In this case, the performance scores will be the values that each wide receiver has for six different
measures. These numbers will then be converted to a normalized value through the use of value
functions and the trade-off preferences, or swing weights, will be incorporated with these normalized
values to create the single-dimensional value function for each player. In the next sub-section, the
model hierarchy will be discussed in detail which will provide the performance scores for the multiple
objectives.
2.2 Model Hierarchy
As stated, the Multiple Objective Decision Analysis model uses six different statistics in order to
compute the value of a player at the wide receiver position. These statistics were found to be the most
important raw metrics, with these being Snap Count, Targets, Receptions, Yards per Reception,
Touchdowns, and Fumbles. There are many other statistics used to measure a player’s performance, but
these are some of the most widely used statistics. While these metrics were deemed to be the most
important, there was some manipulation performed on each one in order to give players of different
“calibers” a fair opportunity within the model. Each statistic follows a logical flow starting with snap
counts, which measures the amount of snaps a player plays for a given year, in other words their overall
game time. Maximizing their targets per snap count is the next metric portraying the ability for a player
to get open and targeted by the quarterback, in other words their receiving opportunities. Maximizing
opportunity conversion is the next category, and this is measured by receptions per target, or how many
times they catch the ball when they are targeted. This leads to yards per reception, the only metric that
was not manipulated in its raw form, measuring the yardage a receiver is able to gain off a converted
opportunity. The last two metrics are touchdowns and fumbles, both on a per reception basis, with
touchdowns being maximized to benefit a team’s scoring output and fumbles being minimized to reduce
the chance of a team turnover. The hierarchy of criteria, objectives, and values measures can be seen in
Table 1. Raw data was collected for Targets, Receptions, Receiving Yards per Reception, Touchdowns,