This is a review of the research using predictive model weights conducted by Joel Brooks, Matthew Kerr, and John Guttag.
Soccer is the world’s most popular sport, however, its statistics are not as sophisticated as those used in other sports. Players are evaluated using simple stats like number of goals, shots and assists. A new method for player evaluation would be to measure their passing performance. Currently, players who are credited with assists are considered to be good passers. This needs to be expanded to include all passes, not only those that lead to goals. In order to do this sets of passes that lead to a shot on net are grouped together. Then the location of where each pass originates and ends is taken into account as well as the player involved and the outcome. In order to incorporate pass locations the field is divided up into 18 zones.
After the data is inputted into the model that relates passing strategy in a possession to shots taken. Looking at the data shows that many shot opportunities originate in the zone centered in the field in front of the penalty box. The corner areas are also featured prominently in those possessions that lead to shots on goal.
The model is then able to rank players based on their ability to complete passes that lead to a shot. In soccer most possessions do not lead to shots on goal. Therefore, players will typically have a negative ranking as the majority of their possessions will not lead to shots on goal. Only a few outstanding players will manage to achieve a positive ranking.
Using predictive model weights facilitates analysis of key areas on the field that lead to shots on goal. If a team were to collect large amounts of data regarding their own plays and enter them into the model, they would be able to build a team specific model. This would allow an in-depth analysis regarding which strategies are the most effective for their situation and players. Teams then can write their playbook to feature these strategies more prominently. The team specific model would also outline which areas of the field lead to the greatest number of shots on goal for their players. Players could then be coached in how to utilize these areas of the field more effectively when they have the ball as well as when they are not directly involved with the ball. As very few players will achieve a positive ranking it will be obvious which players to focus on during a game and which players’ skills should be analyzed in order to improve the skills of other players. Soccer is a game of strategy and this model would determine the best strategy for all players on the field to ensure the greatest number of shots.
How mature is your team’s analytics program? Take the Sports Analytics Maturity Assessment.
Learn about the Groundbreaking Sports Analytics Model coaches and sports analysts are talking about!
Learn all about Sports Analytics here.