An erratic but improving offensive line played a big part in Denver's championship win.
25 Apr 2013
by Danny Tuccitto
In this fifth (and final) episode of our historical draft efficiency series, I'll attempt to determine the best and worst individual picks since the AFL-NFL merger. At this point, I assume you know what value above expectation (VAE) and return on investment (ROI) measure. And given the difficulty of today's task, I also assume that you realize I'll be employing a format similar to the one I used to rank draft classes: present all the stats, offer my opinion about who should be in the conversation for best and worst, and have at it in the comments section. If any of what you just read seems mysterious, here are links to previous parts in the series so you can get up to speed:
Now that we're all on the same page (both literally and figuratively), I'm going to make one minor change to how I present the stats today. It only takes a little bit of thought on the topic to realize that, because there have been a ton of draft picks with essentially no NFL career to speak of (2,244 of 8,436 selected in the top 222), simple math dictates that there will be a 2,244-player tie for last place in the rankings for Adjusted Career AV per Year (CarAV/Yr) and ROI. I don't think you would be satisfied if my answer to, "What was the worst pick since the merger?" was, "These 2,000-or-so guys."
You would be even less satisfied with that answer when I also told you almost every high-profile draft bust you can think of off the top of your head is not among those 2,000-or-so guys. For instance, the NFL Network counted down their Top 10 draft busts a few years ago, and not a single one of the players on their list are among the cast of thousands tied for last place according to Adjusted CarAV/Yr and ROI. Therefore, I'll only be using VAE to determine the handful of candidates for worst pick.
Here is the table showing the top 20 players from 1970 to 2007 according to Adjusted CarAV/Yr:
As Adjusted CarAV/Yr is simply the measure of value in my model, it's no surprise that nearly every player on this list is a Hall of Famer, would-be Hall of Famer, or should-be Hall of Famer. If that weren't the case, we would be in trouble. No, the surprise comes from the fact this group has an average career length of 10.5 years.
One of the criticisms of dividing Career AV by career length was that supernova-like flashes in the pan could break the system. The table clearly shows otherwise. Yes, it's true that Billy Sims and Terrell Davis had shortened careers, but few, if any, people would label them as "flashes in the pan." Meanwhile, people think Barry Sanders had a shortened career, but that's more perception than reality: The man did play 10 seasons. At the end of the day, even at a position as notorious for short careers as any, the nine running backs at the top of Adjusted CarAV/Yr averaged 9.0 seasons.
The other thing I'll draw your attention to is that, given where the players in the table were selected, you can see the VAE and ROI rankings already taking shape. The average draft slot for these players is about the 41st pick, but that's only because of the trio taken after No. 150. Removing those outliers, the remaining 17 players were selected, on average, about 15th. Again, this is the reason why VAE and ROI are better measures for our little relative-value exercise: Over a large sample, and over time, better players tend to get picked earlier in the draft. Sanders was a great pick, but not as great of a pick as Terrell Davis, Tom Brady, or Jake Scott (a two-time All Pro safety for the early 70's Dolphins).
And who the heck is Lydell Mitchell, you may ask? We'll get to him in a bit.
Here is the table showing the top 20 picks according to ROI:
Davis, Brady, and Scott are the only picks that show up on this list too, so let's set them aside for reasons that should be obvious. (Hint: They show up again later.)
For all I said about how Adjusted CarAV/Yr didn't elevate flashes in the pan as much one might have expected, it did still happen. The player ranked fourth in ROI is a perfect example. Shane Olivea was a steal for the Chargers at No. 209. Of his four years as their starting right tackle, San Diego's offensive line ranked in the top eight of ALY twice: seventh in 2005 and first in 2006. Unfortunately, thanks to injuries and a subsequent pain-pill-popping problem, Olivea washed out of the league in 2008. Based on that resume, do I agree with his ROI ranking? No, I do not.
At No. 6 is center Tom Nalen, who you will notice is one of three players in the top 11 from the back-to-back Super Bowl-winning Broncos teams of the late 1990s (along with Davis and Shannon Sharpe). Unlike Olivea, Nalen definitely deserves to be considered one of the most efficient picks since the merger. Nalen was a 13-year starter, made five Pro Bowls, and was voted first team All-Pro twice. And when you consider that the 1994 draft only had 222 picks, he's one of the closest things to a star Mr. Irrelevant that the league has ever seen. Despite Mike Shanahan being seen as one of the best late-round drafters in league history, a reputation further cemented by his selection of Alfred Morris last year, both Nalen and Sharpe were taken before he took the reins in 1995. (I'm not giving him credit for the Sharpe pick when he was a mere "offensive assistant.")
Most of the rest of the list should be familiar to those who have been following this series over the past month, either because I've specifically discussed them or because they were part of great draft classes. However, one final player worth noting is David Woodley, who checks in as the second-most efficient quarterback pick since the merger. As a lifelong resident of Miami save for a longer-than-anticipated stint in Gainesville, I can tell you that the Woodley era is all but forgotten down here. Granted, it's tough to recall a member of the Dolphins quarterbacking relay team who took the baton from Bob Griese and passed it to Dan Marino. Nevertheless, in his two full years as the starter, he helped Miami win two division titles and reach Super Bowl XVII. Even knowing as a stathead that wins is a bad thing by which to judge quarterbacks, that's still one hell of a return on the 214th pick.
Before proceeding, let me just point out that, although Tampa Bay getting nothing from their No. 1 pick in 1986 is technically the worst relative value since 1970, it just doesn't feel right to have Bo Jackson on the list of worsts -- let alone atop it. Don't worry, though. For any Bo-philes out there, the list does have one, just not attached to the same surname.
I know this exception might also apply to Kelly Stouffer, but come on, it's Kelly Friggin' Stouffer. The same lack of sympathy goes for Russell Erxleben. As the Saints punter for four years, Erxleben's career should not be worth 0 AV, which is what PFR defaults to for specialists. Nevertheless, I kept him on the list because (a) New Orleans had no business taking a punter 11th overall; and (b) Erxleben pulled a Garo Yepremian in his first NFL game -- in overtime no less. Dapper Dick Nolan was not pleased, banishing the rookie to Quarantine Bay until Week 1 of the following season. Also, there's this sordid tale. (Seriously, click that link!)
So, with that said, here is the table showing the top-20 and bottom-20 picks according to VAE:
|Best Picks by VAE||Worst Picks by VAE|
|Lydell Mitchell||BALC||1972||48||+9.1||Ki-Jana Carter||CIN||1995||1||-6.0|
|Jake Scott||MIA||1970||159||+9.0||Ryan Leaf||SD||1998||2||-5.9|
|Terrell Davis||DEN||1995||196||+8.9||Akili Smith||CIN||1999||3||-5.7|
|Tom Brady||NE||2000||199||+8.5||Charles Rogers||DET||2003||2||-5.3|
|Jack Lambert||PIT||1974||46||+8.1||JaMarcus Russell||OAK||2007||1||-5.3|
|Jack Ham||PIT||1971||34||+7.1||Aundray Bruce||ATL||1988||1||-5.0|
|Jared Allen||KC||2004||126||+6.9||Bruce Pickens||ATL||1991||3||-5.0|
|Patrick Willis||SF||2007||11||+6.8||Steve Emtman||IND||1992||1||-4.9|
|Lawrence McCutcheon||LARM||1972||70||+6.8||Steve Niehaus||SEA||1976||2||-4.8|
|Domanick Davis||HOU||2003||101||+6.8||Reggie Rogers||DET||1987||7||-4.8|
|Mike Singletary||CHI||1981||38||+6.7||Larry Stegent||STLC||1970||8||-4.8|
|Charlie Johnson||PHI||1977||175||+6.7||Bo Matthews||SD||1974||2||-4.7|
|Curtis Martin||NE||1995||74||+6.6||Kelly Stouffer||STLC||1987||6||-4.6|
|Derrick Brooks||TB||1995||28||+6.4||Rich Campbell||GB||1981||6||-4.6|
|Jahri Evans||NO||2006||108||+6.4||Art Schlichter||IND||1982||4||-4.5|
|Lemar Parrish||CIN||1970||163||+6.3||Heath Shuler||WAS||1994||3||-4.4|
|Eric Martin||NO||1985||179||+6.3||Mike Junkin||CLE1||1987||5||-4.4|
|Zach Thomas||MIA||1996||154||+6.2||Russell Erxleben||NO||1979||11||-4.4|
|Barry Sanders||DET||1989||3||+6.2||Brian Jozwiak||KC||1986||7||-4.3|
|William Andrews||ATL||1979||79||+6.2||Alonzo Highsmith||HOIL||1987||3||-4.1|
The best 20 picks have already been accounted for, so let's focus on the bottom 20. Going back to the NFL Network, it turns out that, unlike some of the laughable inclusions on their top 10 draft classes, nearly all of their top 10 draft busts were worthy of such ignominy according to VAE. Ryan Leaf (first on their list), Aundray Bruce (sixth), Art Schlichter (seventh), and Heath Shuler (ninth) overlap with the above table. Tony Mandarich (second on their list, 30th on mine), Rick Mirer (fourth, 23rd), and the Cougars combination of Andre Ware (10th, 29th) and David Klingler (10th, 51st) are distinctions without much of a difference.
That leaves only three players in their top 10 draft busts that VAE says didn't belong. NFL Network ranked Lawrence Phillips eighth, but his -1.9 VAE only ranks 577th-worst according to my model. Tim Couch was fifth on their list, but his -1.8 VAE ranks 748th on mine. Granted, with over 8,000 picks in my database, both of these players are in the top 10 percent of VAE busts; they're just not anywhere near the top 10.
NFL Network's final, and easily worst, sin of commission had Brian Bosworth as its victim. Because he was a supplemental draft pick, he technically isn't included in my rankings. However, the 1988 pick Seattle forfeited to select him in 1987 would have been No. 19 overall. Adjusting for the fact that the 1988 draft was 6.2 percentages points more valuable than the 1970 baseline, if Seattle's 19th pick had produced the same Career AV/Yr as Bosworth (3.5), that player's VAE would have been a mere -0.4. To reach his name, the above table would have required 4,318 rows. Now, maybe Bosworth would have been taken before 19th in 1988 (or a Boz-less 1987 Seahawks would have picked earlier than 19th in the following draft), but he would have had to have been taken in the top five to produce a VAE as "bad" as that of the also-undeserving Couch.
In terms of NFL Network's sins of omission, there aren't many, but one is a real doozy. If we include players added to the their updated top 10 in April 2010, Akili Smith (eighth on their newer list) and Charles Rogers (seventh) also overlap with the table. And considering that their update was in black celebration of JaMarcus Russell's release, we can check him off as an oversight as well.
If you haven't figured it out already, NFL Network's mindboggling omission is the player that sits atop my list of worst picks according to VAE: Ki-Jana Carter. The expectation for a No. 1 pick is 7.2 CarAV/Yr; after adjusting for the strength of the 1995 draft, Carter's CarAV/Yr was 1.2. He only qualified for our DYAR rankings once in his seven-year career (1997), and finished 28th out of 45 running backs that season. To be fair, for all intents and purposes, his career ended before it ever began. It was 1995, not 2012, so a Purple Jesus-esque recovery simply was not possible. Nevertheless, Carter did play again -- for seven more seasons. He could have put together some kind of decent career, but he didn't. If NFL Network wants to give him a pass for his knee injury, fine; I'm nowhere near as accommodating.
Given everything I've presented here, I think we have a pretty clear picture of which players should be in the conversation for best and worst picks since the merger. Regarding the worst, I'll go with Ki-Jana Carter and the picks that overlap between NFL Network's top 10 and the top of the VAE list: Ryan Leaf, Akili Smith, Charles Rogers, JaMarcus Russell, Aundray Bruce, and Steve Emtman. There's only a 1.0 VAE difference separating those seven, so absolute worst is probably in the eyes of the beholder.
However, perhaps we can help those eyes by also looking at ROI. Among the sad seven, here's how they stack up against one another: Smith (-96.2% ROI), Leaf (-92.9%), Carter (-83.9%), Rogers (-83.3%), Russell (-73.3%), Bruce (-70.2%), and Emtman (-68.4%). Armed with this information, it becomes pretty clear that Smith, Leaf, and Carter were feet and ankles below the rest.
On the flip side, my candidates for absolute best include three obvious players and one wild card. Terrell Davis, Tom Brady, and Jake Scott appear in all three tables. Despite producing an incredible career as a seventh-round pick, I'm inclined to put Scott behind Davis and Brady; however, not by much. And between the latter two, I think it's important to point out that Davis ranks above Brady according to all three stats.
As a wild card, I'm going with Lydell Mitchell for a couple of reasons. First, unlike the other three, he wasn't a late-round pick. There's something to be said for massively exceeding high expectations. He's in the top two of both Adjusted Car AV/Yr and VAE, and it's not because he benefited from a shortened career. And although he didn't make the top 20 in ROI, he's ranked just outside the top one-tenth of one percent (96th). Second, just look at the guy's career. In 1974 and 1977, he led the NFL in receptions -- at running back. In 1976, he had the highest AV of any player in the league. And if you scroll down to the bottom of that link, you'll see that he was basically the equivalent of Terrell Davis and Priest Holmes two decades beforehand.
So there you have it. Feel free to agree or disagree with me in the comments, and ask about players who aren't listed above, but whom you think are worthy of being in the conversation. I hope you enjoyed this series. If there's enough of a demand for it, I'd love to indulge everyone's draft history sensibilities in the future, so let me know if that's something you'd like to see.
50 comments, Last at 29 Dec 2015, 2:06pm by galaxyapple