Battle of the NFL Drafts

One of my favorite NFL writers, Robert Mays, recently wrote a column “Crowning the NFL Draft of the Generation” for The Ringer which analyzed the success of recent NFL drafts.  In his column, Mays posited that the 2007 and 2011 drafts were a cut above the rest “as any debate about the best draft in recent years comes down to these two years.”  To determine which draft was better, he used a completely arbitrary methodology (by his own omission) and formula to determine the stronger draft.

With all the lunatic draft posts and mock drafts circulating around the internet as click bait (unlike this post, of course), I understand that this is an easy topic to write about and get people to read, without worrying too much about statistical truthfulness.  But, here at Pro Sports Analytics, we worry about boring stuff like that!  In order to answer this question in an unbiased manner, let’s think about two questions:

  1. Is it valid for Mays to exclude other recent drafts from his analysis?

As Mays points out, the 2007 and 2011 drafts produced a high number of future Hall of Famers, including Adrian Peterson, Calvin Johnson, and Joe Thomas in 2007 and Cam Newton, Von Miller, and J.J. Watt in 2011 and a number of multi-year starters.  However, this approach shows significant bias.  It doesn’t let the data do the talking.  What if another draft was as strong or stronger and was not considered?

  1. Is there a better methodology to apply than Mays’ arbitrary points system?

His methodology awards players’ points for receiving offseason awards (pro bowls, 1st team All-Pro, MVP, etc.) and arbitrary criteria such as if a player is a probable Hall of Famer or is the defining player of the generation at a position.  If we want to understand the top end strength of each draft (i.e., how good the best players are in each draft), this approach would effectively answer that question.  But it wouldn’t touch on the depth of a draft as it ignores the contributions of good or average players who couldn’t quite receive any offseason awards (though, let’s be honest, everyone makes the Pro Bowl at this point).  To me, the best draft is not the draft with the most top end talent but a deep draft with a respectable amount of top end talent.  It’s easy to point to drafts that produced an inordinate number of great players, but that doesn’t speak to the overall strength of the draft, especially if the value in that draft is the substantial number of players who consistently start but don’t generate any offseason accolades.

Pro Football Reference constructed a statistic (though somewhat flawed) that values every current or past NFL player regardless of position called Approximate Value (“AV”).  I’ve discussed the background on AV, the positive aspects of AV, and its inherent limitations in previous posts on this subject (which you should all read, of course). Using AV, we can calculate the total value that the players in each draft generate.


In order to better answer the question that Mays presents on the strongest draft in recent history, I summed the total AV generated by players in each draft.  I’ve limited it to the first five years of a player’s career, since the lies in the immediate returns that players provide to teams.  Given that the average career length for an NFL player is 3.3 years, most teams don’t draft players expecting them to spend their entire careers with that team.  Instead, teams strive to maximize the value they receive from their drafted players for the years in which they have control over that player (I feel unclean just writing this – control, maximize value, short term gains – ugh, NFL owners don’t sound like great employers…).  Teams tend to draft players, especially early selections, to play immediately.  Few teams are content grooming players to start down the road (Olivia Munn’s ex-boyfriend being one of the biggest exceptions).

Since we want to compare 2007 and 2011 (I’ve included all drafts from 2002 to 2012 in this analysis), we have to cut this off at some place and five years is a reasonable cutoff point.  The table below displays the total amount of AV acquired in each draft from 2002 to 2012 (the last draft year in which the players drafted have played five years in the NFL – well, those still in the NFL…).

Looking at this table, the first thing that jumps out is the low AV per pick from the 2007 draft.  According to Mays, 2007 and 2011 were the surefire top two drafts in recent memory and he concluded that the 2007 was stronger!  However, 2007 produced the least amount of AV and only 2009 produced less AV per pick than 2007.

Why does stand out so much?  2007 did produce some generational players – Calvin Johnson, Joe Thomas, Adrian Peterson, Patrick Willis, Darrelle Revis – but it was not a deep draft.  Only 47 of the 255 players drafted in 2007 produced 15 or more points of AV in the first five years of in the NFL.  Strong at the top but lacking in depth.  Plus, it had some serious early duds.  The most obvious is the workout warrior, JaMarcus Russell.

Let’s look at the defensive linemen selected in the first round in 2007.

Let’s also not forget that JaMarcus Russell wasn’t the only quarterback dud in this draft.  Brady Quinn, the other first round pick (who some thought might go #1) started 20 games, won four of them, and averaged 127 passing yards per game.  The best quarterback in this draft was Trent Edwards who started 33 games (including 14 for the Bills in his second season) but only won 14 of those games while throwing for 26 touchdowns and 30 interceptions.  Ouch.

The other reason that this draft stands out is that two of the more notable players in this draft were not standout players until later in their career.  Marshawn Lynch didn’t become Beast Mode until his second season with the Seahawks, his sixth season in the NFL.  Marshall Yanda, one of the best offensive guards of all time, didn’t start more than 12 games until his fourth season, meaning he didn’t hit his prime until well into his career.

Another theory: maybe he had his years confused and thought he was talking about 2006?

I doubt it, but I was pretty surprised to see that the 2006 draft produced the highest amount of AV above replacement (I rechecked the data three times to make sure I didn’t screw anything up).  This draft is remembered for the Texans’ surprising choice of Mario Williams over Reggie Bush, who the Saints delightfully scooped up with the second pick.  Both of these players went on to have fairly successful, but not standout careers.  Neither of these player’s nor the draft’s best player’s (Haloti Ngata) career rivals the success of some of the top players in 2007 like Joe Thomas or Adrian Peterson.  But the 2006 draft was deeper than 2007.  As I mentioned earlier, 47 players in the 2007 draft produced 15 or more points of AV in the first five years of in the NFL compared to 65 players in the 2006 draft.  2007 produced a stronger collection of high end players, but 2006 produced more above average players.  In essence, it was a deeper draft.

And the Answer is…

Mays does get one thing right – the 2011 draft was really strong, especially the first round.  The table below shows the total AV produced by first round picks from 2002 to 2012.

2011 had the highest amount of AV produced from the players selected in the first round.  If you had a top 12 pick in 2011 and didn’t desperately grab a quarterback (and stayed away from anyone who didn’t go crazy), you were practically guaranteed an impact starter and perennial Pro Bowler.  The table below shows the players selected in the first round in 2011 and their total AV above replacement over the first five years of their career.

The first six picks all produced at least 38 points above replacement (all SEC products too), which is just ridiculous.  As great as Christian Ponder’s first throw as a Viking was, it doesn’t make up for the lost opportunity here for the Vikings (and the Titans and Jaguars).  This is a draft where any team outside the top 12 would have been completely justified in bucking the draft value chart and trading up to land one of the studs, given the talent available.

However, even without a top 12 pick, there was still value to be had later in the first round.  The Saints selected Cameron Jordan (not to be confused with Jordan Cameron) 24th overall (the 9th defensive linemen drafted!) who has thus far produced 46.5 sacks in six seasons.  The Saints doubled down when they chose Mark Ingram, a former Heisman winner, 28th overall.  Ingram has averaged 4.4 yards per carry over his career with one 1,000 yard season, which I would consider a success for a late first round pick.

As I was describing this post to one of my friends, he asked me a hypothetical question.  If you could construct a 53 man roster from one draft, which would you choose?  I think the answer has to be the 2011 draft.  Even though the 2006 draft produced more AV above replacement than any other draft, there are two mitigating factors here that point this in favor of 2011.

First, I only have 53 roster spots.  I don’t want to choose from the deepest draft but from the draft that will provide the most value from #1 to #53.  Anyone after 53 won’t matter.  This is similar to the argument against surplus value.  While teams can generate the most value per dollar from a second round pick, NFL teams need high end talent to compete.  The 2011 draft offered a lot of high end talent.  I would rather have Patrick Peterson and Richard Sherman as my two starting cornerbacks than a young Antonio Cromartie, Cortland Finnegan, Johnathan Joseph, and Cedric Griffin as my top four cornerbacks.

The second factor is that the 2006 draft doesn’t have a quarterback.  Sorry to all the Jay Cutler fans out there (I’m sure Chicago hates me now…just kidding), but I think I’d rather have Cam Newton.  Or Andy Dalton.  Or Tyrod Taylor.  Or even that one guy who caused a controversy because he doesn’t like to stand.

Leave a Reply

Your email address will not be published.