Future Research Activities

• Coverage Areas

  Our research on edge construction project is appropriate for coverage areas of various shapes. However, to date, we have not considered what shape is most realistic or how the parameters of the shape vary with the individual player. We intend to address these questions in the future.

• Skill Differences

  To date, our research on pass probabilities has not considered the abilities of individual players. Instead, it has considered the average behavior of all players at the macroscopic level. In the future, we intend to conduct more microscopic analyses.

• Offensive Strategies

  We intend to consider the ways in which different offensive strategies impact both the number and quality of scoring attempts (and other measures of offensive pressure).

• Automated Tagging

  Most companies currently tag/categorize events manually. While this is almost certainly the most accurate approach, it is both time consuming and tedious. In the future we intend to investigate technologies for generating tags for at least some kinds of events (including sport-specific events like through balls, personal passes, back passes in soccer and more generic events like deflected shots) automatically.

• Defensive Strategies

  We intend to consider the effectiveness of existing defensive strategies and the ways in which network optimization algorithms might be used to develop new defensive strategies. This research will ask both microscopic questions (e.g., does the number of passing lanes obstructed vary with field position, holding the "spread" of the offense constant) and macroscopic questions (e.g., does the strategy reduce the number of goals scored).

• Passes and Sequences of Passes

  We intend to consider such things as: the number of passes a team remains in control after events of different kinds (e.g., goal kicks, throw-ins, etc... in soccer), the number of passes that are to the open field (including different definitions of the open field), the probability of a shot as a function of the number of available passing lanes, what makes a receiver an attractive target (including the probability of success of subsequent passes, the number of open passing lanes for subsequent passes),