• Shoutbox
    Active Users: 0
     
  • Notice: N/A
    NFLfans processing...
 
  • Trade Table Documentation


    ---------------------------------------Bill Dow Sports Mathematics-----------------------------------
    159 Cumberland Way Alameda CA 94502

    contact: Phred at NFLfans http://nflfans.com/x/private.php?do=newpm&u=498

    Trade Table
    Documentation


    General Approach

    Instead of trying to construct a draft pick trade value table that worked for all trades – something I determined was impossible – I decided to construct individual tables for each year and then average the values for each pick across the years. In this way, the relative depth of the draft in each year can be captured and the trades made to fit much better. Thus, the trades for 2005 all fit the 2005 table and do not necessarily fit the composite table which is a reflection of the varying depth of various drafts.

    The range of values used 10000 to 1 was taken from Jimmy Johnson’s comments at the 2002 draft. However the value of the #1 pick can vary from year to year as San Diego received 12,950 points in value from the New York Giants in 2004, and in 2005 it was valued at 8500 points in a very weak draft.

    Trades involving future draft picks were valued using the “NFL Discount System.” A one for one trade of a pick from this year for a next year pick from the previous round was considered equal. Thus a 3rd rounder from the current year was equal to a second rounder next year. In trades involving more than one current year pick and a future pick, the future pick was valued at the average value of the round before the pick. Thus a trade involving a third rounder this year for a current year 6th and a 2nd next year, would value the future pick at mid 3rd round value. Trades involving players were not used because determining player value in terms of picks is left to the experts.

    In addition, trades involving an extra pick, i.e. three for one or 4 for 2 trades, generally come up 100 points short per extra pick and so that was incorporated in the methodology.

    The general methodology used was to identify the picks involved in trades and then assign them values. The other picks were then given prorated values which equated or showed an exponential decline in the values between the picks. An effort was made to have an exponential decline in value between each pick – so that the difference between picks was equal to or less than the difference between the two preceding picks. However in many cases, a number of trades within a range of a dozen or so picks defined a range of picks as equal in value generating “flat spots” in each of the annual tables. In this case there usually was a sharp decline in values directly before and after the “flat spot”. These “flat spots” appear in different parts of the tables in different years so that over 10 years they average out. The resulting composite table in general shows the exponential form but does have some minor exceptions. Picks that were traded multiple times also presented problems. In most of these cases the trades assigned very different values to the pick. In those cases in which I couldn’t make a single value work for both trades, a value was assigned that minimized the variance in the trades.

    Incomplete data was found for the 1997 and 2000 drafts. As a result, the trades were pieced together from listings of picks and news articles. I am confident that I have all of the day one trades from these years, but I’m probably missing some day two trades. The day two trades generally don’t do much to define the table and the impact of any changes from incorporating the additional trades on the composite table would be minimal.

    Day one trades generally fit the table better than day two trades because on day two the teams have fewer picks to maneuver with to equate the trade. In particular, a number of trades were made to get into the top 10 picks in the 4th round in which the team moving up appears to have overpaid. To make the trades equal I would have to greatly increase the value of all day two picks which would throw off too many other trades. As a result, I left them out of balance. It seems that teams are willing to trade up for 3rd round value on day two. Two examples are New England (Assante Samuel) and San Francisco (Isaac Sopoaga). Note that these top ten day two picks can be different overall picks because of the varying number of 3rd round compensation picks.

    Documentation for each of the yearly tables follows.

    Bill Dow

Single Sign On provided by vBSSO