and OL, if you want a safety or a LB, you can see a pretty sharp drop off from Round 2 to Round 3. Better grab those guys early or it's going to be an upward climb for the hope for them to pan out.
Positional Value: What Positions Cost The Most?
Using the wonderful NFL salary cap resource OverTheCap, I took the mean (average) of the top 10 salaries at each position. Using the top 10 is completely arbitrary, but we use a lot of measures using "top 10 players" so I thought it would be a good medium.
This is essentially just a rough way of trying to figure out the cost per year of a top 10 player at each position.
Note: "EDGE" players comprise of 4-3 DEs and 3-4 OLBs, "DL" consists of 4-3 DTs and 3-4 NT and DEs. "LB" consists of 4-3 OLBs and ILBs. "OT" and "OG" comprise of both left side and right side players.
|Position||Avg $/yr of top 10||Rank|
So let's take a look at this. The average cost per year for a top 10 pass rusher is almost $13 million, the highest of any non-QB position. That's definitely to be expected. Guys that can get to the quarterback are the ones you're going to pay the most.
It's also no surprise that offensive tackle (more specifically, left tackle, as the top 10 guys that were paid were LTs) is among the top five position groups to get paid. Wide receivers surprised me by being the second-highest paid position in the league. No wonder Jimmy Graham wanted to consider himself a wide receiver!
This is why asset allocation is so important. According to my calculations, the average top 10 wide receiver makes around $11.7 million/year. We have a top 10 wide receiver in Odell Beckham that makes $2.6 mill/year. That's huge savings on a premium position!
Defensive line is actually a bit skewed. It probably falls down to four or five if it wasn't for Ndamukong Suh's insane $19 million/year deal. You can see a huge and complete drop from position five to position six. That's where the differentiation between "premium positions" and "non-premium positions" clearly is delineated. Based on just this:
- Premium Positions: EDGE, WR, DL, CB, OT
- Non-Premium Positions: LB, S, RB, OC, TE, OG
For further proof, the highest paid safety in the NFL is Earl Thomas, making a cool $10 million/year. That would put him squarely in 10th when it comes to wide receivers or edge rushers. The theory then of positional value from a financial standpoint is simple.
The draft nets you cheap talent at any position. So aim for positions that cost a ton on average in the open market, and fill in the rest of your roster with non-premium positions in free agency because you'll get more bang for your buck.
Putting It All Together
Alright, so let's put both concepts -- hit rate for positions and financial feasibility -- together and see what conclusions we can gather from this.
1) The Riskiest Proposition is Drafting A Guard In the First Round
All throughout the draft process, I've been of the opinion that if you value a player like Brandon Scherff as only a guard, he probably shouldn't be the pick.
Looking at the "MarginalAV3" and "Hit Rate" plots, it looks like offensive linemen get significantly over-drafted and a lot of them, as we discussed, struggle in the first round relative to expectations. That makes it in itself a pretty risky proposition. Add in the fact that guards are the cheapest position in the NFL (discounting special teams), and it makes no sense to draft one so early with your most premium draft assets. Mind you, this is without data including the slow careers of Chance Warmack and Jonathan Cooper (who were drafted in 2013), so if anything, it'll get worse.
That being said, looking at the hit rates, if you do take an offensive lineman, you want to take one before Round 3. I'm not saying ignore OL. I think if you draft an OT and deal with their struggles because the hit rate in round one is still pretty high. After that, from the end of Round 3 to Round 7, the odds of you hitting on an OL are about the same.
2) The Best Investment With A Top Pick Might Just Be Corner
With the average top 10 salary for a corner nearing $11.5 million/year, it's clear that the NFL has decided that cornerback is a premium position. That means getting an elite player at this position on a rookie contract should be of utmost importance.
Looking at the "MarginalAV3" plot for defense. Look at how well corners do relative to expectations among picks 1-16. They do outrageously well. Why is that? I'm not sure, but I figure the "can't miss corners" are athletic, scheme versatile, and scouted heavily. There isn't much projection involved, and if they can shut down players in college, chances are they can do it in the pros. Of course you have guys like Morris Claiborne but you also have Joe Haden and Patrick Peterson. The hit rate for corners in the top 16 picks also approaches 80 percent, meaning that you're likely to hit there and they'll exceed expectations.
3) When In Doubt, Draft Defensive Line or Edge Rushers In the Mid-Rounds
Out of everything to take out of this piece, I think this has the strongest evidence. I'd probably avoid any offense at all in the mid rounds. Look at the offensive plot. Everybody performs under expectations. Then you look at defensive line and edge rushers.
Remember the hit rate for these guys? How from Round 2 to Round 5 or 6, it was pretty much similar? Well then no wonder that these guys perform better than expectations later on. Given that defensive line and edge rushers are often times the most expensive non-QBs on the team (ranked first and third among all positions in terms of expense), there's really no reason not to invest in mid-round defensive line. The hit rates are on par (and in some cases higher) than everybody else and you've got a better chance of exceeding expectations.
4) If You Want a WR, The Earlier The Better
Wide receiver probably follows expectations to the dot. You can see early-round wide receivers actually exceed expectations as compared to a lot of other positions. The hit rate is pretty much a step-wise, steady decrease. Like OL, we're at a hit rate of only 50 percent by the second round. The elite WRs are the ones that mostly make it in this league. There's not much hiding true talent in college. If you're drafted early, you probably deserve it. If you aren't, then the odds are stacked against you.
Considering that wide receivers comprise the second most expensive non-QB position, if you're going to invest in them, make it early if you want to be confident in those savings.
Wrapping Up: Caveats
First of all, this isn't the definitive strategy that teams should use. Even if it might be deemed inefficient, sometimes, if you need a player at a certain position ... you need them at a certain position. The Giants need a safety in the worst way. This analysis shows a huge drop in hit rate between the latter half of the first round and the second round, not to mention that safety is a pretty cheap position. You still need it, though.
Also, while this analysis is tremendous, Whitman himself points out that "Approximate Value" is by no means a perfect measurement of player success. I make no overtures about my rudimentary financial analysis being the tool to use to measure how much a position is essentially worth. There's exceptions, and there's exceptions to exceptions.
However, this data is useful. It provides some backbone to an often ignored part of draft strategy. It adds legitimacy to the concept of positional value and it makes us more cognizant of trends that occur and gives us ideas about why such trends occur. If we can harness this, we can draft smarter and build better. That's the goal, isn't it?