Writers of Pro Football Prospectus 2008

Most Recent FO Features

WentzCar16-1.jpg

» Week 11 DVOA Ratings

DVOA has finally climbed on board the Wentz Wagon! The Eagles move into the No. 1 spot, but they aren't the only strong, well-balanced team in the NFL this year. New Orleans, Pittsburgh, and the Los Angeles Rams make this one of the best seasons ever for multiple teams over 30% in DVOA, and Minnesota isn't far behind.

13 May 2016

What Does the NFL Draft Really Produce? (Part III)

by Scott Kacsmar

In Part I of our study on the 1994-1999 NFL drafts, we focused on the career lengths and value of the players selected. In Part II, we turned our attention towards what the drafting team got out of their picks. This final part continues the team-focused results.

Average Draft Haul

At the end of the day, what can a team really expect to get out of its draft class? Where the team is picking obviously matters, but we broke down the 179 draft classes from 1994-1999 to see how many players hit certain benchmarks for just their drafted team.

Play in a game: Asking for one regular-season appearance is not a lot, even for seventh-round picks, but we found that, on average, only 77.0 percent of a team's draft class will play in at least one game for the team. Only five teams fell below 50 percent, which includes the 1999 Redskins (2-for-6) after that haul from New Orleans in the Ricky Williams trade. Fortunately, those two players (Champ Bailey and Jon Jansen) were good ones. Twenty-four draft classes had every draft pick play for them, but sometimes that is just the case of being a bad team starved for talent. For example, no team had a higher percentage than the expansion 1999 Browns, getting at least a game from all 11 of their draft picks.

Start a game: There are many ways to get into the game as a reserve, but what about starting on offense or defense? The rate here drops to 59.0 percent of a team's draft class (4.8 players, on average) that will start at least one game. The 1996 49ers bring up the rear, only getting 1-of-7 picks to start for them, but at least the one was Terrell Owens. The 1996 Rams actually had all 10 of their picks start for them, though that draft was marred by the No. 6 overall pick used on troubled running back Lawrence Phillips, who was found dead in his prison cell back in January.

Start at least 33 games: Why 33? That means the player had to start for more than two seasons with his drafted team in this 16-game era. The average draft class produced 2.1 such players, or 27.0 percent of a team's class. No team found more than five such players in a draft, and only six teams had the maximum five, including the 1996 and 1997 Buccaneers. A total of 56 draft classes (31.3 percent) produced one or none of these players. A bit troubling is the fact that numbers of players with at least 17 starts (so, more than one season) were not much higher at 2.8 players and 35.7 percent of the class.

Start at least 50 games: Yes, 50 is a nice, round number, but it also means at least four seasons with starts. When contracts are at least four years long, you definitely want to get the most of your high draft picks in this time period. We found that only 1.5 players per class (18.0 percent of picks) make it that long. Four teams (the 1995 Steelers, Vikings, and Lions, as well as the 1998 Bengals) were able to land four such players. If we labeled this the long-term starter metric, then only 84 of the 179 drafts (46.9 percent) produced multiple long-term starters for the team.

Accumulate at least 18 DrAV: We saw in Part I that the average drafted player accumulates about 18 AV for his career. For reference, Chance Warmack has 18 AV through three seasons with the Titans. So what happens if we set 18 as the benchmark for a player's draft value to his drafted team? We find the average draft class produces 1.8 such players, or 22.7 percent of the picks. Again, only the 1997 Buccaneers and 1995 Packers found five such players, as those have been two of the very best drafts from this study.

But generally speaking, if you can find two long-term starters and two contributors, then you are kicking ass in your draft haul. It is just not realistic to draft multiple difference-makers in one year, but finding a few in consecutive years can quickly build a great team with young, cheap talent. Look at the construction of the Seahawks' current DVOA dynasty. In 2010, they drafted Russell Okung, Earl Thomas, Golden Tate, and Kam Chancellor. In 2011, they drafted K.J. Wright, Richard Sherman, Byron Maxwell, and Malcolm Smith. To top it off, in 2012 they drafted Bruce Irvin, Bobby Wagner, and Russell Wilson (the all-important quarterback pickup), and off they went to a Super Bowl win. Seattle has since lost some of those players, but that is the reality of today's NFL as you will see in our final section.

Cumulative Team Draft Results, 1994-1999

The following table summarizes how much CarAV and DrAV each team earned for all of its draft picks in 1994-1999. Since the number of picks varies, we also included the average value per pick. The last two columns show the percentage of CarAV that was accumulated by the player's drafted team. Remember, the Browns get credit for what their players did as Ravens, and "TEN" includes the Houston Oilers years. We also used "OAK" for the Raiders and "STL" for the Rams even though both teams spent 1994 in Los Angeles.

NFL Draft: Career AV and Draft AV, 1994-1999
Team Picks CarAV Rk AVG Rk DrAV Rk AVG Rk Dr Pct. Rk
PIT 58 1234 1 21.3 7 848 2 14.6 4 68.7% 13
GB 54 1211 2 22.4 6 720 7 13.3 8 59.5% 24
TB 49 1181 3 24.1 3 900 1 18.4 2 76.2% 3
STL 51 1163 4 22.8 5 726 6 14.2 5 62.4% 21
IND 45 1133 5 25.2 1 794 3 17.6 3 70.1% 11
PHI 56 1069 6 19.1 11 765 4 13.7 7 71.6% 9
TEN 53 1050 7 19.8 9 746 5 14.1 6 71.0% 10
NE 57 1021 8 17.9 16 659 9 11.6 16 64.5% 19
MIN 48 995 9 20.7 8 628 10 13.1 9 63.1% 20
MIA 58 965 10 16.6 20 715 8 12.3 13 74.1% 6
BUF 53 948 11 17.9 17 564 15 10.6 22 59.5% 23
SEA 39 929 12 23.8 4 410 27 10.5 24 44.1% 31
DEN 46 875 13 19.0 12 569 12 12.4 11 65.0% 17
CIN 47 870 14 18.5 14 510 18 10.9 20 58.6% 26
DAL 52 831 15 16.0 22 569 13 10.9 18 68.5% 15
Team Picks CarAV Rk AVG Rk DrAV Rk AVG Rk Dr Pct. Rk
CHI 53 820 16 15.5 24 471 20 8.9 26 57.4% 28
WAS 42 809 17 19.3 10 535 17 12.7 10 66.1% 16
ARI 57 806 18 14.1 28 418 26 7.3 30 51.9% 29
NYG 46 802 19 17.4 18 558 16 12.1 14 69.6% 12
JAC 46 787 20 17.1 19 568 14 12.3 12 72.2% 8
OAK 42 783 21 18.6 13 465 21 11.1 17 59.4% 25
KC 46 759 22 16.5 21 442 23 9.6 25 58.2% 27
BAL 30 747 23 24.9 2 600 11 20.0 1 80.3% 1
NYJ 52 744 24 14.3 26 385 28 7.4 29 51.7% 30
ATL 46 679 25 14.8 25 495 19 10.8 21 72.9% 7
SF 37 667 26 18.0 15 431 25 11.6 15 64.6% 18
CAR 40 635 27 15.9 23 435 24 10.9 19 68.5% 14
DET 42 592 28 14.1 29 445 22 10.6 23 75.2% 5
NO 41 582 29 14.2 27 361 29 8.8 27 62.0% 22
SD 50 373 30 7.5 31 281 30 5.6 31 75.3% 4
CLE 23 248 31 10.8 30 193 31 8.4 28 77.8% 2

Go figure, no teams found more CarAV in the draft than Pittsburgh and Green Bay, but Tampa Bay had the best draft value in this time. If the Buccaneers had done better at quarterback than Trent Dilfer, then this team may have been a dynasty after adding players such as Derrick Brooks, Warren Sapp, Mike Alstott, Donnie Abraham, Warrick Dunn, Ronde Barber, and Anthony McFarland.

The Colts also fared well in this time, but had the highest standard deviation (41.4) in career AV. That was bound to happen for a team that drafted Peyton Manning, but also used a top-five pick on Trev Alberts. The Chargers (13.0) and Browns (15.6) had the lowest standard deviation, but that is because their picks were consistently poor -- the least valuable of any team.

San Diego especially botched these drafts, which might explain why the team did nothing to build on its surprising Super Bowl appearance in the 1994 season. The Chargers drafted 50 players in these six drafts, and only safety Rodney Harrison, a fifth-round gem, exceeded 40 CarAV. Vaughn Parker (40 AV) was the only other player San Diego drafted to exceed 32 CarAV or 23 DrAV. It is worth noting that San Diego's only first-round pick in this time was trading up for Ryan Leaf in 1998, but that is just the centerpiece of what was a remarkably poor draft period for the franchise.

As far as getting the most value out of your players, the Ravens lead the way with 80.3 percent of their picks' value coming while in Baltimore. It helps when Jonathan Ogden and Ray Lewis played their entire Hall of Fame careers with the team. The split in the top four here is amusing, given that Baltimore and Tampa Bay drafted multiple Hall of Famers and eventually won Super Bowls with incredible defenses. The Chargers and Browns drafted players that were of little value to anyone, but at least most of it was spent in San Diego and Cleveland.

Every team was above 50 percent except for Seattle, which only gained 44.1 percent of the value from its players. It was the era of Seattle being an AFC West team, playing in the Kingdome, and trying to make Rick Mirer happen. So who left Seattle for greener pastures? Kevin Mawae, a likely future Hall of Fame center, spent just four seasons with Seattle. He never made a Pro Bowl with the Seahawks, but went on to make eight of them with the Jets and Titans. In fact, Mawae's 95 AV after leaving his drafted team was the highest of the 1,459 players in this study. The sixth-highest difference was another Seattle refugee: running back Ahman Green was traded to Green Bay, where he became a star. Sam Adams and Phillip Daniels were solid defensive linemen who shined more elsewhere, but that gives Seattle four of the 28 players to break 50 AV after leaving their drafted team. Joey Galloway, Shawn Springs, Todd Weiner, and Pete Kendall also fared pretty well away from Seattle.

The Ones That Got Away

Our last table looks at the player for each team with the largest difference between his CarAV and DrAV.

Player Team Round CarAV DrAV Diff.
Simeon Rice ARI 1 88 28 60
Ephraim Salaam ATL 7 59 24 35
Brandon Stokley BAL 4 45 5 40
Antoine Winfield BUF 1 62 15 47
Kerry Collins CAR 1 81 12 69
Chris Draft CHI 6 39 0 39
Kelly Gregg CIN 6 66 0 66
Derrick Alexander CLE 1 56 26 30
Randall Godfrey DAL 2 72 28 44
Nate Wayne DEN 7 35 0 35
Jeff Hartings DET 1 74 25 49
Matt Hasselbeck GB 6 88 0 88
Marshall Faulk IND 1 133 52 81
Renaldo Wynn JAC 1 55 24 31
Joe Horn KC 5 69 5 64
Player Team Round CarAV DrAV Diff.
Norman Hand MIA 5 38 1 37
Dewayne Washington MIN 1 67 21 46
Curtis Martin NE 3 101 26 75
Ricky Williams NO 1 76 26 50
Scott Gragg NYG 2 60 20 40
James Farrior NYJ 1 95 12 83
La'Roi Glover OAK 5 91 0 91
Charlie Garner PHI 2 73 11 62
Mike Vrabel PIT 3 66 2 64
Rodney Harrison SD 5 72 42 30
Kevin Mawae SEA 2 109 14 95
Terrell Owens SF 3 119 74 45
Eddie Kennison STL 1 70 11 59
Al Harris TB 6 53 0 53
Jon Runyan TEN 4 79 22 57
Champ Bailey WAS 1 112 37 75

For all the players Seattle lost, Matt Hasselbeck was a huge pickup in a trade from Green Bay. Of course, Mike Holmgren was there in 1998 when the Packers drafted Hasselbeck as one of Brett Favre's many notable backups. When Holmgren went to coach Seattle, he soon traded for Hasselbeck, who led Seattle to its first Super Bowl appearance in the 2005 season. That is one Green Bay fans can live with, since Favre remained the quarterback throughout the 2007 season. Similarly, Indianapolis fans cannot be too salty about the Marshall Faulk trade to St. Louis, since Edgerrin James was drafted in 1999 to be his replacement. That was a move that favored both sides, as it seems unlikely that Faulk would have exploded in Indianapolis the way he did with the Rams and the Greatest Show on Turf. James was a great addition to the new "triplets" with Peyton Manning and Marvin Harrison on board.

Pittsburgh rarely makes free-agency splashes, but Jeff Hartings (2001) and James Farrior (2002) were big additions to help the team towards a Super Bowl in 2005. New Orleans may have had the best haul here with La'Roi Glover and Joe Horn helping to resurrect the Saints in the Jim Haslett era.

Among the six players with zero DrAV, cornerback Al Harris may be the best example of a team throwing gold in the trash. Harris was a sixth-round pick in that prolific 1997 draft class from the Buccaneers, but this was a Tampa-2 defense with Tony Dungy and Monte Kiffin stressing zone coverage. Harris would soon make his mark as a man-to-man corner, known for getting very physical with receivers -- traits not necessarily suited for Tampa Bay's system. Harris first made his mark in Philadelphia before a trade to Green Bay where he scored the first and only overtime pick-six in playoff history. That pass was thrown by Hasselbeck, who officially became the last drafted player of the 1990s to announce his retirement on March 9.

Conclusion

Since we had to leave out a lot of the data results due to space, we are providing a downloadable spreadsheet that will allow you to play around with the data. Just be courteous to credit Football Outsiders if you publish something that used this file.

I look forward to adding the next draft years to this study. The 2000 draft is just waiting for Tom Brady and Oakland's two kicking choices (Sebastian Janikowski and Shane Lechler) to hang up the cleats. The 2001 draft is close to finished with Drew Brees, Michael Vick (currently an unsigned free agent), and Steve Smith remaining. The 2002 draft, known for some huge busts in the top 10, is also down to three players: Julius Peppers, Dwight Freeney (another unsigned free agent), and, of all people, Josh McCown. There were 14 players from the 2003 draft on a 2015 roster, so we have a good wait there, but we are starting to unearth the production of a seven-round draft in the salary-cap era.

No matter how well teams think they have the draft planned out, molding that college talent into a successful pro career remains very difficult. If the 2013 draft's terrible first round is any indication, busts are still high in quantity even with far more information available today than there was in the 1990s. Draft success is a necessity for any team to become a consistent winner, but it cannot be your only method of team-building. Free agency is a different kind of gamble, but you wouldn't go to a casino and only play the blackjack tables, would you? Spread the wealth, spread the risk, and see what happens in this crazy league.

Posted by: Scott Kacsmar on 13 May 2016

17 comments, Last at 03 Jun 2016, 7:18am by bengt

Comments

1
by Steve in WI :: Fri, 05/13/2016 - 6:14pm

"Asking for one regular-season appearance is not a lot, even for seventh-round picks, but we found that, on average, only 77.0 percent of a team's draft class will play in at least one game for the team." This is one of the most surprising things I learned from these articles.

I wonder if there is any correlation between teams who have a higher than average percentage of draft picks who contribute in some way, and overall quality of draft/talent evaluation. The idea of a pick, even a 7th round pick, being so bad that he never sees the field in a regular season game strikes me as a really wasted pick. I'm surprised that almost a quarter of picks end up like this.

2
by Bright Blue Shorts :: Sat, 05/14/2016 - 12:57pm

It does seem surprising but a draft pick has to displace an established player who is already a known quantity. He understands the playbook, he understands special teams. I guess sometimes you get a guy into camp and he really just turns out to not know what he's up to. Some positions like QB, K, P may be drafted with no real chance to displace the incumbent.

Part I has a nice table showing starts by draft round. The first two rounds have pretty much everyone play, it's the 6th & 7th rounds where over a quarter of players don't play a game.

I wonder how much lack of scouting of lower draft picks plays into it by then. Being on the clock to to weigh up the strengths and weaknesses of your guys at a variety of positions vs the remaining available players.

3
by Arkaein :: Sat, 05/14/2016 - 2:25pm

I don't think it's surprising at all. There's a baseline level of talent needed to play in the NFL, and by the time the 7th round comes most of the incoming rookies who meet or exceed that baseline are gone.

The line has to be drawn somewhere. If there were 10 rounds, would you see players in the last round not making the team being a wasted pick? 15 rounds? At some point you are making picks with the expectation that they will never contribute, but hoping that they will beat the odds.

12
by Steve in WI :: Tue, 05/17/2016 - 1:21pm

Both of you make good points and I agree that it's not hard to understand why it's so hard for late round picks to replace established players as significant contributors to a team.

But unless you've already decided by the preseason that the guy is hopeless, wouldn't you want to see him in a regular season game at some point? Even if it's only for a couple of snaps or special-teams plays at the end of a game where the outcome is decided? I guess a guy could theoretically be the lowest guy on the depth chart for a team that has good health at his position and the team just never needs to play him.

To put it another way, I find it surprising that the group of drafted players who never see the field is larger than the group of drafted players who never start a game, but play in at least one game. Especially when the latter group includes a significant number of guys who might be decent special-teamers even for a few years, but never compete for a starting job at any position.

13
by Bright Blue Shorts :: Wed, 05/18/2016 - 8:02am

Personally I don't see it as that surprising once you hit the low rounds because the players are probably all one-trick ponies. Good at one particular part of the game but flawed and exposable at the rest.

When I wrote about "replacing an established player" I wasn't referring to a longterm veteran starter, I was talking about replacing a 2nd/3rd year guy who is also flawed. The sort of player who is already playing special teams and a backup and isn't going to last past his 4-year rookie contract.

Unless the 7th rounder comes in and shows he has something outstanding*, he's going to struggle to displace that 2nd-3rd year guy who has similar skill levels but is already a known quantity to the coaches.

I'd say teams would prefer to take the low risk option and deal with what they already know than just cut one average guy for another average guy.

* Terrell Davis - 6th round pick - might be the poster child here ... https://www.youtube.com/watch?v=_if_zC44WlI

4
by bobrulz :: Mon, 05/16/2016 - 3:58am

No surprise to find that the teams that are at the top of the draft value during this period were the most successful in the late 90s/early 00s - Bucs, Steelers, Packers, Rams, Colts, Eagles, Titans...

Even the Patriots are up there despite not having drafted Tom Brady yet, which shows they had the foundation of a good team before he came along.

Can't wait until other years start becoming available! Although Brady and Brees probably aren't retiring in the next 2-3 years, and Josh McCown could very well end up being that spunky 40-year old backup gritting his teeth through hits on a mediocre team (2015 Matt Hasselbeck).

5
by Babylon :: Mon, 05/16/2016 - 12:23pm

The Patriots are a weird case because Belichick cut 60% of the team after 2000 and replaced them with bargain basement free agents. However, they still had Willie McGinest, Tedy Bruschi, Ty Law, Lawyer Milloy, Ted Johnson and Tebucky Jones which was a pretty good defensive base to build around, and a few complimentary pieces for the offense like Kevin Faulk and Damien Woody and also Adam Vinatieri. Also Terry Glenn if you want to count the like 6 games he got in before Belichick sent him on a rocket ship into the sun.

The last two Parcells draft were pretty much a work of art though. Rounds 1-3 in '95 had Ty Law, Ted Johnson and Curtis Martin then in '96 Terry Glenn, Lawyer Milloy and Tedy Bruschi.
Then Parcells leaves over a dispute over roster control and they get three years of Bobby Grier's barren drafts leading to the aforementioned cleansing by Belichick.

7
by Bright Blue Shorts :: Mon, 05/16/2016 - 2:09pm

Seems to me that cutting 60% of a team has become a standard turnaround technique especially in the win-now NFL. It happened with Parcells in 2008 at the Dolphins, it happened at the Raiders in 2012 under Reggie McK. Seattle made 280 roster moves in their first year under Pete Carroll. Looks back to 1989 and Dallas replaced most of the Cowboys from the Landry era (even Herschel) with young guys.

It's easier to take a boatload of new guys and get them to do what you want than to spend a season or two meeting resistance from guys who say "That's not how we used to do things under coach Bob". A bit of fear that cuts are around the corner also keeps everybody on their toes and playing hard!

6
by Scott Kacsmar :: Mon, 05/16/2016 - 2:07pm

Updating this in the future, I'd probably remove Leon Bender and Brandon Burlsworth from the data. Untimely deaths before their careers even started just don't speak to anything they or their drafting teams did wrong. Just two very unfortunate results unrelated to the game of football.

8
by Bright Blue Shorts :: Mon, 05/16/2016 - 2:30pm

It was a good series of articles Scott. From my perspective Part I could probably have been split into two for easier reading. It's always nice to have something to read when there's not much else on the boards.

I guess it's beyond your desire or remit but I was briefly looking at Bill Walsh's GMing at the 49ers from 1979-88. Obviously a different era but he seemed to be very good at selecting players who contributed. 55 of 60 players drafted in rounds 1-7 played at least 1 NFL game, 8 of 14 in rounds 8-9, it was only rounds 10-12 (picks 253-336) where he found almost no-one but even Dwight Clark and Jesse Sapolu were long term starters. His 1986 draft was one of the best ever with eight great picks.

9
by ChrisS :: Mon, 05/16/2016 - 5:56pm

Good set of articles Scott. I wanted to post this after part I but forgot. After looking at the futility of rounds 6 and 7 I was wondering if UDFA's might actually do better than low draft choices. I would think that the teams don't spend much time on the low end of the draft chart and take less care in making the selections (if you have 10 free minutes you would most likely spend it determining who is the 3rd or 4th best guard, not the 20th or 21st best). But it seems that the UDFA process is more of a dialog (recruiting process?) between the team and the player where the player can figure out what the team wants and can try to go to a team that would/could use his skill set. And the teams have more time to focus on the few guys that are available and look for skills that fill a missing niche.

10
by Jerry :: Tue, 05/17/2016 - 7:37am

A mildly interesting project for someone might be to count the number of UDFAs on 90-man rosters right now, and what percentage of them end up on opening day rosters. I'm guessing that success rate will make the seventh round look a lot better.

What we get from the logarithmic curve of draft pick vs. value is that there's generally a lot more difference between the 3rd and 4th best guards than there is between the 20th and 21st. And in the same way the curve flattens at about the fourth round, "eighth round picks" (i.e. the top UDFAs) aren't much different from seventh-rounders, but the more UDFAs teams sign, the worse they get.

So predicting that a team will have UDFAs on the roster is safe, but predicting which one(s) right now is much harder.

11
by tuluse :: Tue, 05/17/2016 - 11:57am

There is actually some evidence that an UDFA who makes it to camp has better odds of making final 53 than 7th round picks.

So either it's just a weird quirk, or maybe when you get past the first 200 players or so each year, team's have no more idea than the rest of us about potential talent and a few days to do some research helps them out.

17
by bengt :: Fri, 06/03/2016 - 7:18am

I guess one reason why UDFAs might have better chances to make a roster than players drafted very low could be that UDFAs can get signing bonuses nowadays. so they would incur dead money when cut.

14
by Hurt Bones :: Wed, 05/18/2016 - 12:16pm

Most teams sign about a dozen or more UDFAs. So nearly 400 UDFAs every year. Yeah, some make it. Some are damn good, 90% are fodder.

15
by LionInAZ :: Wed, 05/18/2016 - 8:17pm

Like tuluse, I wondered about this also. Like you, I concluded that the percentage of UDFAs that make the roster is much smaller than 7th rounders even.

16
by justanothersteve :: Thu, 05/19/2016 - 12:09pm

Just curious how many more Super Bowls Manning might have been in for the Colts had they kept Faulk and drafted Holt instead of James. The Greatest Show on Turf may have been 240 miles further east.