Simple in concept but actually quite complex, curling is often called "chess on ice". Like innings in baseball, curling takes place over "ends" (8 or 10). Each end comprises of 16 shots, with each shot a decision point in the contest. In a 10-end game the captain of each team, known as the "skip", makes 80 decisions. This makes for 160 decisions between two teams. In many cases the decision is universally accepted, but very often there can be disagreement on the correct called shot and one poor choice can turn the course of the game. Because of the way curling is played, each decision can be examined using analytics. What is the chance my teammate will successfully execute this shot? What will my opponent likely do next? What are the conditions of the ice and how will that effect our chance of making this shot and each subsequent shot? Based on the outcome of my decision, what is my probability of winning the game? Given the shot-by-shot structure of curling, many would consider statistical analysis as both desired and perhaps necessary for success. This is traditionally not the case. Curling is a game steeped in tradition, where parents and grandparents pass on experience to the next generation. In many instances, some of the best teams in the world stop and consider several options (while a chess timer is ticking away) when pre-planning could help clarify the correct decision much faster. Sometimes teams may make a clear strategic mistake because they simply don`t know the numbers and rely instead on gut or feel. Teams can struggle to properly evaluate their own level of play while also failing to recognize opportunities to out maneuver opponents because they lack the information or incorrectly interpret small sample sizes from their own visual analysis. In this paper we will present the current information available to teams and how these datasets can be applied to analyze their own play, prepare for the opposition and evaluate in-game decision making. We will conclude with considerations on possible next stages for curling analytics and what is needed to achieve these goals.
© Copyright 2019 MIT Sloan Sports Analytics Conference 2019. Kaikki oikeudet pidätetään. / Kääntänyt https://www.DeepL.com/Translator