Answering detailed "what if" questions to help guide decision making is the core purpose of data analytics. Until now, insights extracted from player tracking data were only available post game—primarily because of the complexity of the algorithms and the domain-specific knowledge required to interface with the systems (i.e. scripting/coding). In this work, we bring analytics courtside for use in in-game decisions by combining data-driven ghosting with a digital sketching interface. Our framework is highly intuitive—anyone can draw a play and easily understand how a team is likely to defend against it. More importantly, because the algorithms run in realtime and the interface requires only a few stylus strokes, the system can instantly deliver objective analysis for very specific queries. To create this capability, we adapted data-driven ghosting from soccer to basketball, and devised algorithms to translate sketches into hallucinated tracking data. Our experiments illustrate the ability to which we can accurately predict how specific teams will respond in specific game situations. As a result, coaches, fans and commentators can use our tool to explore an endless number of scenarios: sketching plays to find weaknesses in a particular defensive system, editing existing tracking data to see if a better decision could have been made, or evaluating similarities/differences between how teams react to the same attack.
© Copyright 2018 Alle Rechte vorbehalten.