≡ Menu

Creating a Draft Value Chart, Part II

Last week, I took another stab at creating a draft value chart. The biggest modification I made was looking at just the value provided by a player in his first five years to a team. As a result, the graph flattened, compressing the difference between the value provided by the top and bottom picks.

I’m not sure if there’s a right answer, or even what the “market” is. Almost no one thinks the high values assigned by the Jimmy Johnson draft chart are “correct.” Still, I think advanced analysts can sometimes get carried away with the idea of trading down. You may not need a dozen superstars to win, but you probably need a few, especially on offense.

Last time, I stated that we needed to take the marginal value of a player compared to that of an undrafted pick. That’s true, but after thinking it over some more, we need to do this on a yearly basis. A player providing 1 or 2 points of AV in a season obviously isn’t doing much. Therefore, I reconstructed the draft value chart by giving a player credit only for the AV he produced after 2 points of AV in each season. So if a player had an AV of 10 in each of his first 5 seasons, he gets credit for 40 points of value. Using these values produces the following chart, along with a logarithmic trendline:

After again removing the marginal value for the draft slots on the “career” level — even undrafted players, on average, will have some seasons with AV over two in their first five years — we can finalize this draft value chart. You can view all the values here. We can also compare this (in blue) to the Jimmy Johnson draft chart (in red):

We can also use this draft value chart and the Jimmy Johnson chart to examine some recent high-profile draft trades. Let’s start with two trades involving top five picks in the 2012 draft:

  • The Browns traded the 4th pick, along with picks 118, 139 and 211, to move up to the 3rd pick to select Trent Richardson. According to my data, the Browns gave up too much, giving up 33.7 points of value in exchange for 27.6 points. But the NFL chart says the Vikings actually got fleeced, as the #3 pick is worth 2200 points, compared to just 1901.5 points from the other four picks.
  • Jacksonville traded the 7th pick and the 101st pick to move up to the fifth overall slot to select Justin Blackmon. My chart says the Jaguars slightly overpaid, giving up 27.4 points of value for 24.3 points of value. Conversely, the NFL chart says the Jaguars made out well, giving up 1596 points in exchange for 1700.
  • In 2003, the Jets traded the 13th, 22nd, and 116 picks to move up to the #4 pick to select Dewayne Robertson. My chart says the Jets significantly overpaid — their three picks equaled 37.5 points of value, while the 4th pick is worth only 25.8 points. The NFL draft value chart agrees, saying New York gave up 1992 points of value for the rights to draft “Baby Sapp”, who was worth 1800 points of draft value and much less than that in terms of real NFL value.
  • In April, the Eagles sent the 15th, 114th and 172nd picks to Seattle for the right to get the 12th pick and draft Fletcher Cox. This is another situation where my chart says the Eagles overpaid (23.6 points for 18.8 points) and the NFL draft chart says the Seahawks didn’t ask for enough (giving up 1200 points for 1139 points).
  • Earlier in that draft, the Cowboys desired Morris Claiborne so strongly that they sent the 14th and 45th picks to the Rams for the 6th pick. That one is very close on the NFL chart — the 6th pick is worth 1600 points, while the 14th and 45th are worth 1550 points — but my chart says Dallas overpaid (28.2 points for 23.2 points).
  • In 2011, the Redskins had the 10th pick and could have taken Blaine Gabbert. For many reasons, they’re glad they’re not; instead, they sent that pick to Jacksonville for the 16th and 49th picks in the draft. The Jimmy Johnson chart says Jacksonville slightly overpaid to get a franchise quarterback (1410 for 1300), while the AV chart says they significantly overpaid (26.8 for 19.9).
  • Do you remember the Ravens also traded up to try to get a franchise quarterback in Joe Flacco? They had the 26th pick, and packaged the 89th and 173rd picks to move up to 18 to get Flacco. The NFL chart says this was pretty close to even — Baltimore slightly won the deal, 900 to 867 — while my chart thinks Baltimore lost out, giving up 21.7 points in exchange for 16.2.
  • In their never-ending attempt to get a pass rusher in the late ’00s, Denver sent the 21st, 86th, and 198th picks to Jacksonville for the 17th slot to grab Jarvis Moss. Both charts say the Broncos overpaid, the Johnson chart by the count of 972 for 950, and the AV chart by 22.4 for 16.6.
  • In 2008, the Jaguars coveted local star Derrick Harvey so much that they sent picks number 26, 71, 89 and 215 to Baltimore for the 8th slot. The official chart says Jacksonville won the trade, giving up 1085 points of value for 1400 points. My chart says Jacksonville lost (in more ways than one), giving up 27.7 points for 21.4.
  • One of the more complex first round trades came in 2004, when Houston sent the 40th, 71st, 103rd, and 138th picks to Tennessee for the 27th and 159th picks. The Texans wanted Jason Babin and were willing to overpay to get him. The NFL draft chart says the Texans gave up 860 points in exchange for 708, while my chart says the Titans really fleeced Houston, getting 26.9 points in exchange for only 15.9 points.

I think you can get a sense of how this works, even though it is a far from complete look at draft trades. And obviously each situation is unique, with actual teams with actual needs and roster spots making these moves in reality, not in the abstract. For the most part, it doesn’t look like the Jimmy Johnson chart is followed to the letter. Part of that is due to the rising salaries that went to rookies, and part of that is simply due to the fact that teams knew the chart was out of whack. It seems like they use that chart as a starting point, and then give a haircut to the value of the top picks. Interestingly, my chart seems to always advocate for the tram that traded down as the winner.

  • Richie

    According to my data, the Browns gave up too much, giving up 33.7 points of value in exchange for 27.6 points. But the NFL chart says the Vikings actually got fleeced, as the #3 pick is worth 2200 points, compared to just 1901.5 points from the other four picks.

    I have to believe that some (if not all) teams have developed their own charts. I think that coming up with a chart that shows the “correct” value, like you have done, is not necessarily what an NFL team would want to do. They might design some sort of rough “correct” value chart, and then they might plug in historical trades, to figure out what other teams generally do. They might even plug in past trades for specific teams in order to find opportunities to make trades where they feel they can get an advantage – particularly if it involves a player that they don’t necessary value/need.

    What about the RGIII trade? Who made off better there?

    • Chase Stuart

      I purposely avoided the RGIII trade because I plan to deal with trades of future picks in a future post.

      • Richie

        Nope. I also didn’t know about the directional play data. Thanks.

      • dnabrice

        Did you ever post it?

  • PBizzel

    Looks like the #SEAHAWKS have been Jedi-like wise to move down their 1st pick to the 57th position along w/ their off-season moves…we are right at the equilibrium point of the graph. #12s #12ING

  • Hurrah, that’s what I was looking for, what a stuff! present here at this blog, thanks admin of this web site.

  • mike

    Here’s my first attempt at a dynamic drafting combined chart.
    6 charts combined each based on one of the following:
    1)Start for 3 yrs or more
    2)Start for 5yrs or more
    3)1xpro bowl
    4)3x probowl
    5)1x All pro
    6)3x All pro

    In my opinion, too many value charts focus on a one size fits all approach. I’m not even talking about adjusting the drafting chart based upon some kind of assessment of how the draft class compares with year’s past, but in what the teams are looking for. Someone with 53 very good players all in competition for starting roles, but no starting role with an elite talent, isn’t going to have a ton of value in trading way down like the harvard sports analysis chart suggests, as they can only keep 53 players and only start 11 at a time. Conversely a team with several superstars ready to make a leap but major holes in several other spots and no special teams and depth isn’t going to want to move up and target the best chance at having only one multiple all pros, but instead may want to find lots of role players, special teams and adequate starters. I view a value chart as a rule of thumb or baseline from which to deviate, but you have to know how much deviation is acceptable as a baseline, and how that will change the expectations of what you get.

  • Just a note that similar work on this same subject appeared in the Journal of Quantitative Analysis in Sports in May 2011.

    • Chase Stuart

      Thanks, Michael.

      And not that you were implying to the contrary, but I’ll just note that my first iteration on this topic came in May 2008.

  • Pingback: NFL draft: The Buccaneers and the value of a fourth-round pick • Florida Diary()