Goals - The ultimate objective of a soccer team is to shoot goals and prevent goals by
the opponent so as to win the match. Goals are a good measure of individual output for
forward players in particular.
Shots on goal - This output measure includes goals, but also shots kicked at the
opponent’s goal that were eventually blocked by an opponent, a teammate, the goal post
or crossbar. The advantage of shots on goal as an output measure compared to goals is
that they occur at a much higher frequency, and are less subject to luck. Shots on goal
are again most pertinent as an output measure for forward players.
Passes received - This measure counts the number of passes a player receives from his
teammates. It is a good indicator of how active the player is on the field, which in turn
is related to his fitness, and second, his teammates’ current trust in his ability to make a
valuable contribution to the game.
Ball contacts - Ball contacts is a more aggregate measure than passes received of how
involved a player is, and also reflects a player’s success in obtaining the ball.
Duels won - A duel is a situation where two players fight for the ball in direct con-
frontation. A duel counts as won if the player himself or one of his teammates finally
obtains the ball. Duels won measures physical fitness, and dedication.
Cross passes - Long and usually high pass played from the sidelines intended for
attacking players close to the opponent’s goal. This a measure of running effort and
offensive motivation.
Minutes played - Our data also include detailed information on player substitutions
during a match. The rules of the game specify that the coach can make at most three
substitutions per match, and in the majority of matches coaches make two or three sub-
stitutions, in 80% of cases in the last 30 minutes of a match (total match length is 90
minutes plus a few minutes extra time). It makes sense to view a player’s number of
minutes played as a relevant output measure. First, a player’s performance on the field
will influence the coach’s substitution decisions.21 Second, the club coach’s decision to let
a player be one of players starting the match or stay on the field for a long time depends
taken by Garicano and Palacios-Huertas (2005), who provide empirical evidence for Lazear’s (1989)
prediction that relative performance evaluations can lead to undesirable sabotage. Once we control for
constant differences between players by means of player fixed effects, our regressions show no significant
effects of nomination contest participation on either fouls suffered or fouls committed.
21 This implies that an endogeneity problem would arise if we tried to control for the number or minutes
played by using it as an explanatory variable for say a player’s number of shots on goal.
18