Saturday, 25 of February of 2017

Archives from month » September, 2014

Computer science meets football

LTU undergraduate student Bob Strange likes football, and as computer science major decided to apply computational analysis to his favorite sport game. With assistance of LTU faculty he developed a computational model that analyzed all football plays of all NFL teams in the past 10 seasons. The model was then able to predict the type of the next football play (pass or run) based on the status of the game such as the time left, down, distance to the goal, etc’. His model was able to predict the type of play with accuracy significantly higher than random guessing. He also discovered that while some teams are more consistent and therefore easier for the computer to predict, other teams are less predictable and the computer was not able to analyze with accuracy higher than just a random guess. The method can be used by teams during game to better predict the offensive play of the opponent, or to simulate the play of the opponent during practice for a game. Bob Strange’s study was reviewed by expert scientists and was accepted for publication in International Journal of Computer Science in Sport.

Patriots-Steelers_wm_com Patriots-Steelers_wm_com_small