Abstract: A system to rapidly set up new claw machines for different styles of prizes, while also providing an excellent player experience through dynamically changing claw machine behavior through machine learning algorithms. These systems can be readily installed on most any crane machine by replacing the controller card. The system is configured by means of a control wand physically connected to the controller. Teaching the machine about the prize is accomplished by physically inserting a sample prize into the grabber, and executed a command via control wand to tell the machine to learn. Once configured and taught the prize, the machine is autonomous, and will maintain its programmed profit margin throughout the prizes dispensed by learning player behavior and making adjustments after each play.