Welcome to PatsFans.com

NFL Draft Statistical Analysis

Discussion in 'Patriots Draft Talk' started by TeamPats, Apr 10, 2007.

Thread Status:
Not open for further replies.
  1. TeamPats

    TeamPats On the Game Day Roster

    Joined:
    Nov 23, 2004
    Messages:
    438
    Likes Received:
    1
    Ratings:
    +1 / 0 / -0

    This is a somewhat selfish thread however I really do believe it will interest many of you. I am taking a Statistics course in my MBA program and have decided to do my final project on the NFL Draft. I am going to take into account all possible measureables and see there relationship to draft position. These include:

    HT, WT, 40 time, benchpress, Wonderlic, college winning %, Program size, etc.

    I have just begun thinking about this and would love some advice. But ultimately I am going to find trends in the draft for position and overall according to these measurables. I am also going to try and compare our top 100 lists to actual results to get an idea of how strong the projectors were. If anyone knows a great website where I could easily find the data for each draftee please let me know. If people are interested, I would also be happy to post my results.
     
  2. Clonamery

    Clonamery PatsFans.com Supporter PatsFans.com Supporter

    Joined:
    Feb 23, 2005
    Messages:
    3,357
    Likes Received:
    42
    Ratings:
    +136 / 0 / -1

    You'd better post your results.

    Statistics make my head hurt real bad. Scanners bad.
     
  3. Rainy-Day Patriot

    Rainy-Day Patriot PatsFans.com Supporter PatsFans.com Supporter

    Joined:
    Dec 21, 2004
    Messages:
    48
    Likes Received:
    2
    Ratings:
    +2 / 0 / -0

    #70 Jersey

    Sounds fun!

    What tools do you have in your stats toolbox?

    Simple Linear Regression?
    Multiple Linear Regression?
    Transformations? (Log? Quadratic?)
    Influence Statistics?
    Non-Linear Regression?
    Logistic Regression?
    Composites? (Alpha? PCA? Cluster?)

    Whatever you have in your toolbox, you should have plenty to find something interesting. Do you have a theory that you want to test out?

    I'd love to see what you do.
    Good luck,
    Sean :)
     
  4. TeamPats

    TeamPats On the Game Day Roster

    Joined:
    Nov 23, 2004
    Messages:
    438
    Likes Received:
    1
    Ratings:
    +1 / 0 / -0

    Probably gonna go the route of Multiple Linear Regression as this is only an MBa course. I am actually a Electrical Engineer by trade so I know some other anylsis as well however I don't think I can use them. Should be interesting though, I am curious to know how the 40 time vs the wonderlic plays out in which is more influential and other things of that nature.
     
  5. PonyExpress

    PonyExpress In the Starting Line-Up

    Joined:
    Feb 12, 2006
    Messages:
    4,661
    Likes Received:
    21
    Ratings:
    +25 / 3 / -0

    Are you trying to determine how much athletic measurables play into where a player actually gets drafted, or using the study to form a value judgment of your own on where a player should be drafted? Either way, i look forward to your study.:)
     
  6. Box_O_Rocks

    Box_O_Rocks PatsFans.com Supporter PatsFans.com Supporter

    Joined:
    Mar 13, 2005
    Messages:
    20,550
    Likes Received:
    25
    Ratings:
    +25 / 0 / -0

    I'll see Clonamery's headache and raise him one pained *****!

    NFLDraftScout.com ($25/yr) has the best player measurables of any draftsite, though even there you will get holes in your data. I assume you are going to control for team system Tampa-2 vs. 46 vs. San Diego 3-4 vs. Pats' 3-4?
     
  7. bunzoburns

    bunzoburns Practice Squad Player

    Joined:
    Jul 9, 2005
    Messages:
    214
    Likes Received:
    0
    Ratings:
    +0 / 0 / -0

    Here is a little bit of data I pulled together for the WR position.
     

    Attached Files:

  8. TeamPats

    TeamPats On the Game Day Roster

    Joined:
    Nov 23, 2004
    Messages:
    438
    Likes Received:
    1
    Ratings:
    +1 / 0 / -0

    I'm not going to lie to you Box, ideally the Tampa-2, 46, 3-4 and so forth would be a better overall way to go, I just don't know if I would be able to do a credible comparison with team's broken down by system. I know it's a shortcut, but in one of these kinds of projects there is only so much you can do. I would have to presume that the worth of one player is way more to one team then another, however within a draft format either team could either trade up or down accordingly to take advantage of this discrepency. Hopefully this will be a sort of "middle ground" to get an overall view of the Draft as seen by the NFL. If say Cinci see Gaines available at # 12 or so they may move up in front of some 3/4 teams out of fear that other 4/3 teams might do the same.

    But overall, the biggest reason is I'm not sure I could analyze via system is I don't think I have enough football knowledge, however I would be HAPPY to keep this study a growing project and continue to post it here for any critique offered, I certainly am looking for any feedback that could help make this be worthy tool to observe Draft trends.
     
    Last edited: Apr 10, 2007
  9. TeamPats

    TeamPats On the Game Day Roster

    Joined:
    Nov 23, 2004
    Messages:
    438
    Likes Received:
    1
    Ratings:
    +1 / 0 / -0

    Thanks Bunzo, this is exactly the kind of stuff I am looking to start with. I also plan on dropping the $25 on NFLDraftScout.
     
  10. TeamPats

    TeamPats On the Game Day Roster

    Joined:
    Nov 23, 2004
    Messages:
    438
    Likes Received:
    1
    Ratings:
    +1 / 0 / -0

    Oh and ultimately the study will reflect on where a player got drafted and what reasons why, however going up until the draft I intend on trying my best to set my own overall rankings system, which is obviously gonna be a load of cr@p but the intent is to to have some form of baseline so as to when it happens I can compare it to where people went and why. I also plan on trying to set my own rankings via measurables and compare it to the experts. Again to recap though, the study will be numbers only and not take into account any game footage, but it will be interesting if perhaps some teams tend to trend towards great measurables while others do not, ie. Bengals taking talent but not character.
     
  11. Box_O_Rocks

    Box_O_Rocks PatsFans.com Supporter PatsFans.com Supporter

    Joined:
    Mar 13, 2005
    Messages:
    20,550
    Likes Received:
    25
    Ratings:
    +25 / 0 / -0

    I'm not surprised that controlling for system is beyond the current scope, that would have been a bear to tackle. As for data, it occurs to me that the scouting service running the Combine releases this information to draft sites - though I couldn't say if a fee were involved. Perhaps contacting them directly http://www.nflcombine.net/ is the better way to ensure the quality of your data and to obtain data from back years. The two scouting services the majority of teams use also should have Pro-day data...surely a struggling post-grad student could get some help from these kind folks! Good luck!
     
Thread Status:
Not open for further replies.

Share This Page

unset ($sidebar_block_show); ?>