≡ Menu

The 2000 NFL Draft was supposed to bring an incredible infusion of wide receiver talent. Peter Warrick, Plaxico Burress, and Travis Taylor were top-10 picks, making it one of only four classes since 1970 were three wide receivers drafted in the top ten. In addition, Sylvester Morris, R. Jay Soward, Dennis Northcutt, and Todd Pinkston all went in the top 36 picks, one of only seven classes since the merger with seven wide receivers in the top 36. Avion Black was the 20th wide receiver taken with the 121st pick: add it all up, and the 2000 draft had unmatched levels of quality and quantity. The graph below shows the amount of draft value spent on wide receivers (you can click here for value spent on wide receivers and tight ends) in each draft from 1970 to 2011:

wr draft

The 2000 Draft, however, did not live up to the hype, as anyone who read the names in the opening paragraph could have assumed. In the graph below, I’ve shown the average percentage of NFL receiving yards gained by each draft class (only including drafted players) in their first five years. This is calculated by taking the total number of receiving yards by all players in say, the 2000 Draft, in the years 2000, 2001, 2002, 2003, and 2004, and dividing that total by all receiving yards gained in those seasons. By this measure, the 2000 class gained just 9.1% of all receiving yards in their first five seasons, which ranks 37th out of 42 draft classes from 1970 to 2011:

wr production

Let’s put those two charts together:

draft production wr

If you notice anything, you probably notice that the chart seems like a jumbled mess. But that’s the point: one would think that the classes with the most draft value (like the 2000 or 2004 classes) would produce a lot of value, but that hasn’t been the case.  In fact, the correlation coefficient between draft value spent on receivers and NFL production has been negative!  That means NFL teams know even less than you think!  So the fact that this year’s draft class was a terrible one for pass catchers doesn’t mean anything about how they will turn out. The two best draft classes in terms of production were the ’74 and ’76 classes, and both were below average in terms of draft capital spent.

Now, there are some limitations in this study. It’s possible (though I think very unlikely) that teams are better at drafting now, which would invalidate some of these results. And with more receivers on the field, the percentage of yards gained by top players necessarily has been going down. This is not a perfect study by any means, and the amount of draft capital spent on wide receivers has been increasing while the amount of production by receivers has been decreasing. But even after you adjust for era, the correlation coefficient — while now positive — only rises to 0.07, which is essentially zero. Recent years like 2010, when little draft capital was spent but the class was very productive, show that the problem has not been solved.

For those curious, here’s the raw data below. I’d love to hear your thoughts.

YearDraft ValueProductionDV RkProd Rk
1970132.810.9%3522
197117012.7%188
1972143.311%3120
1973164.212.5%2010
1974135.716.6%342
1975109.410.9%4223
1976145.717%281
197712514.6%384
1978176.311.3%1216
1979118.813.3%416
1980124.410.1%3929
1981163.59.1%2138
1982154.916.3%233
198314411.1%2918
1984170.59.8%1732
1985196.912.2%914
198613611.1%3319
1987149.911.2%2617
1988199.513.5%75
1989162.28.7%2240
1990127.112.6%379
1991191.212.9%117
19921287.3%3642
1993153.39.8%2531
1994198.39.6%834
1995143.69.5%3035
1996216.112.4%311
1997141.38.7%3239
1998173.710%1430
1999154.19.3%2436
2000224.99.1%137
2001200.612.3%612
2002170.710.4%1627
2003176.210.1%1328
2004219.310.8%224
2005214.68.1%441
2006120.310.4%4025
2007211.910.4%526
2008173.311.8%1515
2009195.911%1021
2010146.512.2%2713
2011166.19.7%1933
Avg16411.2%
  • Clint

    According to the final chart, receivers are generally drafted too high, correct?

    • I don’t think that’s the right interpretation although that doesn’t mean you’re wrong. I’ll give additional thoughts later but it has to do with marginal value added.

      • Clint

        Hm. If they are drafted high, and the production is generally low, I just take that as bad drafting/drafting guys too high.

        Also, most of those receivers were regular flops, but Peter Warrick just had bad injuries. He was dangerous his first year or two. Also, RJ Soward could’ve done some work, but he was a basket case. Also, this was a real thing: http://www.scout.com/nfl/jaguars/story/1430622-is-the-curse-of-r-jay-soward-a-real-thing

        • I may not quite have understood your position. I thought you were saying teams in general shouldn’t be spending higher picks on WRs. I don’t think that’s been borne out by the data today. Were you suggesting something else?

        • Richie

          Funny. Right after that article was written, Robinson and Hurns had decent rookie seasons and then both bounced back with 1,000+ yard sophomore seasons.

          • Clint

            Exactly. Haha.

  • spidr

    two theories. maybe it’s the effect of previous year(s) WR class. e.g. a good WR entering league in year N may be competing with GREAT WRs from year N-1.

    alternatively, perhaps it’s the effect of QB play? WRs selected with a high draft pick may be playing with poor QBs, and vice-versa.

    • The WR draft value chart is very spiky, so who knows.

      I think the QB effect is probably pretty small, since so many WRs are drafted. This looks at all WRs, not just those with high picks. It’s possible there’s something going on there, but my hunch would be that any effect would be tiny.

  • Phil

    “It’s possible (though I think very unlikely) that teams are better at drafting now, which would invalidate some of these results. ”

    I’m going to put this here, because the linked article was 2 years ago
    —————–
    I’m sort of surprised that there isn’t more tangible evidence of the draft as a whole improving, what I would suggest is to look at the later rounds for additional evidence

    1972 draft had 17 rounds (there were fewer teams back then, so that’s not quite as many picks as it seems)
    did teams get any value out of their 17th round picks?

    has it become any harder to extract value from the 17th round?

    —————

    in the orginal article you posed that Jimmy Graham might have been a FA in a different era, that he lasted only to the 3rd round seems like a pretty big improvement

    if there are fewer and fewer good/great players available in later rounds

    I would suggest that signals improvement

    ————-

    it’d be interesting to know if that was actually the case

    • Thanks, Phil. I’m not quite sure I’m following, though: what sort of study would you suggest?

      • Phil

        Here’s what I was thinking:

        Total draft class realized career AV-

        at what pick is 25% of the available career AV off the board, at what pick is 50% of the available career AV off the board, at what pick is 75% of the available career AV off the board?

        then, has that changed over time?

        ——————–

        if the draft position at which 50% (or any given percentile) of the available career AV has come of the board has a trend of dropping since 1970 (or whenever), I’d suggest that that means the draft as a whole, has improved

        conversely, if the draft positioned at given percentiles, haven’t dropped, or just bounced around randomly, I’d suggest that’s evidence that the draft hasn’t improved

        ——————–

        as a follow up, (if that seems promising [or your curiosity is merely piqued]) I think you could also do a bunch of follow ups, doing the same analysis by position, to see if league as a whole has generally improved evaluating certain positions

        [I’d suspect some pretty oddball results on that one, seems like RB might look especially peculiar
        , though I suspect QB might look there are significant periods where the league shows improvement (though maybe QB would have sample size problems too)]

        • What would you define as career AV? Like, sum all the AV from every draft player, and then figure out what 25% of that number is? So having Brady fall to 199 will make the ’00 class look bad?

          • Phil

            Yeah, exactly,

            Brady probably does make the ’00 class look bad (like Warner and Romo should make the drafts that they weren’t even picked in look bad)

            But if the ’00 class is just an outlier, then that is still something we can work with

            ————–

            I might break the analysis into different eras, before the free agency era and after the free agency era

            Before FA, teams were drafting the lifetime value of guys

            Afterwards, you have to think about whether teams are just drafting the first contract (first 5 years or so)

            —————

            I’d probably start with whether 1968 to 1991 or so saw any improvement, and then think about how to measure the drafts from 1992 on (maybe just looking at AV realized in the first 5 years is the way to do that (idk))

  • TN

    I wonder if the success of the 1974 and 1976 draft classes (closely followed by 1977) had anything to do with the relaxation of the pass-defense rules in 1978. All of a sudden there were a lot more receiving yards to be had, so it makes sense that the young veteran WRs would be best able to grab them.

    • Well that wouldn’t be the case really for ’74; we’d only be looking at one year’s worth of data. And for ’76, I think they generally hit the ground running. I’m also not sure that the young WRs would be best able to grab the new receiving yards — why do you think that’s the case?

      The ’74 Draft had Stallworth, Nat Moore, Lynn Swann, and Roger Carr — those are four really good receivers, and also Dave Casper (who wasn’t included for these purposes). In ’76, Largent and Pat Tilley were 4th round picks, while Henry Marshall and Sammy White were 3rd and 2nd rounders.

      Meanwhile, ’75 was pretty weak; it had just Freddie Solomon and….. Leonard Thompson? Rick Upchurch? ’73 had Isaac Curtis and … Wallace Francis? The The ’77 Draft was a lot better, with Stanley Morgan, Wesley Walker, and Tony Hill, but it had nothing outside of those three, and even those three were more about their longevity than their first five years.

      • TN

        It’s possible I’ve misunderstood the data – are we talking about the percentage of yardage gained by those WRs in their first year, or over their careers? Nevertheless, the fact that three of the top four years happened in a four-season span seems significant.

        As for why that age group would benefit the most, it seems to me that they were entering their primes just as pass-catching was getting easier. I’d think a Nat Moore or Stanley Morgan would be better positioned to take advantage than an older, declining receiver like Biletnikoff or Warfield.

        • In this study, it’s over 5-year periods.

          It’s an interesting point, but I could see the argument that an older WR who maybe otherwise would appeared to have lost a step would benefit from the looser coverage rules.

          • Abe Froman

            That was a big part of the 1996 class longevity. Harrison and Owens were already in their early 30s when the pass interference rules changed in 2004. The next three classes seemed to benefit as well. By the time the 2009 QB protection rules came into effect, it appears the impact had already been absorbed by the “market” in this context.

            Just due to concussion attrition, I doubt we’ll ever see this kind of longevity by classes again.

  • I will say I am surprised that there is no relationship here. It makes me wonder if I am measuring this properly, so I’ll ask: is there another/better way to measure this? One would think that the WR classes with the most draft capital spent on those players would do better here.

    • Richie

      Next step is to do something similar for RB and QB. Maybe WR is an especially difficult position to evaluate.

    • Abe Froman

      It might be the data. AV and yards might not be comparable values. Would it be possible to compare positional draft AV versus % of all AV for all receivers in a class?

  • Andrew Healy

    I’m usually on board with how you approach things, as you know, but this time I’m not sure. If the goal is to test whether teams have gotten better at identifying the correct receivers earlier in the draft, I’m not seeing how this does better than looking at the hit rate for teams on picks in the first round, or first two rounds, etc.

    And there is a danger with looking at receiving yards (although I’m on board that simpler is almost always better). Suppose that teams have gotten better at moving on from sunk costs, better at being willing to give snaps to guys who have shown to be the best at playing receiver in the league, rather than former high picks. That would tend to make the relationship a downward trend, all else equal. I think that’s at least somewhat plausible, too.

    • Your input is always welcome — I am hardly in the conclusions stage, but rather in the fact finding stage.

      That said, you haven’t summarized the goal correctly. At a high level, I was curious as to whether a “bad” draft for WRs, such as 2016, actually means anything. This was one of the worst years ever for WRs in the draft: the first only went at 15, and so on down the line. Only 74% of the draft value spent on the average draft from ’13 to ’15 was spent on WRs in 2016: http://www.footballperspective.com/quarterbacks-and-defensive-tackles-dominated-the-2016-draft/

      And only 177 points of draft value was spent on WRs/TEs in general, the lowest in any draft since 1987:

      http://www.footballperspective.com/the-2016-draft-was-the-worst-for-wrstes-since-1987/

      So that leads to the goal: does a “bad” draft for WRs mean that the WR group is actually bad? There are lots of ways to measure that, but this was the most intuitive one for me. So far, the results indicate that a bad draft doesn’t mean anything. Look at 1979, when only 2 WRs went in the top 30, and only five more went in the top 100. That was a terrible draft for WRs, yet it produced Drew Hill, Roy Green, and Dwight Clark — with picks 328, 89, and 249. Jerry Butler and Earnest Gray were not stars but were solid, too. I have not included UDFAs in this study, but Steve Watson was a great one from this class, too.

      So, at least so far, the evidence makes me think that I shouldn’t be down on the 2016 class because the NFL viewed it as a weak draft. That’s my takeaway. As for your sunk costs paragraph, it’s definitely a “who knows” idea, but my gut would go the other way. It feels like teams are less likely to admit mistakes now, and more concerned with CYA than they used to be.

      • Andrew Healy

        Oh, I see. For that goal, that does seem like an intuitive way to get at it. The only thing to add then is that if teams are getting better at drafting WRs — and I’m probably a little more optimistic than you that this could be the case — the conclusion might not apply now. 25-30 years ago, I don’t see a sub 6-foot receiver being the first guy taken. Now, perhaps the best guy (Coleman) gets picked first. So I guess I’ll take the other side of the bet here: going forward, I’ll predict that classes with less draft value have less NFL production, maybe only a little less, but clearly less.

        We could go double or nothing on that beer you owe me for the Jets point differential getting even worse than its crappy 2014 level.